US20190073788A1 - Self-learning spatial recognition system - Google Patents

Abstract

A method includes detecting a first object entering a first video frame of a plurality of video frames of a view of a geolocation and determining, from the plurality of video frames, that the first object has stopped in an area of the geolocation for at least a threshold amount of time. The method also includes detecting the first object leaving a second video frame of the plurality of video frames, and identifying, by a computer processing device, the area of the geolocation as a region of interest based on the detecting the first object leaving.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS
• This application is a continuation of U.S. patent application Ser. No. 15/982,877, filed May 17, 2018, which claims the benefit of U.S. Provisional Patent Application No. 62/520,328, filed Jun. 15, 2017, the entire contents of each of which are hereby incorporated by reference.
BACKGROUND
• This disclosure relates to a self-learning spatial recognition system and, more particularly, to a self-learning spatial recognition system in a publish-subscribe (or “PubSub”) system.
• The PubSub pattern is a data communication messaging arrangement implemented by software systems where so-called publishers publish messages to topics and so-called subscribers receive the messages pertaining to particular topics to which they are subscribed. There can be one or more publishers per topic and publishers generally have no knowledge of what subscribers, if any, will receive the published messages. Because publishers may publish large volumes of messages and subscribers may subscribe to many topics (or “channels”), the overall volume of messages directed to a particular channel and/or subscriber may be difficult to manage.
BRIEF DESCRIPTION OF THE DRAWINGS
• FIG. 1A illustrates an example system that supports the PubSub communication pattern.
• FIG. 1B illustrates functional layers of software on an example client device.
• FIG. 2 is a diagram of an example messaging system.
• FIG. 3A is a data flow diagram of an example method for writing data to a streamlet.
• FIG. 3B is a data flow diagram of an example method for reading data from a streamlet.
• FIG. 4A is a data flow diagram of an example method for publishing messages to a channel of a messaging system.
• FIG. 4B is a data flow diagram of an example method for subscribing to a channel of a messaging system.
• FIG. 4C is an example data structure for storing messages of a channel of a messaging system.
• FIG. 5A is a data flow diagram of an example method for publishing and replicating messages of a messaging system.
• FIG. 5B is a data flow diagram of an example method for retrieving stored messages in a messaging system.
• FIGS. 5C and 5D are data flow diagrams of example methods for repairing a chain of copies of data in a messaging system.
• FIG. 6 is an example data flow diagram for the application of filtering criteria in a messaging system.
• FIGS. 7A-7D are illustrations of how messages may be processed using query instructions that include a period-based parameter.
• FIG. 8 is a diagram of an example messaging system for self-learning spatial recognition in a PubSub communication system.
• FIG. 9 is a first flowchart of an example method for self-learning spatial recognition in a PubSub communication system.
• FIG. 10 is a second flowchart of an example method for self-learning spatial recognition in a PubSub communication system.
• FIG. 11 is a third flowchart of an example method for self-learning spatial recognition in a PubSub communication system.
• FIG. 12 is a diagram of an example mapping for self-learning spatial recognition in a PubSub communication system.
• FIG. 13 is a block diagram of an example computing device that may perform one or more of the operations described herein.
DETAILED DESCRIPTION
• Elements of examples or embodiments described with respect to a given aspect of the invention can be used in various embodiments of another aspect of the invention. For example, it is contemplated that features of dependent claims depending from one independent claim can be used in apparatus, systems, and/or methods of any of the other independent claims.
• The details of one or more embodiments of the subject matter described in this specification are set forth in the accompanying drawings and the description below. Other features, aspects, and advantages of the subject matter will become apparent from the description, the drawings, and the claims.
• System architecture for self-learning spatial recognition may include a messaging system. The messaging system may support the PubSub communication pattern and may allow publishers and subscribers to publish and receive live messages. Users may include both publishers and subscribers in the messaging system. For example, publishers may publish messages via the messaging system that indicate useful information to subscribers. The subscribers may view the messages and associated information. It is noted the present embodiments may also be implemented using other suitable (non-PubSub) real-time or near real-time network topologies.
• In certain embodiments, systems may track free or open geographical areas (e.g., parking lot spots) with a webcam or other suitable video or camera source if a map of the geographical area (e.g., parking lot) is available. For example, if a map of a parking lot already exists, defined parking spots may be labeled via either human intervention or computer-based artificial intelligence. However, if the parking lot is large and there are no markers or other identifiers on the parking lot surface (or the marks are not obvious or easily seen), detecting free or open parking spots in the parking lot may be problematic. Even if possible using existing methods, such methods of detecting free or open parking spots may be time-consuming, resource, and processor intensive. Accordingly, it would be desirable to provide systems and methods for self-learning spatial recognition, where current approaches cannot provide such spatial recognition without human intervention, and while minimizing computer resources and processing power.
• Advantageously, the systems and methods of self-learning spatial recognition described herein resolve the above problems, and others, by utilizing a PubSub computing network to identify regions of interest (e.g., parking spots, seats, boat slips, spots in a queue, etc.) in a geographical location (e.g., parking lots, parking structures, streets, harbors, restaurants, sports complexes, theatres, parks, open fields, shopping areas, etc.). The present embodiments may provide techniques for resource and processing power minimizing self-learning spatial recognition, and may thereby increase processing capability and/or capacity and data throughput within the PubSub computing network.
• Worth noting is that while the term “vehicle” is used throughout for convenience and simplicity, the term “vehicle” may refer to one of, for example, a bike, a skateboard, an automotive motor vehicle, a motorcycle, a scooter, a boat, an airplane, a train, a jet, or other suitable conveyance. Furthermore, the term “pedestrian” as used throughout may include both humans and animals.
• FIG. 1A illustrates anexample system100 that supports the PubSub communication pattern. Theexample system100 may support video streaming by providing messages from publishers to subscribers that include a video frame, for instance. Components of theexample system100 may also provide support for querying and updating live video streams. Publisher clients (e.g., Publisher 1) can publish messages to named channels (e.g., “Channel 1”) by way of thesystem100. A message can comprise any type of information including one or more of the following: text, image content, sound content, multimedia content, video content, binary data, and so on. Other types of message data are possible. Subscriber clients (e.g., Subscriber2) can subscribe to a named channel using thesystem100 and start receiving messages which occur after the subscription request or from a given position (e.g., a message number or time offset). A client can be both a publisher and a subscriber.
• Depending on the configuration, a PubSub system can be categorized as follows:
• • One to One (1:1). In this configuration there is one publisher and one subscriber per channel. A typical use case is private messaging.
  • One to Many (1:N). In this configuration there is one publisher and multiple subscribers per channel. Typical use cases are broadcasting messages (e.g., stock prices).
  • Many to Many (M:N). In this configuration there are many publishers publishing to a single channel. The messages are then delivered to multiple subscribers. Typical use cases are map applications.
• There is no separate operation needed to create a named channel. A channel is created implicitly when the channel is subscribed to or when a message is published to the channel. In some implementations, channel names can be qualified by a name space. A name space comprises one or more channel names. Different name spaces can have the same channel names without causing ambiguity. The name space name can be a prefix of a channel name where the name space and channel name are separated by a dot or other suitable separator. In some implementations, name spaces can be used when specifying channel authorization settings. For instance, themessaging system100 may have app1.foo and app1.system.notifications channels where “app1” is the name of the name space. The system can allow clients to subscribe and publish to the app1.foo channel. However, clients can only subscribe to, but not publish to the app1.system.notifications channel.
• FIG. 1B illustrates functional layers of software on an example client device. A client device (e.g., client102) is a data processing apparatus such as, for example, a personal computer, a laptop computer, a tablet computer, a smart phone, a smart watch, or a server computer. Other types of client devices are possible. Theapplication layer104 comprises the end-user application(s) that will integrate with thePubSub system100. Themessaging layer106 is a programmatic interface for theapplication layer104 to utilize services of thesystem100 such as channel subscription, message publication, message retrieval, user authentication, and user authorization. In some implementations, the messages passed to and from themessaging layer106 are encoded as JavaScript Object Notation (JSON) objects. Other message encoding schemes are possible.
• Theoperating system108 layer comprises the operating system software on theclient102. In various implementations, messages can be sent and received to/from thesystem100 using persistent or non-persistent connections. Persistent connections can be created using, for example, network sockets. A transport protocol such as TCP/IP layer112 implements the Transport Control Protocol/Internet Protocol communication with thesystem100 that can be used by themessaging layer106 to send messages over connections to thesystem100. Other communication protocols are possible including, for example, User Datagram Protocol (UDP). In further implementations, an optional Transport Layer Security (TLS)layer110 can be employed to ensure the confidentiality of the messages.
• FIG. 2 is a diagram of anexample messaging system100. Thesystem100 provides functionality for implementing PubSub communication patterns. The system comprises software components and storage that can be deployed at one or more data centers122 in one or more geographic locations, for example. The system comprises MX nodes (e.g., MX nodes ormultiplexer nodes202,204 and206), Q nodes (e.g., Q nodes orqueue nodes208,210 and212), one or more configuration manager nodes (e.g., configuration manager214), and optionally one or more C nodes (e.g., C nodes orcache nodes220 and222). Each node can execute in a virtual machine or on a physical machine (e.g., a data processing apparatus). Each MX node can serve as a termination point for one or more publisher and/or subscriber connections through theexternal network216. The internal communication among MX nodes, Q nodes, C nodes, and the configuration manager can be conducted over aninternal network218, for example. By way of illustration,MX node204 can be the terminus of a subscriber connection fromclient102. Each Q node buffers channel data for consumption by the MX nodes. An ordered sequence of messages published to a channel is a logical channel stream. For example, if three clients publish messages to a given channel, the combined messages published by the clients comprise a channel stream. Messages can be ordered in a channel stream, for example, by time of publication by the client, by time of receipt by an MX node, or by time of receipt by a Q node. Other ways for ordering messages in a channel stream are possible. In the case where more than one message would be assigned to the same position in the order, one of the messages can be chosen (e.g., randomly) to have a later sequence in the order. Each configuration manager node is responsible for managing Q node load, for example, by assigning channels to Q nodes and/or splitting channel streams into so-called streamlets. Streamlets are discussed further below. The optional C nodes provide caching and load removal from the Q nodes.
• In theexample messaging system100, one or more client devices (publishers and/or subscribers) establish respective persistent connections (e.g., TCP connections) to an MX node (e.g., MX node204). The MX node serves as a termination point for these connections. For instance, external messages (e.g., between respective client devices and the MX node) carried by these connections can be encoded based on an external protocol (e.g., JSON). The MX node terminates the external protocol and translates the external messages to internal communication, and vice versa. The MX nodes publish and subscribe to streamlets on behalf of clients. In this way, an MX node can multiplex and merge requests of client devices subscribing for or publishing to the same channel, thus representing multiple client devices as one, instead of one by one.
• In theexample messaging system100, a Q node (e.g., Q node208) can store one or more streamlets of one or more channel streams. A streamlet is a data buffer for a portion of a channel stream. A streamlet will close to writing when its storage is full. A streamlet will close to reading and writing and be de-allocated when its time-to-live (TTL) has expired. By way of illustration, a streamlet can have a maximum size of 1 MB and a TTL of three minutes. Different channels can have streamlets limited by different sizes and/or by different TTLs. For instance, streamlets in one channel can exist for up to three minutes, while streamlets in another channel can exist for up to 10 minutes. In various implementations, a streamlet corresponds to a computing process running on a Q node. The computing process can be terminated after the streamlet's TTL has expired, thus freeing up computing resources (for the streamlet) back to the Q node, for example.
• When receiving a publish request from a client device, an MX node (e.g., MX node204) makes a request to a configuration manager (e.g., configuration manager214) to grant access to a streamlet to write the message being published. Note, however, that if the MX node has already been granted write access to a streamlet for the channel (and the channel has not been closed to writing), the MX node can write the message to that streamlet without having to request a grant to access the streamlet. Once a message is written to a streamlet for a channel, the message can be read by MX nodes and provided to subscribers of that channel.
• Similarly, when receiving a channel subscription request from a client device, an MX node makes a request to a configuration manager to grant access to a streamlet for the channel from which messages are read. If the MX node has already been granted read access to a streamlet for the channel (and the channel's TTL has not been closed to reading), the MX node can read messages from the streamlet without having to request a grant to access the streamlet. The read messages can then be forwarded to client devices that have subscribed to the channel. In various implementations, messages read from streamlets are cached by MX nodes so that MX nodes can reduce the number of times needed to read from the streamlets.
• By way of illustration, an MX node can request a grant from the configuration manager that allows the MX node to store a block of data into a streamlet on a particular Q node that stores streamlets of the particular channel. Example streamlet grant request and grant data structures are as follows:

