US20030016668A1 - Method of device-to-device communications in hybrid distributed device control networks - Google Patents

Abstract

The present invention comprises a method of source routing to implement device-to-device communications across a hybrid distributed device control network. The method is based in packet communications in which packets are structured so that they can be readily converted between communications protocols, and in which packets enclose routing information and parameters.

Description

BACKGROUND OF INVENTIONBACKGROUND RELATED APPLICATIONS
• This invention uses the concepts of true distributed control and distributed device control network of our co-pending applications. It also uses the concepts of device controller and network-enabled devices.[0001]
BACKGROUND FIELD OF INVENTION
• This invention relates to device-to-device network communication methods and systems, specifically to device-to-device communications across a hybrid distributed device control network.[0002]
BACKGROUND DISCUSSION OF PRIOR ART
• The Cambridge Dictionary of American English defines a “device” to be an object or machine invented to fulfill a particular purpose. According to the present invention, the term “device” is not limited to physical apparati, but is considerably expanded to comprise abstract or virtual devices, such as system operators, that partake in network communications. One fundamental characteristic of devices is that they comprise a finite set of states associated with their operation.[0003]
• Source routing is packet routing in which a source node has a priori knowledge of the network path to a destination node, i.e., a complete sequence of intermediary network nodes a packet must traverse to reach its destination. This routing information is included in the header of every sent packet.[0004]
• Source routing is implemented in many network communication environments (e.g., Internet Protocol). Yet, its use is deemed inefficient for regular network package routing due to the overhead of increased packet header size it involves, and it is virtually never used, except for network mapping and troubleshooting instances.[0005]
• Nonetheless, the simplicity of using source routing communications makes it positively suitable and advantageous for some applications, specifically hybrid distributed device control networks, since its use greatly reduces the cost and complexity of hardware logic of network routing devices and allows for faster networks.[0006]
• According to the present invention, a hybrid distributed device control network comprises a set of interconnected subnetworks of arbitrary topology, each containing several interconnected device controllers and/or network-enabled devices. The term “hybrid” refers to a network that comprises several subnetworks interconnected across dissimilar communication media (e.g., Ethernet, RF, etc.), and using different communication protocols (e.g., LONtalk, UDP/IP, etc.). Subnetworks using incompatible protocols are interconnected at network routers which act as dual nodes, translating and transferring packets between subnetworks and communication protocols. There are two types of source routing, strict and loose. In strict source routing, every intermediary network node is explicitly specified in the packet header at the source. In loose source routing, a set of intermediary nodes through which the packet must pass are specified in the packet header, but each of these may reside several hops away from one another.[0007]
• Implementing a communication system over a hybrid distributed device control network using source routing requires simpler processing that results in reduced implementation complexity and cost. Source routing is also compatible with the philosophy of true distributed control because packet routing information is contained at each node where packets originate.[0008]
• In the recent past, there have been several attempts to create a method of source routing to communicate across hybrid networks (e.g., U.S. Pat. 5,570,084). However, the resulting methods are definitely inadequate for device-to-device communications because they do not implement important communications services, namely, multicasting and broadcasting, which are absolutely required for interdevice communications. Other attempts have been made to propose generic methods for routing packets in communications networks (e.g., U.S. Pat. 5,353,283). However, the resulting methods either have been developed for networks with limited topologies or cannot be used in networks comprising subnetworks using incompatible communications protocols and/or dissimilar communications media.[0009]
• Hence, it is an object of the present invention to overcome the disadvantages of the prior art, presenting a source routing communication method specifically designed for interdevice packet routing across networks comprising different communication protocols (e.g., LONtalk, UDP/IP, etc.) and media (e.g., Ethernet, RF, etc.), and providing several communication services.[0010]
SUMMARY OF INVENTION
• The present invention comprises a method of source routing to implement device-to-device communications across a hybrid distributed device control network. The method is based in packet communications in which packets are structured so that they can be readily converted between communications protocols, and in which packets enclose routing information and parameters.[0011]
OBJECTS AND ADVANTAGES
• Accordingly, several objects and advantages of the present invention are:[0012]
• a) To provide a flexible method designed and developed specifically for interdevice communications, which addresses and overcomes the limitations of existing communications methods;[0013]
• b) to provide a method of source routed communications designed to fulfill requirements of device-to-device communications, which reduces the complexity and cost of system implementation of existing communications methods;[0014]
• c) to provide a method of device-to-device communications which can be used in networks independently of the network communication protocols and/or media used, and which indicates how a packet should be transferred and translated between networks using different network communication protocols and/or media;[0015]
• d) to provide a flexible method of device-to-device communications which supports dynamic structuring of communication packets, depending on the communication services required;[0016]
• e) to provide a method of device-to-device communications which supports one-to-many communication services, including multicasting and broadcasting;[0017]
• f) to provide a method of device-to-device communications in which all network nodes and routers, once configured, operate autonomously in agreement with the philosophy of true distributed control; Other objects and advantages of this invention will become apparent from a consideration of the ensuing description and drawings.[0018]