• 	StreamletGrantRequest = {
  	“channel”: string( )
  	“mode”: “read” | “write”
  	“position”: 0
  	}
  	StreamletGrantResponse = {
  	“streamlet-id”: “abcdef82734987”,
  	“limit-size”: 2000000, # 2 megabytes max
  	“limit-msgs”: 5000, # 5 thousand messages max
  	“limit-life”: 4000, # the grant is valid for 4 seconds
  	“q-node”: string( )
  	“position”: 0
  	}

• The StreamletGrantRequest data structure stores the name of the stream channel and a mode indicating whether the MX node intends on reading from or writing to the streamlet. The MX node sends the StreamletGrantRequest to a configuration manager node. The configuration manager node, in response, sends the MX node a StreamletGrantResponse data structure. The StreamletGrantResponse contains an identifier of the streamlet (streamlet-id), the maximum size of the streamlet (limit-size), the maximum number of messages that the streamlet can store (limit-msgs), the TTL (limit-life), and an identifier of a Q node (q-node) on which the streamlet resides. The StreamletGrantRequest and StreamletGrantResponse can also have a position field that points to a position in a streamlet (or a position in a channel) for reading from the streamlet.
• A grant becomes invalid once the streamlet has closed. For example, a streamlet is closed to reading and writing once the streamlet's TTL has expired and a streamlet is closed to writing when the streamlet's storage is full. When a grant becomes invalid, the MX node can request a new grant from the configuration manager to read from or write to a streamlet. The new grant will reference a different streamlet and will refer to the same or a different Q node depending on where the new streamlet resides.
• FIG. 3A is a data flow diagram of an example method for writing data to a streamlet in various embodiments. InFIG. 3A, when an MX node (e.g., MX node202) request to write to a streamlet is granted by a configuration manager (e.g., configuration manager214), as described before, the MX node establishes a Transmission Control Protocol (TCP) connection with the Q node (e.g., Q node208) identified in the grant response received from the configuration manager (302). A streamlet can be written concurrently by multiple write grants (e.g., for messages published by multiple publisher clients). Other types of connection protocols between the MX node and the Q node are possible.
• The MX node then sends a prepare-publish message with an identifier of a streamlet that the MX node wants to write to the Q node (304). The streamlet identifier and Q node identifier can be provided by the configuration manager in the write grant as described earlier. The Q node hands over the message to a handler process301 (e.g., a computing process running on the Q node) for the identified streamlet (306). The handler process can send to the MX node an acknowledgement (308). After receiving the acknowledgement, the MX node starts writing (publishing) messages (e.g.,310,312,314, and318) to the handler process, which in turn stores the received data in the identified streamlet. The handler process can also send acknowledgements (316,320) to the MX node for the received data. In some implementations, acknowledgements can be piggy-backed or cumulative. For instance, the handler process can send to the MX node an acknowledgement for every predetermined amount of data received (e.g., for every 100 messages received) or for every predetermined time period (e.g., for every one millisecond). Other acknowledgement scheduling algorithms, such as Nagle's algorithm, can be used.
• If the streamlet can no longer accept published data (e.g., when the streamlet is full), the handler process sends a Negative-Acknowledgement (NAK) message (330) indicating a problem, following by an EOF (end-of-file) message (332). In this way, the handler process closes the association with the MX node for the publish grant. The MX node can then request a write grant for another streamlet from a configuration manager if the MX node has additional messages to store.
• FIG. 3B is a data flow diagram of an example method for reading data from a streamlet in various embodiments. InFIG. 3B, an MX node (e.g., MX node204) sends to a configuration manager (e.g., configuration manager214) a request for reading a particular channel starting from a particular message or time offset in the channel. The configuration manager returns to the MX node a read grant including an identifier of a streamlet containing the particular message, a position in the streamlet corresponding to the particular message, and an identifier of a Q node (e.g., Q node208) containing the particular streamlet. The MX node then establishes a TCP connection with the Q node (352). Other types of connection protocols between the MX node and the Q node are possible.
• The MX node then sends to the Q node a subscribe message (354) with the identifier of the streamlet (in the Q node) and the position in the streamlet from which the MX node wants to read (356). The Q node hands over the subscribe message to ahandler process351 for the streamlet (356). The handler process can send to the MX node an acknowledgement (358). The handler process then sends messages (360,364,366), starting at the position in the streamlet, to the MX node. In some implementations, the handler process can send all of the messages in the streamlet to the MX node. After sending the last message in a particular streamlet, the handler process can send a notification of the last message to the MX node. The MX node can send to the configuration manager another request for another streamlet containing a next message in the particular channel.
• If the particular streamlet is closed (e.g., after its TTL has expired), the handler process can send an unsubscribe message (390), followed by an EOF message (392), to close the association with the MX node for the read grant. The MX node can close the association with the handler process when the MX node moves to another streamlet for messages in the particular channel (e.g., as instructed by the configuration manager). The MX node can also close the association with the handler process if the MX node receives an unsubscribe message from a corresponding client device.
• In various implementations, a streamlet can be written into and read from at the same time instance. For example, there can be a valid read grant and a valid write grant at the same time instance. In various implementations, a streamlet can be read concurrently by multiple read grants (e.g., for channels subscribed to by multiple publisher clients). The handler process of the streamlet can order messages from concurrent write grants based on, for example, time-of-arrival, and store the messages based on the order. In this way, messages published to a channel from multiple publisher clients can be serialized and stored in a streamlet of the channel.
• In themessaging system100, one or more C nodes (e.g., C node220) can offload data transfers from one or more Q nodes. For instance, if there are many MX nodes requesting streamlets from Q nodes for a particular channel, the streamlets can be offloaded and cached in one or more C nodes. The MX nodes (e.g., as instructed by read grants from a configuration manager) can read the streamlets from the C nodes instead.
• As described above, messages for a channel in themessaging system100 are ordered in a channel stream. A configuration manager (e.g., configuration manager214) splits the channel stream into fixed-sized streamlets that each reside on a respective Q node. In this way, storing a channel stream can be shared among many Q nodes; each Q node stores a portion (one or more streamlets) of the channel stream. More particularly, a streamlet can be stored in, for example, registers and/or dynamic memory elements associated with a computing process on a Q node, thus avoiding the need to access persistent, slower storage devices such as hard disks. This results in faster message access. The configuration manager can also balance load among Q nodes in themessaging system100 by monitoring respective workloads of the Q nodes and allocating streamlets in a way that avoids overloading any one Q node.
• In various implementations, a configuration manager maintains a list identifying each active streamlet, the respective Q node on which the streamlet resides, an identification of the position of the first message in the streamlet, and whether the streamlet is closed for writing. In some implementations, Q nodes notify the configuration manager and/or any MX nodes that are publishing to a streamlet that the streamlet is closed due to being full or when the streamlet's TTL has expired. When a streamlet is closed, the streamlet remains on the configuration manager's list of active streamlets until the streamlet's TTL has expired so that MX nodes can continue to retrieve messages from the streamlet.
• When an MX node requests a write grant for a given channel and there is not a streamlet for the channel that can be written to, the configuration manager allocates a new streamlet on one of the Q nodes and returns the identity of the streamlet and the Q node in the StreamletGrantResponse. Otherwise, the configuration manager returns the identity of the currently open for writing streamlet and corresponding Q node in the StreamletGrantResponse. MX nodes can publish messages to the streamlet until the streamlet is full or the streamlet's TTL has expired, after which a new streamlet can be allocated by the configuration manager.
• When an MX node requests a read grant for a given channel and there is not a streamlet for the channel that can be read from, the configuration manager allocates a new streamlet on one of the Q nodes and returns the identity of the streamlet and the Q node in the StreamletGrantResponse. Otherwise, the configuration manager returns the identity of the streamlet and Q node that contains the position from which the MX node wishes to read. The Q node can then begin sending messages to the MX node from the streamlet beginning at the specified position until there are no more messages in the streamlet to send. When a new message is published to a streamlet, MX nodes that have subscribed to that streamlet will receive the new message. If a streamlet's TTL has expired, thehandler process351 can send an EOF message (392) to any MX nodes that are subscribed to the streamlet.
• In some implementations, themessaging system100 can include multiple configuration managers (e.g.,configuration manager214 plus one or more other configuration managers). Multiple configuration managers can provide resiliency and prevent single point of failure. For instance, one configuration manager can replicate lists of streamlets and current grants it maintains to another “slave” configuration manager. As another example, multiple configuration managers can coordinate operations between them using distributed consensus protocols, such as, for example, Paxos or Raft protocols.
• FIG. 4A is a data flow diagram of an example method for publishing messages to a channel of a messaging system. InFIG. 4A, publishers (e.g.,publisher clients402,404,406) publish messages to themessaging system100 described earlier in reference toFIG. 2. For instance,publishers402 respectively establish connections411 and send publish requests to theMX node202.Publishers404 respectively establishconnections413 and send publish requests to theMX node206.Publishers406 respectively establishconnections415 and send publish requests to theMX node204. Here, the MX nodes can communicate (417) with a configuration manager (e.g., configuration manager214) and one or more Q nodes (e.g.,Q nodes212 and208) in themessaging system100 via theinternal network218.
• By way of illustration, each publish request (e.g., in JSON key/value pairs) from a publisher to an MX node includes a channel name and a message. The MX node (e.g., MX node202) can assign the message in the publish request to a distinct channel in themessaging system100 based on the channel name (e.g., “foo”) of the publish request. The MX node can confirm the assigned channel with theconfiguration manager214. If the channel (specified in the subscribe request) does not yet exist in themessaging system100, the configuration manager can create and maintain a new channel in themessaging system100. For instance, the configuration manager can maintain a new channel by maintaining a list identifying each active streamlet of the channel's stream, the respective Q node on which the streamlet resides, and identification of the positions of the first and last messages in the streamlet as described earlier.
• For messages of a particular channel, the MX node can store the messages in one or more buffers or streamlets in themessaging system100. For instance, theMX node202 receives from thepublishers402 requests to publish messages M11, M12, M13, and M14 to a channel foo. TheMX node206 receives from thepublishers404 requests to publish messages M78 and M79 to the channel foo. TheMX node204 receives from thepublishers406 requests to publish messages M26, M27, M28, M29, M30, and M31 to the channel foo.
• The MX nodes can identify one or more streamlets for storing messages for the channel foo. As described earlier, each MX node can request a write grant from theconfiguration manager214 that allows the MX node to store the messages in a streamlet of the channel foo. For instance, theMX node202 receives a grant from theconfiguration manager214 to write messages M11, M12, M13, and M14 to astreamlet4101 on theQ node212. TheMX node206 receives a grant from theconfiguration manager214 to write messages M78 and M79 to thestreamlet4101. Here, thestreamlet4101 is the last one (at the moment) of a sequence of streamlets of thechannel stream430 storing messages of the channel foo. Thestreamlet4101 has messages (421) of the channel foo that were previously stored in thestreamlet4101, but is still open, i.e., thestreamlet4101 still has space for storing more messages and the streamlet's TTL has not expired.
• TheMX node202 can arrange the messages for the channel foo based on the respective time that each message was received by theMX node202, e.g., M11, M13, M14, M12 (422), and store the received messages as arranged in thestreamlet4101. That is, theMX node202 receives M11 first, followed by M13, M14, and M12. Similarly, theMX node206 can arrange the messages for the channel foo based on their respective time that each message was received by theMX node206, e.g., M78, M79 (423), and store the received messages as arranged in thestreamlet4101. Other arrangements or ordering of the messages for the channel are possible.
• The MX node202 (or MX node206) can store the received messages using the method for writing data to a streamlet described earlier in reference toFIG. 3A, for example. In various implementations, the MX node202 (or MX node206) can buffer (e.g., in a local data buffer) the received messages for the channel foo and store the received messages in a streamlet for the channel foo (e.g., streamlet4101) when the buffered messages reach a predetermined number or size (e.g., 100 messages) or when a predetermined time (e.g., 50 milliseconds) has elapsed. For instance, theMX node202 can store in thestreamlet 100 messages at a time or in every 50 milliseconds. Other appropriate algorithms and techniques, such as Nagle's algorithm, can be used for managing the buffered messages.
• In various implementations, the Q node212 (e.g., a handler process) stores the messages of the channel foo in thestreamlet4101 in the order as arranged by theMX node202 andMX node206. TheQ node212 stores the messages of the channel foo in thestreamlet4101 in the order theQ node212 receives the messages. For instance, assume that theQ node212 receives messages M78 (from the MX node206) first, followed by messages M11 and M13 (from the MX node202), M79 (from the MX node206), and M14 and M12 (from the MX node202). TheQ node212 stores in thestreamlet4101 the messages in the order as received, e.g., M78, M11, M13, M79, M14, and M12, immediately after themessages421 that are already stored in thestreamlet4101. In this way, messages published to the channel foo from multiple publishers (e.g.,402,404) can be serialized in a particular order and stored in thestreamlet4101 of the channel foo. Different subscribers that subscribe to the channel foo will receive messages of the channel foo in the same particular order, as will be described in more detail in reference toFIG. 4B.
• In the example ofFIG. 4A, at a time instance after the message M12 was stored in thestreamlet4101, theMX node204 requests a grant from theconfiguration manager214 to write to the channel foo. Theconfiguration manager214 provides the MX node204 a grant to write messages to thestreamlet4101, as thestreamlet4101 is still open for writing. TheMX node204 arranges the messages for the channel foo based on the respective time that each message was received by theMX node204, e.g., M26, M27, M31, M29, M30, M28 (424), and stores the messages as arranged for the channel foo.
• By way of illustration, assume that the message M26 is stored to the last available position of thestreamlet4101. As thestreamlet4101 is now full, theQ node212 sends to the MX node204 a NAK message, following by an EOF message, to close the association with theMX node204 for the write grant, as described earlier in reference toFIG. 3A. TheMX node204 then requests another write grant from theconfiguration manager214 for additional messages (e.g., M27, M31, and so on) for the channel foo.
• Theconfiguration manager214 can monitor available Q nodes in themessaging system100 for their respective workloads (e.g., how many streamlets are residing in each Q node). Theconfiguration manager214 can allocate a streamlet for the write request from theMX node204 such that overloading (e.g., too many streamlets or too many read or write grants) can be avoided for any given Q node. For instance, theconfiguration manager214 can identify a least loaded Q node in themessaging system100 and allocate a new streamlet on the least loaded Q node for write requests from theMX node204. In the example ofFIG. 4A, theconfiguration manager214 allocates anew streamlet4102 on theQ node208 and provides a write grant to theMX node204 to write messages for the channel foo to thestreamlet4102. As shown inFIG. 4A, the Q node stores in thestreamlet4102 the messages from theMX node204 in an order as arranged by the MX node204: M27, M31, M29, M30, and M28 (assuming that there is no other concurrent write grant for thestreamlet4102 at the moment).
• When theconfiguration manager214 allocates a new streamlet (e.g., streamlet4102) for a request for a grant from an MX node (e.g., MX node204) to write to a channel (e.g., foo), theconfiguration manager214 assigns to the streamlet its TTL, which will expire after TTLs of other streamlets that are already in the channel's stream. For instance, theconfiguration manager214 can assign to each streamlet of the channel foo's channel stream a TTL of 3 minutes when allocating the streamlet. That is, each streamlet will expire 3 minutes after it is allocated (created) by theconfiguration manager214. Since a new streamlet is allocated after a previous streamlet is closed (e.g., filled entirely or expired), in this way, the channel foo's channel stream comprises streamlets that each expires sequentially after its previous streamlet expires. For instance, as shown in anexample channel stream430 of the channel foo inFIG. 4A,streamlet4098 and streamlets before4098 have expired (as indicated by the dotted-lined gray-out boxes). Messages stored in these expired streamlets are not available for reading for subscribers of the channel foo.Streamlets4099,4100,4101, and4102 are still active (not expired). Thestreamlets4099,4100, and4101 are closed for writing, but still are available for reading. Thestreamlet4102 is available for reading and writing, at the moment when the message M28 was stored in thestreamlet4102. At a later time, thestreamlet4099 will expire, following by thestreamlets4100,4101, and so on.
• FIG. 4B is a data flow diagram of an example method for subscribing to a channel of a messaging system. InFIG. 4B, asubscriber480 establishes aconnection462 with anMX node461 of themessaging system100.Subscriber482 establishes aconnection463 with theMX node461.Subscriber485 establishes aconnection467 with anMX node468 of themessaging system100. Here, theMX nodes461 and468 can respectively communicate (464) with theconfiguration manager214 and one or more Q nodes in themessaging system100 via theinternal network218.
• A subscriber (e.g., subscriber480) can subscribe to the channel foo of themessaging system100 by establishing a connection (e.g.,462) and sending a request for subscribing to messages of the channel foo to an MX node (e.g., MX node461). The request (e.g., in JSON key/value pairs) can include a channel name, such as, for example, “foo.” When receiving the subscribe request, theMX node461 can send to the configuration manager214 a request for a read grant for a streamlet in the channel foo's channel stream.
• By way of illustration, assume that at the current moment the channel foo'schannel stream431 includesactive streamlets4102,4103, and4104, as shown inFIG. 4B. Thestreamlets4102 and4103 each are full. Thestreamlet4104 stores messages of the channel foo, including the last message (at the current moment) stored at aposition47731.Streamlets4101 and streamlets before4101 are invalid, as their respective TTLs have expired. Note that the messages M78, M11, M13, M79, M14, M12, and M26 stored in thestreamlet4101, described earlier in reference toFIG. 4A, are no longer available for subscribers of the channel foo, since thestreamlet4101 is no longer valid, as its TTL has expired. As described earlier, each streamlet in the channel foo's channel stream has a TTL of 3 minutes, thus only messages (as stored in streamlets of the channel foo) that are published to the channel foo (i.e., stored into the channel's streamlets) no earlier than 3 minutes from the current time can be available for subscribers of the channel foo.
• TheMX node461 can request a read grant for all available messages in the channel foo, for example, when thesubscriber480 is a new subscriber to the channel foo. Based on the request, theconfiguration manager214 provides the MX node461 a read grant to the streamlet4102 (on the Q node208) that is the earliest streamlet in the active streamlets of the channel foo (i.e., the first in the sequence of the active streamlets). TheMX node461 can retrieve messages in thestreamlet4102 from theQ node208, using the method for reading data from a streamlet described earlier in reference toFIG. 3B, for example. Note that the messages retrieved from thestreamlet4102 maintain the same order as stored in thestreamlet4102. However, other arrangements or ordering of the messages in the streamlet are possible. In various implementations, when providing messages stored in thestreamlet4102 to theMX node461, theQ node208 can buffer (e.g., in a local data buffer) the messages and send the messages to theMX node461 when the buffer messages reach a predetermined number or size (e.g., 200 messages) or a predetermined time (e.g., 50 milliseconds) has elapsed. For instance, theQ node208 can send the channel foo's messages (from the streamlet4102) to theMX node461 200 messages at a time or in every 50 milliseconds. Other appropriate algorithms and techniques, such as Nagle's algorithm, can be used for managing the buffered messages.
• After receiving the last message in thestreamlet4102, theMX node461 can send an acknowledgement to theQ node208, and send to theconfiguration manager214 another request (e.g., for a read grant) for the next streamlet in the channel stream of the channel foo. Based on the request, theconfiguration manager214 provides the MX node461 a read grant to the streamlet4103 (on Q node472) that logically follows thestreamlet4102 in the sequence of active streamlets of the channel foo. TheMX node461 can retrieve messages stored in thestreamlet4103, e.g., using the method for reading data from a streamlet described earlier in reference toFIG. 3B, until it retrieves the last message stored in thestreamlet4103. TheMX node461 can send to theconfiguration manager214 yet another request for a read grant for messages in the next streamlet4104 (on Q node474). After receiving the read grant, theMX node461 retrieves messages of the channel foo stored in thestreamlet4104, until the last message at theposition47731. Similarly, theMX node468 can retrieve messages from thestreamlets4102,4103, and4104 (as shown with dotted arrows inFIG. 4B), and provide the messages to thesubscriber485.
• TheMX node461 can send the retrieved messages of the channel foo to the subscriber480 (via the connection462) while receiving the messages from theQ nodes208,472, or474. In various implementations, theMX node461 can store the retrieved messages in a local buffer. In this way, the retrieved messages can be provided to another subscriber (e.g., subscriber482) when the other subscriber subscribes to the channel foo and requests the channel's messages. TheMX node461 can remove messages stored in the local buffer that each has a time of publication that has exceeded a predetermined time period. For instance, theMX node461 can remove messages (stored in the local buffer) with respective times of publication exceeding 3 minutes. In some implementations, the predetermined time period for keeping messages in the local buffer onMX node461 can be the same as or similar to the time-to-live duration of a streamlet in the channel foo's channel stream, since at a given moment, messages retrieved from the channel's stream do not include those in streamlets having respective times-to-live that had already expired.
• The messages retrieved from thechannel stream431 and sent to the subscriber480 (by the MX node461) are arranged in the same order as the messages were stored in the channel stream, although other arrangements or ordering of the messages are possible. For instance, messages published to the channel foo are serialized and stored in thestreamlet4102 in a particular order (e.g., M27, M31, M29, M30, and so on), then stored subsequently in thestreamlet4103 and thestreamlet4104. The MX node retrieves messages from thechannel stream431 and provides the retrieved messages to thesubscriber480 in the same order as the messages are stored in the channel stream: M27, M31, M29, M30, and so on, followed by ordered messages in thestreamlet4103, and followed by ordered messages in thestreamlet4104.
• Instead of retrieving all available messages in thechannel stream431, theMX node461 can request a read grant for messages stored in thechannel stream431 starting from a message at particular position, e.g.,position47202. For instance, theposition47202 can correspond to an earlier time instance (e.g., 10 seconds before the current time) when thesubscriber480 was last subscribing to the channel foo (e.g., via a connection to theMX node461 or another MX node of the messaging system100). TheMX node461 can send to the configuration manager214 a request for a read grant for messages starting at theposition47202. Based on the request, theconfiguration manager214 provides the MX node461 a read grant to the streamlet4104 (on the Q node474) and a position on thestreamlet4104 that corresponds to thechannel stream position47202. TheMX node461 can retrieve messages in thestreamlet4104 starting from the provided position, and send the retrieved messages to thesubscriber480.
• As described above in reference toFIGS. 4A and 4B, messages published to the channel foo are serialized and stored in the channel's streamlets in a particular order. Theconfiguration manager214 maintains the ordered sequence of streamlets as they are created throughout their respective times-to-live. Messages retrieved from the streamlets by an MX node (e.g.,MX node461, or MX node468) and provided to a subscriber can be, in some implementations, in the same order as the messages are stored in the ordered sequence of streamlets. In this way, messages sent to different subscribers (e.g.,subscriber480,subscriber482, or subscriber485) can be in the same order (as the messages are stored in the streamlets), regardless which MX nodes the subscribers are connected to.
• In various implementations, a streamlet stores messages in a set of blocks of messages. Each block stores a number of messages. For instance, a block can store two hundred kilobytes of messages (although other sizes of blocks of messages are possible). Each block has its own time-to-live, which can be shorter than the time-to-live of the streamlet holding the block. Once a block's TTL has expired, the block can be discarded from the streamlet holding the block, as described in more detail below in reference toFIG. 4C.
• FIG. 4C is an example data structure for storing messages of a channel of a messaging system. As described with the channel foo in reference toFIGS. 4A and 4B, assume that at the current moment the channel foo'schannel stream432 includesactive streamlets4104 and4105, as shown inFIG. 4C.Streamlet4103 and streamlets before4103 are invalid, as their respective TTLs have expired. Thestreamlet4104 is already full for its capacity (e.g., as determined by a corresponding write grant) and is closed for additional message writes. Thestreamlet4104 is still available for message reads. Thestreamlet4105 is open and is available for message writes and reads.
• By way of illustration, the streamlet4104 (e.g., a computing process running on theQ node474 shown inFIG. 4B) currently holds two blocks of messages.Block494 holds messages from channel positions47301 to47850.Block495 holds messages from channel positions47851 to48000. The streamlet4105 (e.g., a computing process running on another Q node in the messaging system100) currently holds two blocks of messages.Block496 holds messages from channel positions48001 to48200.Block497 holds messages starting fromchannel position48201, and still accepts additional messages of the channel foo.
• When thestreamlet4104 was created (e.g., by a write grant), a first block (sub-buffer)492 was created to store messages, e.g., from channel positions47010 to47100. Later on, after theblock492 had reached its capacity, anotherblock493 was created to store messages, e.g., from channel positions47111 to47300.Blocks494 and495 were subsequently created to store additional messages. Afterwards, thestreamlet4104 was closed for additional message writes, and thestreamlet4105 was created with additional blocks for storing additional messages of the channel foo.
• In this example, the respective TTL's ofblocks492 and493 had expired. The messages stored in these two blocks (from channel positions47010 to47300) are no longer available for reading by subscribers of the channel foo. Thestreamlet4104 can discard these two expired blocks, e.g., by de-allocating the memory space for theblocks492 and493. Theblocks494 or495 could become expired and be discarded by thestreamlet4104, before thestreamlet4104 itself becomes invalid. Alternatively,streamlet4104 itself could become invalid before theblocks494 or495 become expired. In this way, a streamlet can hold one or more blocks of messages, or contain no block of messages, depending on respective TTLs of the streamlet and blocks, for example.
• A streamlet, or a computing process running on a Q node in themessaging system100, can create a block for storing messages of a channel by allocating a certain size of memory space from the Q node. The streamlet can receive, from an MX node in themessaging system100, one message at a time and store the received message in the block. Alternatively, the MX node can assemble (i.e., buffer) a group of messages and send the group of messages to the Q node. The streamlet can allocate a block of memory space (from the Q node) and store the group of messages in the block. The MX node can also perform compression on the group of messages, e.g., by removing a common header from each message or performing other suitable compression techniques.
• As described above, a streamlet (a data buffer) residing on a Q node stores messages of a channel in themessaging system100. To prevent failure of the Q node (a single point failure) that can cause messages being lost, themessaging system100 can replicate messages on multiple Q nodes, as described in more detail below.
• FIG. 5A is a data flow diagram of anexample method500 for publishing and replicating messages of themessaging system100. As described earlier in reference toFIG. 4A, theMX node204 receives messages (of the channel foo) from thepublishers406. Theconfiguration manager214 can instruct the MX Node204 (e.g., with a write grant) to store the messages in thestreamlet4102 on theQ node208. InFIG. 5A, instead of storing the messages on a single node (e.g., Q node208), theconfiguration manager214 allocates multiple Q nodes to store multiple copies of thestreamlet4102 on these Q nodes.
• By way of illustration, theconfiguration manager214 allocatesQ nodes208,502,504, and506 in themessaging system100 to store copies of thestreamlet4102. Theconfiguration manager214 instructs theMX node204 to transmit the messages for the channel foo (e.g., messages M27, M31, M29, M30, and M28) to the Q node208 (512). A computing process running on theQ node208 stores the messages in the first copy (copy #1) of thestreamlet4102. Instead of sending an acknowledgement message to theMX node204 after storing the messages, theQ node208 forwards the messages to the Q node502 (514). A computing process running on theQ node502 stores the messages in another copy (copy #2) of thestreamlet4102. Meanwhile, theQ node502 forwards the messages to the Q node504 (516). A computing process running on theQ node504 stores the messages in yet another copy (copy #3) of thestreamlet4102. TheQ node504 also forwards the message to the Q node506 (518). A computing process running on theQ node506 stores the messages in yet another copy (copy #4) of thestreamlet4102. TheQ node506 can send an acknowledgement message to theMX node204, indicating that all the messages (M27, M31, M29, M30, and M28) have been stored successfully instreamlet copies #1, #2, #3 and #4.
• In some implementations, after successfully storing the last copy (copy #4), theQ node506 can send an acknowledgement to its upstream Q node (504), which in turns sends an acknowledgement to its upstream Q node (502), and so on, until the acknowledgement is sent to theQ node208 storing the first copy (copy #1). TheQ node208 can send an acknowledgement message to theMX node204, indicating that all messages have been stored successfully in the streamlet4102 (i.e., in thecopies #1, #2, #3 and #4).
• In this way, four copies of the streamlet4102 (and each message in the streamlet) are stored in four different Q nodes. Other numbers (e.g., two, three, five, or other suitable number) of copies of a streamlet are also possible. In the present illustration, the four copies form a chain of copies including a head copy in thecopy #1 and a tail copy in thecopy #4. When a new message is published to thestreamlet4102, the message is first stored in the head copy (copy #1) on theQ node208. The message is then forwarded downstream to the next adjacent copy, thecopy #2 on theQ node502 for storage, then to thecopy #3 on theQ node504 for storage, until the message is stored in the tail copy thecopy #4 on theQ node506.
• In addition to storing and forwarding by messages, the computing processes running on Q nodes that store copies of a streamlet can also store and forward messages by blocks of messages, as described earlier in reference toFIG. 4C. For instance, the computing process storing thecopy #1 of thestreamlet4102 onQ node208 can allocate memory and store a block of, for example, 200 kilobytes of messages (although other sizes of blocks of messages are possible), and forward the block of messages to the next adjacent copy (copy #2) of the chain for storage, and so on, until the block messages is stored in the tail copy (copy #4) on theQ node506.
• Messages of thestreamlet4102 can be retrieved and delivered to a subscriber of the channel foo from one of the copies of thestreamlet4102.FIG. 5B is a data flow diagram of anexample method550 for retrieving stored messages in themessaging system100. For instance, thesubscriber480 can send a request for subscribing to messages of the channel to theMX node461, as described earlier in reference toFIG. 4B. Theconfiguration manager214 can provide to the MX node461 a read grant for one of the copies of thestreamlet4102. TheMX node461 can retrieve messages of thestreamlet4102 from one of the Q nodes storing a copy of thestreamlet4102, and provide the retrieved messages to thesubscriber480. For instance, theMX node461 can retrieve messages from the copy #4 (the tail copy) stored on the Q node506 (522). As for another example, theMX node461 can retrieve messages from thecopy #2 stored on the Q node502 (524). In this way, the multiple copies of a streamlet (e.g., copies #1, #2, #3, and #4 of the streamlet4102) provide replication and redundancy against failure if only one copy of the streamlet were stored in themessaging system100. In various implementations, theconfiguration manager214 can balance workloads among the Q nodes storing copies of thestreamlet4102 by directing the MX node461 (e.g., with a read grant) to a particular Q node that has, for example, less current read and write grants as compared to other Q nodes storing copies of thestreamlet4102.
• A Q node storing a particular copy in a chain of copies of a streamlet may fail, e.g., a computing process on the Q node storing the particular copy may freeze. Other failure modes of a Q node are possible. An MX node can detect a failed node (e.g., from non-responsiveness of the failed node) and report the failed node to a configuration manager in the messaging system100 (e.g., configuration manager214). A peer Q node can also detect a failed Q node and report the failed node to the configuration manager. For instance, an upstream Q node may detect a failed downstream Q node when the downstream Q node is non-responsive, e.g., fails to acknowledge a message storage request from the upstream Q node as described earlier. It is noted that failure of a Q node storing a copy of a particular streamlet of a particular channel stream does not have to be for publish or subscribe operations of the particular streamlet or of the particular channel stream. Failure stemming from operations on another streamlet or another channel stream can also alert a configuration manager about failure of a Q node in themessaging system100.
• When a Q node storing a particular copy in a chain of copies of a streamlet fails, a configuration manager in themessaging system100 can repair the chain by removing the failed node, or by inserting a new node for a new copy into the chain, for example.FIGS. 5C and 5D are data flow diagrams of example methods for repairing a chain of copies of a streamlet in themessaging system100. InFIG. 5C, for instance, after detecting that theQ node504 fails, theconfiguration manager214 can repair the chain of copies by redirecting messages intended to be stored in thecopy #3 of thestreamlet4102 on theQ node502 to thecopy #4 of thestreamlet4102 on theQ node506. In this example, a message (or a block of messages) is first sent from theMX node204 to theQ node208 for storage in thecopy #1 of the streamlet4102 (572). The message then is forwarded to theQ node502 for storage in thecopy #2 of the streamlet4102 (574). The message is then forwarded to theQ node506 for storage in thecopy #4 of the streamlet4102 (576). TheQ node506 can send an acknowledgement message to theconfiguration manager214 indicating that the message has been stored successfully.
• Here, a failed node can also be the node storing the head copy or the tail copy of the chain of copies. For instance, if theQ node208 fails, theconfiguration manager214 can instruct theMX node204 first to send the message to theQ node502 for storage in thecopy #2 of thestreamlet4102. The message is then forwarded to the next adjacent copy in the chain for storage, until the message is stored in the tail copy.
• If theQ node506 fails, theconfiguration manager214 can repair the chain of copies of thestreamlet4102 such that thecopy #3 on theQ node504 becomes the tail copy of the chain. A message is first stored in thecopy #1 on theQ node208, then subsequently stored in thecopy #2 on theQ node502, and thecopy #3 on theQ node504. TheQ node504 then can send an acknowledgement message to theconfiguration manager214 indicating that the message has been stored successfully.
• InFIG. 5D, theconfiguration manager214 replaces the failednode Q node504 by allocating anew Q node508 to store acopy #5 of the chain of copies of thestreamlet4102. In this example, theconfiguration manager214 instructs theMX node204 to send a message (from the publishers406) to theQ node208 for storage in thecopy #1 of the streamlet4102 (582). The message is then forwarded to theQ node502 for storage in thecopy #2 of the streamlet4102 (584). The message is then forwarded to theQ node508 for storage in thecopy #5 of the streamlet4012 (586). The message is then forwarded to theQ node506 for storage in thecopy #4 of the streamlet4102 (588). TheQ node506 can send an acknowledgement message to theconfiguration manager214 indicating that the message has been stored successfully.
• FIG. 6 is a data flow diagram600 illustrating the application of selective filtering, searching, transforming, querying, aggregating and transforming of messages in real time to manage the delivery of messages into and through each channel and on to individual subscribers. Users operating applications on client devices, such as, for example, smartphones, tablets, and other internet-connected devices, act as subscribers (e.g.,subscriber480 inFIG. 4B,subscriber602 inFIG. 6). The applications may be, for example, consumers of the messages to provide real-time information about news, transportation, sports, weather, or other subjects that rely on published messages attributed to one or more subjects and/or channels.Message publishers604 can be any internet-connected service that provides, for example, status data, transactional data or other information that is made available to thesubscribers602 on a subscription basis. In some versions, the relationship between publishers and channels is 1:1, such that is there is one and only one publisher that provides messages into that particular channel. In other instances, the relationship may be many-to-one (more than one publisher provides messages into a channel), one-to-many (a publisher's messages are sent to more than one channel), or many-to-many (more than one publisher provides messages to more than one channel). Typically, when a subscriber subscribes to a channel, they receive all messages and all message data published to the channel as soon as it is published. The result, however, is that many subscribers can receive more data (or data that requires further processing) than is useful. The additional filtering or application of functions against the data places undue processing requirements on the subscriber application and can delay presentation of the data in its preferred format.
• Afilter606 can be created by providing suitable query instructions at, for example, the time thesubscriber602 subscribes to thechannel608. Thefilter606 that is specified can be applied to all messages published to the channel608 (e.g., one message at a time), and can be evaluated before thesubscriber602 receives the messages (e.g., block2 inFIG. 6). By allowingsubscribers602 to create query instructions a priori, that is upon subscribing to thechannel608 and before data is received into thechannel608, the burden of filtering and processing messages moves closer to the data source, and can be managed at the channel level. As a result, the messages are pre-filtered and/or pre-processed before they are forwarded to thesubscriber602. Again, the query instructions need not be based on any a priori knowledge of the form or substance of the incoming messages. The query instructions can be used to pre-process data for applications such as, for example, real-time monitoring services (for transportation, healthcare, news, sports, weather, etc.) and dashboards (e.g., industrial monitoring applications, financial markets, etc.) to filter data, summarize data and/or detect anomalies. One ormore filters606 can be applied to eachchannel608.
• The query instructions can implement real-time searches and queries, aggregate or summarize data, or transform data for use by a subscriber application. In some embodiments, including those implementing JSON formatted messages, the messages can be generated, parsed and interpreted using the query instructions, and the lack of a pre-defined schema (unlike conventional RDBMS/SQL-based applications) means that the query instructions can adapt to changing business needs without the need for schema or application layer changes. This allows the query instructions to be applied selectively at the message level within a channel, thus filtering and/or aggregating messages within the channel. In some instances, the queries may be applied at the publisher level—meaning channels that receive messages from more than one publisher may apply certain filters against messages from specific publishers. The query instructions may be applied on a going-forward basis, that is on only newly arriving messages, and/or in some cases, the query instructions may be applied to historical messages already residing in the channel queue.
• The query instructions can be applied at either or both of the ingress and egress side of the PubSub service. On the egress side, the query instructions act as a per-connection filter against the message channels, and allow each subscriber to manage their own set of unique filters. On the ingress side, the query instructions operate as a centralized, system-wide filter that is applied to all published messages.
• For purposes of illustration and not limitation, examples of query instructions that may be applied during message ingress include:
• • A message may be distributed to multiple channels or to a different channel (e.g., based on geo-location in the message, or based on a hash function of some value in the message).
  • A message may be dropped due to spam filtering or DoS rules (e.g., limiting the number of messages a publisher can send in a given time period).
  • An alert message may be sent to an admin channel on some event arriving at any channel (e.g., cpu_temp>threshold).
• For purposes of illustration and not limitation, examples of query instructions that may be applied during message egress include:
• • Channels that contain events from various sensors where the user is only interested in a subset of the data sources.
  • Simple aggregations, where a system reports real time events, such as cpu usage, sensor temperatures, etc., and we would like to receive some form of aggregation over a short time period, irrespective of the number of devices reporting or the reporting frequency, e.g., average (cpu_load), max(temperature), count(number_of_users), count(number_of_messages) group by country.
  • Transforms, where a system reports real time events and metadata is added to them from mostly static external tables, e.g., adding a city name based on IP address, converting an advertisement ID to a marketing campaign ID or to a marketing partner ID.
  • Adding default values to event streams where such values do not exist on certain devices.
  • Advanced aggregations, where a system reports real time events, and combines some mostly static external table data into the aggregation in real time, e.g., grouping advertisement clicks by partners and counting number of events.
  • Counting number of user events, grouping by a/b test cell allocation.
• In some embodiments, the query instructions may be used to define an index or other suitable temporary data structure, which may then be applied against the messages as they are received into the channel to allow for the reuse of the data element(s) as searchable elements. In such cases, a query frequency may be maintained to describe the number of times (general, or in a given period) that a particular data element is referred to or how that element is used. If the frequency that the data element is used in a query exceeds some threshold, the index may be stored for subsequent use on incoming messages, whereas in other instances in which the index is used only once (or infrequently) it may be discarded. In some instances, the query instruction may be applied to messages having arrived at the channel prior to the creation of the index. Thus, the messages are not indexed according to the data elements described in the query instructions but processed using the query instructions regardless, whereas messages arriving after the creation of the index may be filtered and processed using the index. For queries or other subscriptions that span the time at which the index may have been created, the results of applying the query instructions to the messages as they are received and processed with the index may be combined with results of applying the query instructions to non-indexed messages received prior to receipt of the query instructions.
• For purposes of illustration and not limitation, one use case for such a filtering application is a mapping application that subscribes to public transportation data feeds, such as the locations of all buses across a city. The published messages may include, for example, geographic data describing the location, status, bus agency, ID number, route number, and route name of the buses. Absent pre-defined query instructions, the client application would receive individual messages for all buses. However, query instructions may be provided that filter out, for example, inactive routes and buses and aggregate, for example, a count of buses by agency. The subscriber application receives the filtered bus data in real time and can create reports, charts and other user-defined presentations of the data. When new data is published to the channel, the reports can be updated in real time based on a period parameter (described in more detail below).
• The query instructions can be provided (e.g., at the time the subscriber subscribes to the channel) in any suitable format or syntax. For example, the following illustrates the structure of several fields of a sample subscription request Protocol Data Unit (PDU) with the PDU keys specific to adding a filter to a subscription request:

• 	{
  	″action″: ″subscribe″,
  	“body”: {
  	″channel″: “ChannelName”
  	″filter″: “QueryInstructions”
  	″period″: [1-60, OPTIONAL]
  	}
  	}

  
  In the above subscription request PDU, the “channel” field can be a value (e.g., string or other appropriate value or designation) for the name of the channel to which the subscriber wants to subscribe. The “filter” field can provide the query instructions or other suitable filter commands, statements, or syntax that define the type of key/values in the channel message to return to the subscriber. The “period” parameter specifies the time period in, for example, seconds, to retain messages before returning them to the subscriber (e.g., an integer value from 1 to 60, with a default of, for example, 1). The “period” parameter will be discussed in more detail below. It is noted that a subscription request PDU can include any other suitable fields, parameters, or values.
• One example of a query instruction is a “select” filter, which selects the most recent (or “top”) value for all (e.g., “select.*”) or selected (e.g., “select.name”) data elements. In the example below, the Filter column shows the filter value sent in the query instructions as part of a subscription as the filter field. The Message Data column lists the input of the channel message data and the message data sent to the client as output. In this example, the value for the “extra” key does not appear in the output, as the “select” filter can return only the first level of results and does not return any nested key values.

• Filter	Message Data
  SELECT *	Input
  	{“name”: “art”, “eye”: “blue”},
  	{“name”: “art”, “age”: 11},
  	{“age”: 12, “height”: 190}
  	Output
  	{“name”: “art”, “age”: 12, “eye”: “blue”, “height”: 190}
  SELECT top.*	Input
  	{“top”: {“age”: 12, “eyes”: “blue”}},
  	{“top”: {“name”: “joy”, “height”: 168}, “extra”: 1},
  	{“top”: {“name”: “art”}}
  	Output
  	{“name”: “art”, “age”: 12, “eye”: “blue”, “height”: 168}

• For aggregative functions, all messages can be combined that satisfy the query instructions included in the GROUP BY clause. The aggregated values can then be published as a single message to the subscriber(s) at the end of the aggregation period. The number of messages that are aggregated depends on, for example, the number of messages received in the channel in the period value for the filter. For instance, if the period parameter is set to 1, and 100 messages are received in one second, all 100 messages are aggregated into a single message for transmission to the subscsriber(s). As an example, a query instruction as shown below includes a filter to aggregate position data for an object, grouping it by obj_id, with a period of 1:
• • SELECT*WHERE (<expression with aggregate function>) GROUP BY obj_id
    In this example, all messages published in the previous second with the same obj_id are grouped and sent as a batch to the subscriber(s).
• In some embodiments, a MERGE(*) function can be used to change how aggregated message data is merged. The MERGE(*) function can return a recursive union of incoming messages over a period of time. The merge function may be used, for example, to track location data for an object, and the subscriber is interested in the most recent values for all key/value pairs contained in a set of aggregated messages. The following statement shows an exemplary syntax for the MERGE(*) function:
• • SELECT [expr] [name,]MERGE(*)[.*] [AS name] [FROM expr] [WHERE expr] [HAVING expr] GROUP BY name
• The following examples illustrate how the MERGE(*) function may be applied within query instructions to various types of channel messages. In the following examples, the Filter column shows the filter value included in the query instructions as part of a subscription request as the FILTER field. The Message Data column lists the Input channel message data and the resulting message data sent to the subscriber as Output. The filter returns the most recent values of the keys identified in the input messages, with the string MERGE identified as the column name in the output message data. The first example below shows the MERGE(*) function in a filter with a wildcard, for the message data is returned using the keys from the input as column names in the output.

• Filter	Message Data
  SELECT	Input
  MERGE(*)	{“name”: “art”, “age”: 10},
  	{“name”: “art”, “age”: 11, “items”: [0]}
  	Output
  	{“MERGE”: {“name”: “art”, “age”: 11, “items”: [0]}}

  
  The next example illustrates the use of the MERGE(*) function in a filter using a wildcard and the “AS” statement with a value of MERGE. The output data includes MERGE as the column name.

• Filter	Message Data
  SELECT	Input
  MERGE(*).*	{
  	“name”: “art”,
  	“age”: 12,
  	“items”: [0],
  	“skills”: {
  	“work”: [“robots”]
  	}
  	},
  	{
  	“name”: “art”,
  	“age”: 13,
  	“items”: [“car”],
  	“skills”: {
  	“home”: [“cooking”]
  	}
  	}
  	Output
  	{
  	“name”: “art”,
  	“age”: 13,
  	“items”: [“car”],
  	“skills”: {
  	“work”: [“robots”],
  	“home”: [“cooking”]
  	}
  	}
  SELECT	Input
  MERGE(top.*)	{“top”: { }, “garbage”: 0},
  AS merge	{“top”: {“name”: “art”, “eyes”: “blue”}},
  	{“top”: {“name”: “joy”, “height”: 170}}
  	Output
  	{“merge”: {“name”: “joy”, “eyes”: “blue”, “height”:
  	170}}

• Generally, for aggregative functions and for filters that only include a SELECT(expr) statement, only the latest value for any JSON key in the message data from the last message received can be stored and returned. Therefore, if the most recent message received that satisfies the filter statement is missing a key value identified in a previously processed message, that value is not included in the aggregate, which could result in data loss. However, filters that also include the MERGE(*) function can retain the most recent value for all keys that appear in messages to an unlimited JSON object depth. Accordingly, the most recent version of all key values can be retained in the aggregate.
• The MERGE(*) function can be used to ensure that associated values for all keys that appear in any message during the aggregation period also appear in the final aggregated message. For example, a channel may track the physical location of an object in three dimensions: x, y, and z. During an aggregation period of one second, two messages are published to the channel, one having only two parameters: OBJ{x:1, y:2, z:3} and OBJ{x:2, y:3}. In the second message, the z value did not change and was not included in the second message. Without the MERGE(*) function, the output result would be OBJ{x:2, y:3}. Because the z value was not present in the last message in the aggregation period, the z value was not included in the final aggregate. However, with the MERGE(*) function, the result is OBJ{x:2, y:3, z:3}.
• The following table shows one set of rules that may be used to aggregate data in messages, depending on the type of data. For arrays, elements need not be merged, but instead JSON values can be overwritten for the array in the aggregate with the last array value received.

• Type of JSON	Data to Aggregate	Without	With
  Data	{msg1}, {msg2}	MERGE(*)	MERGE(*)
  Additional	{a: 1, b: 2}, {c: 3}	{c: 3}	{a: 1, b: 2,
  key/value			c: 3}
  Different value	{a: 2}, {a: “2”}	{a: “2”}	{a: “2”}
  datatype
  Missing key/value	{a: 2}, { }	{a: 2}	{a: 2}
  null value	{a: 2}, {a: null}	{a: null}	{a: null}
  Different key	{a: {b: 1}}, {a: {c: 2}}	{a: {c: 2}}	{a: {b: 1,
  value			c: 2}}
  Arrays	{a: [1, 2]}, {a: [3, 4]}	{a: [3, 4]}	{a: [3, 4]}

• The query instructions can be comprised of one or more suitable filter commands, statements, functions, or syntax. For purposes of illustration and not limitation, in addition to the SELECT and MERGE functions, the query instructions can include filter statements or functions, such as, for example, ABS(expr), AVG(expr), COALESCE(a[, b . . . ]), CONCAT(a[, b . . . ]), COUNT(expr), COUNT_DISTINCT(expr), IFNULL(expr1, expr2), JSON(expr), MIN(expr[, expr1, . . . ]), MAX(expr[, expr1, . . . ]), SUBSTR(expr, expr1 [, expr2]), SUM(expr), MD5(expr), SHA1(expr), FIRST_VALUE(expr) OVER (ORDER BY expr1), and/or LAST_VALUE(expr) OVER (ORDER BY expr1), where “expr” can be any suitable expression that is capable of being processed by a filter statement or function, such as, for example, a SQL or SQL-like expression. Other suitable filter commands, statements, functions, or syntax are possible for the query instructions.
• According to the present invention, non-filtered queries can translate to an immediate copy of the message to the subscriber, without any JSON or other like processing. Queries that include a SELECT filter command (without aggregation) can translate into an immediate filter. In instances in which the messages are formatted using JSON, each message may be individually parsed and any WHERE clause may be executed directly on the individual message as it arrives, without the need for creating indices or other temporary data structures. If the messages pass the WHERE clause filter, the SELECT clause results in a filtered message that can be converted back to its original format or structure (e.g., JSON) and sent to the subscriber.
• Aggregative functions, such as, for example, COUNT( ), SUM( ), AVG( ), and the like, can translate into an immediate aggregator. In instances in which the messages are formatted using JSON, each message may be individually parsed and any WHERE clause may be executed directly on the individual message as it arrives, without the need for creating indices or other temporary data structures. If a WHERE clause is evaluated, messages passing such criteria are aggregated (e.g., aggregates in the SELECT clause are executed, thereby accumulating COUNT, SUM, AVG, and so forth) using the previous accumulated value and the value from the individual message. Once per aggregation period (e.g., every 1 second), the aggregates are computed (e.g., AVG=SUM/COUNT), and the SELECT clause outputs the aggregated message, which can be converted to its original format or structure (e.g., JSON) and sent to the subscriber.
• More complex aggregative functions, such as, for example, GROUP BY, JOIN, HAVING, and the like, can be translated into a hash table aggregator. Unlike SELECT or other like functions that can use a constant memory, linearly expanding memory requirements can be dependent upon the results of the GROUP BY clause. At most, grouping by a unique value (e.g., SSN, etc.) can result in a group for each individual message, but in most cases grouping by a common data element (e.g., user_id or other repeating value) can result in far fewer groups. In practice, each message is parsed (from its JSON format, for example). The WHERE clause can be executed directly on the individual message as it arrives, without creating indices or other temporary structures. If the WHERE clause is satisfied, the GROUP BY expressions can be computed directly and used to build a hash key for the group. The aggregative functions in the SELECT clause can be executed, accumulating COUNT, SUM, AVG, or other functions using the previous accumulated value specific for the hash key (group) and the value from the individual message. Once per aggregation period (e.g., every 1 second), the aggregates are computed (e.g., AVG=SUM/COUNT) for each hash key (group), and the SELECT clause can output the aggregated message for each hash key to be converted back to its original format or structure (e.g., JSON) and sent to the subscriber (e.g., one message per hash key (group)).
• In embodiments in which the aggregation period is limited (e.g., 1 second-60 seconds) and the network card or other hardware/throughput speeds may be limited (e.g., 10/gbps), the overall maximal memory consumption can be calculated as time*speed (e.g., 1 GB per second, or 60 GB per minute). Hence, the upper bound is independent of the number of subscribers. In certain implementations, each message only need be parsed once (e.g., if multiple filters are set by multiple clients) and only if needed based on the query instructions, as an empty filter does not require parsing the message.
• Referring toFIG. 7A, subscriptions can include a “period” parameter, generally defined in, for example, seconds and in some embodiments can range from 1 to 60 seconds, although other time increments and time ranges are possible. The period parameter(s) can be purely sequential (e.g., ordinal) and/or time-based (e.g., temporal) and included in the self-described data and therefore available for querying, aggregation, and the like. For example,FIG. 7A illustrates the filter process according to the present invention for the first three seconds with a period of 1 second. In the present example, the subscription starts at t=0. The filter created from the query instructions is applied against all messages received during each 1-second period (e.g., one message at a time). The results for each period are then batched and forwarded to the subscriber. Depending on the query instructions used, the messages can be aggregated using the aggregation functions discussed previously before the message data is sent to the subscriber.
• In some cases, the process defaults to sending only new, incoming messages that meet the query instructions on to the subscriber. However, a subscriber can subscribe with history and use a filter, such that the first message or messages sent to the subscriber can be the historical messages with the filter applied. Using the period of max_age and/or a “next” parameter provides additional functionality that allows for retrieval and filtering of historical messages.
• More particularly, a max_age parameter included with the query instructions can facilitate the retrieval of historical messages that meet this parameter.FIG. 7B illustrates an example of a max_age parameter of 2 seconds (with a period of 1 second) that is provided with the query instructions. The filter created from the query instructions is applied to the historical messages from the channel that arrived from t−2 through t=0 (t=0 being the time the subscription starts), and to the messages that arrived in the first period (from t=0 to t+1). These messages can be sent in a single batch to the subscriber (as Group 1). The filter is applied to each message in each subsequent period (e.g., from t+1 to t+2 as Group 2) to batch all messages that meet the query instructions within that period. Each batch is then forwarded on to the subscriber.
• When a subscriber subscribes with a “next” parameter to a channel with a filter, the filter can be applied to all messages from the next value up to the current message stream position for the channel, and the results can be sent to the subscriber in, for example, a single batch. For example, as illustrated inFIG. 7C, a next parameter is included with the query instructions (with a period of 1 second). The next parameter instructs the process to apply the filter created from the query instructions to each message from the “next position” up through the current stream position (e.g., up to t=0) and to the messages that arrived in the first period (from t=0 to t+1). These messages can be sent in a single batch to the subscriber (as Group 1). The filter is applied to each message in each subsequent period (e.g., from t+1 to t+2 as Group 2) to batch all messages that meet the query instructions within that period. Each batch is then forwarded on the subscriber.
• When a subscriber subscribes with a next parameter, chooses to receive historical messages on a channel, and includes a filter in the subscription, the subscriber can be updated to the current message stream position in multiple batches.FIG. 7D illustrates an example of a max_age parameter of 2 seconds (with a period of 1 second) and a next parameter that can be combined into one set of query instructions. The filter created from the query instructions is applied to the historical messages from the channel that arrived from the end of the history to the “next” value of the subscription (i.e., from 2 seconds before the next value up to the next value), to the messages from the next value to the current stream position (e.g., up to t=0), and to the messages that arrived in the first period (from t=0 to t+1). These messages can be sent in a single batch to the subscriber (as Group 1). The filter is applied to each message in each subsequent period (e.g., from t+1 to t+2 as Group 2) to batch all messages that meet the query instructions within that period. Each batch is then forwarded on the subscriber. Consequently, historical messages can be combined with messages that start at a particular period indicator and batched for transmission to the subscriber.
• The query instructions can define how one or more filters can be applied to the incoming messages in any suitable manner. For example, the resulting filter(s) can be applied to any or all messages arriving in each period, to any or all messages arriving across multiple periods, to any or all messages arriving in select periods, or to any or all messages arriving on a continuous or substantially continuous basis (i.e., without the use of a period parameter such that messages are not retained before returning them to the subscriber). Such filtered messages can be batched in any suitable manner or sent individually (e.g., one message at a time) to subscribers. In particular, the filtered messages can be sent to the subscriber in any suitable format or syntax. For example, the following illustrates the structure of several fields of a sample channel PDU that contains the message results from a filter request:

• {
  ″action″: ″channel/data″,
  “body”: {
  ″channel″: ChannelName
  ″next″: ChannelStreamPosition
  ″messages″: [ChannelData]+   // Can be one or more
  messages
  }
  }

  
  In the above channel PDU, the “channel” field can be a value (e.g., string or other appropriate value or designation) of the channel name to which the subscriber has subscribed. The “next” field can provide the channel stream position of the batch of messages returned in the channel PDU. The “messages” field provides the channel data of the messages resulting from application of the specified filter. One or more messages can be returned in the “messages” field in such a channel PDU. It is noted that a channel PDU can include any other suitable fields, parameters, values, or data.
• FIG. 8 is a diagram of anexample messaging system800 for self-learning spatial recognition in a PubSub communication system. Thesystem800 may include adata analyzer830 that analyzes data frames811 that are encoded in amessage810. Themessage810, including the data frames811, may be provided through a PubSub system as described herein. In some implementations, the data frames811 ofmessage810 may include live video frames812 and/or adata813 of the client device sending themessage810. In one embodiment,message810 may be sent from a device including a camera, directed to a geographical region, as described herein. Additional details regarding live video frames812 anddata813 are provided with respect toFIGS. 9-11.
• In one embodiment, themessage810 may pass through anetwork820, such as through a PubSub system, before arriving at adata analyzer830. The data analyzer830 may be a subscriber to a channel on which themessage810 is provided or otherwise published. The data analyzer830 may perform one or more analysis processes to analyze thedata frame811. For example, thedata analyzer830 may apply one or more machine learning models to the live video frames812 to identify one or more objects (e.g., elements) within thelive video frame812.Data analyzer830 may apply one or more machine learning models to the data frames811 to determine their content (e.g., whether alive video frame812 exists, whetheradditional data813 exists, etc.). The data analyzer830 may be one of a plurality of data analyzers830 that each may analyze the data frames811 and/or live video frames812 in similar or different manners. The data analyzer830 may include one ormore analysis components831 that perform different analyses. In some implementations, the analysis component (e.g.,831) and/or the data analyzer830 may provide additional metadata that indicates whether a particular analysis has been performed so that thedata analyzer830 does not repeat potentially time consuming or processor consuming analysis. Based on the analysis performed by thedata analyzer830, asecond message840 may be generated to include new or additional data, such as a processeddata844. In one embodiment, processeddata844 corresponds to any elements of the data frames811 identified bydata analyzer830. In one embodiment, the processeddata844 may include regions of interests, the status of regions of interest, maps, locations of objects, types of objects, etc. as described herein. The data analyzer830 may provide thesecond message840 overnetwork820. Asdifferent data analyzers830 oranalysis components831 perform additional analyses, the data stored in thesecond message840 may be increase or decrease. Accordingly, thedata analyzer830 may forward thesecond message840 on to subscribers of the channels at some or any point after analysis has been completed (e.g., in real time, substantially real time, or sometime later).
• FIG. 9 is a first flowchart of anexample method900 for self-learning spatial recognition in a PubSub communication system. Themethod900 can be implemented using, for example, an MX node (e.g.,MX node204, MX node461) and a Q node (e.g.,Q node212, Q node208) of themessaging system100, for example. The method begins inblock902 by a computer processing device receiving a plurality of video frames from a video source. In one embodiment, each video frame of the plurality of video frames includes a view of a geolocation. For example, the plurality of video frames may correspond to a parking lot, a parking structure, a street, an arena, a stadium, a theater, a park, a harbor, a field, a shopping center etc. The geographical location may include any suitable location for self-learning spatial recognition, as described herein. In one embodiment, the plurality of video frames from the video source is a plurality of live video frames from a live video source. Live video frames from a live video source may be processed in real time (e.g., as they are captured) by the PubSub system described herein. In another embodiment, the live video frames may be processed at a suitable, later time.
• Atblock904, processing logic detecting a first object entering a first side of a first video frame of the plurality of video frames. In one embodiment, video frames include four sides, corresponding to top, bottom, left, and right orientations. In other embodiments, video frames may include less or more than four sides. In one embodiment, the first object may be classified as an object type. For example, object types may include vehicles, pedestrians, etc. In another embodiment, the object is not classified as a type, but is recognized as a moving object.
• Atblock906, processing logic determines, from the plurality of video frames, that the first object has stopped in an area of the geolocation for at least a threshold amount of time. In one embodiment, the threshold of time is one hour. In another embodiment, the threshold of time is 30 minutes. In various other embodiments, any suitable threshold of time may be used. Advantageously, by determining that the first object has remained in the area of the geolocation for at least a threshold of time, regions of interest may be more accurately determined (e.g., areas that are occupied for less than the threshold of time are less likely to be regions of interest, as defined herein).
• Atblock908, processing logic detects the first object leaving a second side of a second video frame of the plurality of video frames. In one embodiment, the second side of the second video frame is the same side from which the first object entered the first video frame. In another embodiment, the second side of the second video frame is a different side than that from which the first object entered the first video frame. Optionally, processing logic at block912 may determine that a size of the first object changes less than a defined threshold between the first video frame and the second video frame. In one embodiment, the defined threshold may be a ten percent change in size. In various other embodiments, any other suitable defined threshold may be used. For a large field of view (e.g., from wide-angle cameras), there may be a difference between object size in the center of a view and on its edge. Advantageously, by determining that the size of the first object changes less than a defined threshold between the first video frame and the second video frame (e.g., due to perspective), processing logic is able to more accurately determine that the first object is of interest, and therefore at the area occupied by the first object is a region of interest. For example, a bird flying through the view of the video frames may change in size by more than the defined threshold, and may therefore be ruled out as occupying a region of interest.
• At block910, processing logic identifies, by a computer processing device, the area of the geolocation as a region of interest based on the detecting the first object leaving the second side of the video frame and further based on the determining that the first object was stopped in the area of the geolocation for at least the threshold amount of time. In another embodiment, if the first object was stopped in the area of the geolocation for less than the threshold amount of time before leaving, the area is not identified as a region of interest. In one embodiment, a region of interest may correspond to a parking sport, a chair, a bench, a boating slip, or a position in a queue, etc., as described herein. Worth noting is that processing logic may also perform background removal according to any suitable method to assist in identifying moving objects.
• FIG. 10 is a second flowchart of anexample method1000 for self-learning spatial recognition in a PubSub communication system. Themethod1000 can be implemented using, for example, an MX node (e.g.,MX node204, MX node461) and a Q node (e.g.,Q node212, Q node208) of themessaging system100, for example. The method begins inblock1002 by processing logic generating, based on a region of interest, a map comprising the geolocation. In one embodiment, the map may include regions of interest identified in the geolocation. The map may include alphanumeric labels for the regions of interest identified in the map. The map may further include status identifiers, to indicate whether regions of interest are occupied or unoccupied. Alternatively or in addition to generating a map, processing logic may generate a list of regions of interest in the geographical location, with corresponding labels and status identifiers.
• Atblock1004, processing logic determines whether a second object occupies the region of interest. If a second object does occupy the region of interest, processing logic atblock1006 may mark the region the region of interest on the map (or list) as occupied. If a second object does not occupy the region of interest, processing logic may mark the region of interest as unoccupied (or leave the status identifier corresponding to the region of interest as unoccupied, if that is already the current designation.).
• FIG. 11 is a third flowchart of anexample method1100 for self-learning spatial recognition in a PubSub communication system. Themethod1100 can be implemented using, for example, an MX node (e.g.,MX node204, MX node461) and a Q node (e.g.,Q node212, Q node208) of themessaging system100, for example. The method begins inblock1102 by processing logic generating, based on a region of interest, a map comprising the geolocation. Alternatively or in addition to the map, a list may be generated, as discussed above.
• Atblock1104, processing logic determines whether the region of interest has been unoccupied for a defined threshold of time. In one embodiment, the defined threshold of time may be 3 days. In another embodiment, the defined threshold of time may be one week. In various other embodiments, any other suitable defined threshold of time may be used. If the region of interest has been unoccupied for the defined threshold of time, atblock1106 processing logic removes the region of interest from the map. Advantageously, the features ofmethod1100 allow for a region of interest (e.g., a parking spot) to be declassified as a region of interest, if it is not occupied for a certain period of time (e.g., the defined threshold).
• FIG. 12 is a first diagram of an example mapping for self-learning spatial recognition in a PubSub communication system. As described herein, an object may be classified as “of interest” (e.g., car, van, truck, motorcycle, bicycle, scooter, or other suitable vehicle or pedestrian) if it entered the camera view from a side (e.g., from the left, right, top, or bottom of the video frame), stopped for more than a threshold time, and left the camera view by crossing at one of its sides. The place where the object stopped may then be classified as a parking spot (or other type of region of interest). For example, an object (vehicle1202) may entergeographical region1204 from the right side, (e.g.,1206), park in an area (e.g., area1208) for more than a threshold amount of time, and leave from the right side (e.g.,1206). Thereafterarea1208 may be identified as a parking spot (e.g., #1). As shown, parking spots 2-6 may have been already identified as parking spots, according to the operations described herein.
• It is noted that the threshold of time may be chosen to differentiate between a vehicle (which may remain motionless for long periods of time) and, for instance, a person (which cannot). In other words, the period of time may be chosen to allow differentiation between “interesting” objects (e.g., vehicles) and “non-interesting” objects (e.g., non-vehicular objects) in the context of the present embodiments. Additionally or alternatively, the period of time can be chosen to allow differentiation between a “parked” vehicle and one that is merely momentarily stopped or standing (e.g., a car stopped in the middle of the road waiting to make a turn or standing by the side of a road waiting to pick up a passenger). Whatever period of time is chosen, the present invention may take at least up to that period of time before initially classifying the area as a parking spot.
• Furthermore, as described above with respect toFIG. 9, during object movement, an object can change in size. If a change in size is more than a defined threshold (e.g., ten percent), it may be considered as noise or caused by error in object unification (e.g., when two objects are close enough that they could be unified by mistake). However, when the size of the object changes less than the defined threshold, the change in size may be considered as object sizing due to perspective. Alternatively, instead of using a percentage, a suitable heuristic can be used that can assume that a single object should remain single.
• After a region of interest (e.g., a parking spot) is marked or otherwise identified according to the embodiments described herein, it could be occupied by other objects for a period less than the threshold duration. In other words, once initially identified, the region of interest can be marked as open or occupied after any suitable length of time as cars or other vehicles enter, park, and leave the parking lot.
• According to the one embodiment, the boundaries of a region of interest can be detected by, for example, extracting the difference between an occupied region of interest and an image of the same unoccupied region of interest. Such an exemplary embodiment can act as a type of mask or filter that results in detecting an area that may be of arbitrary shape. For example, it is noted that shadows and other image artifacts could be included as part of the vehicle, which can result in an identified parking spot that is not necessarily rectangular in shape.
• If two regions of interest have more than, for example, 50% overlap, they may be considered to be the same region of interest. In other words, regions of interest should be checked pairwise, such that if region of interest A has merged into region of interest B, then region of interest A is no longer considered separate from region of interest B. Other values for determining overlap can be used. For instance, a 50% overlap has been provided as an illustrative example, because an object cannot overlap with two non-overlapping objects more than 50% at the same time. However, alternative mechanisms can be used to determine overlap, such as with a suitable heuristic or the like.
• If a region of interest has been unoccupied for a period of time that exceeds some defined threshold (e.g., a define threshold could be about a week, although other suitable lengths of time can also be used), it may be removed because the region of interest could have been identified due to camera movement, someone using the region of interest in an improper way (e.g., parking in an emergency-vehicle-only zone), etc.
• In one embodiment, various data may be generated by the embodiments disclosed herein. Data may include, for example: a length of time the region of interest is occupied or unoccupied; an image of the object that occupied the region of interest most recently; a map of all regions of interest in the geographical location, including labels and status indicators (e.g., color coded); a number of current occupied and unoccupied regions of interest; etc.
• As previously disclosed, the embodiments described herein are not limited to use for parking spots in a parking lot. Rather, the present embodiments can be used to identify open and occupied spaces in any large area that includes a reasonably regular and repeating pattern of objects. For purposes of illustration and not limitation, the present embodiments can be used to identify parking spaces lining a road, berths (slips) in a harbor, seats in a stadium, etc.
• FIG. 13 is a block diagram of anexample computing device1300 that may perform one or more of the operations described herein. Thecomputing device1300 may be connected to other computing devices in a LAN, an intranet, an extranet, and/or the Internet. Thecomputing device1300 may operate in the capacity of a server machine in client-server network environment or in the capacity of a client in a peer-to-peer network environment. Thecomputing device1300 may be provided by a personal computer (PC), a set-top box (STB), a server, a network router, switch or bridge, or any machine capable of executing a set of instructions (sequential or otherwise) that specify actions to be taken by that machine. Further, while only asingle computing device1300 is illustrated, the term “computing device” shall also be taken to include any collection of computing devices that individually or jointly execute a set (or multiple sets) of instructions to perform the methods discussed herein.
• Theexample computing device1300 may include a computer processing device (e.g., a general purpose processor, ASIC, etc.)1302, amain memory1304, a static memory1306 (e.g., flash memory and a data storage device1308), which may communicate with each other via abus1310. Thecomputer processing device1302 may be provided by one or more general-purpose processing devices such as a microprocessor, central processing unit, or the like. In an illustrative example,computer processing device1302 may comprise a complex instruction set computing (CISC) microprocessor, reduced instruction set computing (RISC) microprocessor, very long instruction word (VLIW) microprocessor, or a processor implementing other instruction sets or processors implementing a combination of instruction sets. Thecomputer processing device1302 may also comprise one or more special-purpose processing devices such as an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), a digital signal processor (DSP), network processor, or the like. Thecomputer processing device1302 may be configured to execute the operations described herein, in accordance with one or more aspects of the present disclosure, for performing the operations and steps discussed herein.
• Thecomputing device1300 may further include anetwork interface device1312, which may communicate with anetwork1314. Thedata storage device1308 may include a machine-readable storage medium1316 on which may be stored one or more sets of instructions, e.g., instructions for carrying out the operations described herein, in accordance with one or more aspects of the present disclosure.Instructions implementing module1318 may also reside, completely or at least partially, withinmain memory1304 and/or withincomputer processing device1302 during execution thereof by thecomputing device1300,main memory1304 andcomputer processing device1302 also constituting computer-readable media. The instructions may further be transmitted or received over thenetwork1314 via thenetwork interface device1312.
• While machine-readable storage medium1316 is shown in an illustrative example to be a single medium, the term “computer-readable storage medium” should be taken to include a single medium or multiple media (e.g., a centralized or distributed database and/or associated caches and servers) that store the one or more sets of instructions. The term “computer-readable storage medium” shall also be taken to include any medium that is capable of storing, encoding or carrying a set of instructions for execution by the machine and that cause the machine to perform the methods described herein. The term “computer-readable storage medium” shall accordingly be taken to include, but not be limited to, solid-state memories, optical media and magnetic media.
• Embodiments of the subject matter and the operations described in this specification can be implemented in digital electronic circuitry, or in computer software, firmware, or hardware, including the structures disclosed in this specification and their structural equivalents, or in combinations of one or more of them. Embodiments of the subject matter described in this specification can be implemented as one or more computer programs, i.e., one or more modules of computer program instructions, encoded on computer storage medium for execution by, or to control the operation of, data processing apparatus. Alternatively or in addition, the program instructions can be encoded on an artificially-generated propagated signal, e.g., a machine-generated electrical, optical, or electromagnetic signal that is generated to encode information for transmission to suitable receiver apparatus for execution by a data processing apparatus. A computer storage medium can be, or be included in, a computer-readable storage device, a computer-readable storage substrate, a random or serial access memory array or device, or a combination of one or more of them. Moreover, while a computer storage medium is not a propagated signal, a computer storage medium can be a source or destination of computer program instructions encoded in an artificially-generated propagated signal. The computer storage medium can also be, or be included in, one or more separate physical components or media (e.g., multiple CDs, disks, or other storage devices).
• The operations described in this specification can be implemented as operations performed by a data processing apparatus on data stored on one or more computer-readable storage devices or received from other sources.
• The term “computer processing device” encompasses all kinds of apparatus, devices, and machines for processing data, including by way of example a programmable processor, a computer, a system on a chip, or multiple ones, or combinations, of the foregoing. Although referred to as a computer processing device, use of the term also encompasses embodiments that include one or more computer processing devices. The computer processing device can include special purpose logic circuitry, e.g., an FPGA (field programmable gate array) or an ASIC (application-specific integrated circuit). The computer processing device can also include, in addition to hardware, code that creates an execution environment for the computer program in question, e.g., code that constitutes processor firmware, a protocol stack, a database management system, an operating system, a cross-platform runtime environment, a virtual machine, or a combination of one or more of them. The computer processing device and execution environment can realize various different computing model infrastructures, such as web services, distributed computing and grid computing infrastructures.
• A computer program (also known as a program, software, software application, script, or code) can be written in any form of programming language, including compiled or interpreted languages, declarative, procedural, or functional languages, and it can be deployed in any form, including as a stand-alone program or as a module, component, subroutine, object, or other unit suitable for use in a computing environment. A computer program may, but need not, correspond to a file in a file system. A program can be stored in a portion of a file that holds other programs or data (e.g., one or more scripts stored in a markup language resource), in a single file dedicated to the program in question, or in multiple coordinated files (e.g., files that store one or more modules, sub-programs, or portions of code). A computer program can be deployed to be executed on one computer or on multiple computers that are located at one site or distributed across multiple sites and interconnected by a communication network.
• The processes and logic flows described in this specification can be performed by one or more programmable processors executing one or more computer programs to perform actions by operating on input data and generating output. The processes and logic flows can also be performed by, and apparatus can also be implemented as, special purpose logic circuitry, e.g., an FPGA (field programmable gate array) or an ASIC (application-specific integrated circuit).
• Processing devices suitable for the execution of a computer program include, by way of example, both general and special purpose microprocessors, and any one or more processors of any kind of digital computer. Generally, a processing device will receive instructions and data from a read-only memory or a random access memory or both. The essential elements of a computer are a processor for performing actions in accordance with instructions and one or more memory devices for storing instructions and data. Generally, a computer will also include, or be operatively coupled to receive data from or transfer data to, or both, one or more mass storage devices for storing data, e.g., magnetic disks, magneto-optical disks, optical disks, or solid state drives. However, a computer need not have such devices.
• Moreover, a computer can be embedded in another device, e.g., a smart phone, a mobile audio or video player, a game console, a Global Positioning System (GPS) receiver, or a portable storage device (e.g., a universal serial bus (USB) flash drive), to name just a few. Devices suitable for storing computer program instructions and data include all forms of non-volatile memory, media and memory devices, including, by way of example, semiconductor memory devices, e.g., EPROM, EEPROM, and flash memory devices; magnetic disks, e.g., internal hard disks or removable disks; magneto-optical disks; and CD-ROM and DVD-ROM disks. The processing device and the memory can be supplemented by, or incorporated in, special purpose logic circuitry.
• To provide for interaction with a user, embodiments of the subject matter described in this specification can be implemented on a computer having a display device, e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor, for displaying information to the user and a keyboard and a pointing device, e.g., a mouse, a trackball, a touchpad, or a stylus, by which the user can provide input to the computer. Other kinds of devices can be used to provide for interaction with a user as well; for example, feedback provided to the user can be any form of sensory feedback, e.g., visual feedback, auditory feedback, or tactile feedback; and input from the user can be received in any form, including acoustic, speech, or tactile input. In addition, a computer can interact with a user by sending resources to and receiving resources from a device that is used by the user; for example, by sending web pages to a web browser on a user's client device in response to requests received from the web browser.
• Embodiments of the subject matter described in this specification can be implemented in a computing system that includes a back-end component, e.g., as a data server, or that includes a middleware component, e.g., an application server, or that includes a front-end component, e.g., a client computer having a graphical user interface or a Web browser through which a user can interact with an implementation of the subject matter described in this specification, or any combination of one or more such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication, e.g., a communication network. Examples of communication networks include a local area network (“LAN”) and a wide area network (“WAN”), an inter-network (e.g., the Internet), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks).
• The computing system can include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. In some embodiments, a server transmits data (e.g., an HTML page) to a client device (e.g., for purposes of displaying data to and receiving user input from a user interacting with the client device). Data generated at the client device (e.g., a result of the user interaction) can be received from the client device at the server.
• A system of one or more computers can be configured to perform particular operations or actions by virtue of having software, firmware, hardware, or a combination of them installed on the system that in operation causes or cause the system to perform the actions. One or more computer programs can be configured to perform particular operations or actions by virtue of including instructions that, when executed by data processing apparatus, cause the apparatus to perform the actions.
• While this specification contains many specific implementation details, these should not be construed as limitations on the scope of any inventions or of what may be claimed, but rather as descriptions of features specific to particular embodiments of particular inventions. Certain features that are described in this specification in the context of separate embodiments can also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment can also be implemented in multiple embodiments separately or in any suitable subcombination. Moreover, although features may be described above as acting in certain combinations and even initially claimed as such, one or more features from a claimed combination can in some cases be excised from the combination, and the claimed combination may be directed to a subcombination or variation of a subcombination.
• Similarly, while operations are depicted in the drawings in a particular order, this should not be understood as requiring that such operations be performed in the particular order shown or in sequential order, or that all illustrated operations be performed, to achieve desirable results. In certain circumstances, multitasking and parallel processing may be advantageous. Moreover, the separation of various system components in the embodiments described above should not be understood as requiring such separation in all embodiments, and it should be understood that the described program components and systems can generally be integrated together in a single software product or packaged into multiple software products.
• Thus, particular embodiments of the subject matter have been described. Other embodiments are within the scope of the following claims. In some cases, the actions recited in the claims can be performed in a different order and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In certain implementations, multitasking and parallel processing may be advantageous.

Claims (20)

What is claimed is:

1. A method, comprising:

detecting a first object entering a first video frame of a plurality of video frames of a view of a geolocation;

detecting the first object leaving a second video frame of the plurality of video frames; and

identifying, by a computer processing device, an area of the geolocation as a region of interest based on the detecting the first object leaving.

2. The method ofclaim 1, comprising:

generating, based on the region of interest, a map comprising the geolocation.

3. The method ofclaim 2, comprising:

determining that a second object occupies the region of interest; and

marking the region of interest on the map as occupied.

4. The method ofclaim 2, comprising:

determining that the region of interest has not been occupied for a predetermined amount of time; and

removing the region of interest from the map.

5. The method ofclaim 1, wherein the region of interest comprises a parking spot, and wherein the first object comprises a vehicle and wherein the method further comprises determining, from the plurality of video frames, that the first object has stopped in the area of the geolocation for at least a threshold amount of time.

6. The method ofclaim 1, comprising:

receiving the plurality of video frames from a live video source.

7. The method ofclaim 6, wherein the plurality of video frames are received from the live video source via a publish-subscribe communication system.

8. The method ofclaim 1, further comprising determining, from the plurality of video frames, that the first object has stopped in the area of the geolocation for at least a threshold amount of time, wherein the threshold amount of time is selected to differentiate between different objects in the plurality of video frames.

9. The method ofclaim 1, comprising:

detecting boundaries of the region of interest by extracting a difference between an image of an occupied region of interest and an image of a same unoccupied region of interest.

10. The method ofclaim 1, comprising:

determining that a size of the first object changes less than a predetermined threshold between the first video frame and the second video frame.

11. A system, comprising:

a computer processing device programmed to perform operations to:

detect a first object entering a first video frame of a plurality of video frames of a view of a geolocation;

detect the first object leaving a second video frame of the plurality of video frames; and

identify an area of the geolocation as a region of interest based on the detecting the first object leaving.

12. The system ofclaim 11, the computer processing device further to:

generate, based on the region of interest, a map comprising the geolocation.

13. The system ofclaim 12, the computer processing device further to:

determine that a second object occupies the region of interest; and

mark the region of interest on the map as occupied.

14. The system ofclaim 12, the computer processing device further to:

determine that the region of interest has not been occupied for a predetermined amount of time; and

remove the region of interest from the map.

15. The system ofclaim 11, wherein the computer processing device further to determine, from the plurality of video frames, that the first object has stopped in an area of the geolocation for at least a threshold amount of time, wherein the region of interest comprises a parking spot, and wherein the first object comprises a vehicle.

16. The system ofclaim 11, the computer processing device further to:

receive the plurality of video frames from a live video source.

17. The system ofclaim 16, wherein the plurality of video frames are received from the live video source via a publish-subscribe communication system.

18. The system ofclaim 11, wherein the computer processing device further to determine, from the plurality of video frames, that the first object has stopped in an area of the geolocation for at least a threshold amount of time, and wherein the threshold amount of time is selected to differentiate between different objects in the plurality of video frames.

19. The system ofclaim 11, the computer processing device further to:

detect boundaries of the region of interest by extracting a difference between an image of an occupied region of interest and an image of a same unoccupied region of interest.

20. A non-transitory computer-readable medium having instructions stored thereon that, when executed by a computer processing device, cause the computer processing device to:

detect a first object entering a first video frame of a plurality of video frames of a view of a geolocation;

detect the first object leaving a second video frame of the plurality of video frames; and

identify an area of the geolocation as a region of interest based on the detecting the first object leaving.

Images