BRIEF DESCRIPTION OF DRAWINGS
• In the drawings, closely related figures have the same number but different alphabetic suffixes.[0019]
• FIG. 1 shows an exemplary hybrid distributed control network.[0020]
• FIG. 2A illustrates a general configuration of device controllers and network-enabled devices.[0021]
• FIG. 2B shows an exemplary instance of device-to-device communications, between an alarmed door and an alarm siren.[0022]
• FIG. 3 displays the structure of a generic network communications packet.[0023]
• FIG. 4 comprises a simplified version of the network shown in FIG. 1. FIG. 5A illustrates the acknowledge communications service.[0024]
• FIG. 5B illustrates the broadcast communications service.[0025]
• FIG. 5B illustrates the multicast communications service.[0026]
• The terms Type I, II and III Networks are arbitrary terms used to differentiate between exemplary subnetworks using dissimilar communication protocols and media. Any network protocol type may be substituted in their stead.[0027]
LIST OF REFERENCE NUMERALS IN DRAWINGS
• [0028]10 Internetwork
• [0029]12 Network Router Node (Internet to Type III Network)
• [0030]14,18 Network Router Node (Internet to Type II Network)
• [0031]16 Network Router Node (Internet to Type I Network)
• [0032]20 Type III Subnetwork (Wireless)
• [0033]22,26,37 Type II Subnetwork
• [0034]24 Type I Subnetwork
• [0035]28,30 Type III Network End Node
• [0036]32 Network Router Node (Type III Network to Type II Network)
• [0037]34,36,38,40,48,50 Type II Network End Nodes
• [0038]42,43,44 Type I Network End Nodes
• [0039]46 Network Router Node (Type I Network to Type II Network)
• [0040]210 Alarmed electric door
• [0041]212 Device controller controlling
• [0042]210214 Device controller controlling
• [0043]216216 Alarm siren
• [0044]218 Network-enabled Alarm siren
DETAILED DESCRIPTION
• Now, the present invention will be described by referring to the accompanying drawings that illustrate preferred embodiments of the invention.[0045]
• FIG. 1 illustrates an exemplary hybrid distributed device control network in which a plurality of[0046]network router nodes12,14,16 and18, hereafter referred to as routers, are connected to aninternet10. Said routers serve as connection links betweeninternet10 and a set ofsubnetworks20,22,24,26, which use dissimilar communication protocols and/or media. In FIG. 1,subnetwork20 is of hypothetical Type III (e.g., using radio frequency as medium), subnetworks22 and26 are of hypothetical Type II (e.g., using LONtalk protocol), andsubnetwork24 is of hypothetical Type I (e.g., using UDP/IP over Ethernet). There are two further components, namely, a subnetwork37 and arouter46. Subnetwork37 connects to subnetwork20 through arouter32. Likewise,router46 connectssubnetworks22 and24 together.
• Each said subnetwork contains a plurality of network end nodes, hereafter referred to as nodes. For instance,[0047]subnetwork20 comprisesnodes28 and30; subnetwork22 comprisesnodes38 and40;subnetwork24 comprisesnodes42,43 and44;
• subnetwork[0048]26 comprisesnodes48 and50; and subnetwork37 comprisesnodes34 and36.
• In said hybrid distributed device control network, hereafter referred to as network, a packet originating at a source node is directly received at a destination node, provided both source and destination nodes reside in the same subnetwork. For instance, a packet originating at[0049]source node44 will directly reachdestination node42, withinsubnetwork24.
• Furthermore, a packet originating at a source node whose destination node resides in a different subnetwork must traverse a network path through one or more routers to reach its destination. For instance, a packet originating at[0050]source node42 must pass throughrouters16 and14 to reachdestination node38.
• In this method, every router interconnects two subnetworks, and thus exists in two subnetworks simultaneously, transferring packets from one subnetwork into another. Since the two subnetworks interconnected by a router use different communication protocols and/or media, it follows that a router must have two different network addresses, one for each of its associated subnetworks. For example,[0051]router46 must have a Type II Network address seen bynodes38 and40 of subnetwork22, and a Type I Network address seen bynodes42,43 and44 ofsubnetwork24.
• Every node in said network is either a device controller to which several devices may be connected, or a network-enabled device, i.e., a conventional device with added capabilities to communicate across a network. FIG. 2A illustrates this.[0052]
• The main functions of a device controller include controlling the state of all conventional devices connected to it, such as modifying the device″s operating state (e.g., command execution, activations, deactivations, etc.), detecting an operating state change or reporting on the operating state change of a device (e.g., event reporting), among others. Some typical devices include electric doors, thermal sensors, GPS, video cameras, among several others. The main functions of a network-enabled device include reporting on its operating state, among others.[0053]
• This invention concerns communication between device controllers and/or network-enabled devices across networks. FIG. 2B illustrates an exemplary case: an[0054]electric door210 controlled by adevice controller212 is opened. Saidcontroller212 detects the state change in saidelectric door210 and sends two messages or packets across the network. The first packet reachesdevice controller214, containing a command to activate analarm siren216. The second packet reaches network-enabledalarm siren218 and activates it.
• In source routing, each packet or message originating at a network node contains packet routing information, including a list of all network routers a packet must traverse in order to reach its destination node. The structure of a packet is described next, and is shown in FIG. 3.[0055]
• According to the present invention, a packet consists of three fundamental sections: header, network path and data. The headers section contains general information about the packet, including a packet type and a packet identification number, among several others. The network path section comprises a list of all network node addresses a packet must traverse to reach its final destination. The data section contains the actual packet content information.[0056]
• The header section contains several information fields: • Packet Type (PT): a number which specifies the type of said packet, and which indicates the type of communication services that said packet requires. Packet types include ACK (i.e., referring to a packet sent to acknowledge receipt of another packet), NO_ACK (i.e., referring to a packet comprising data which requires an ACK packet response), UNACK (i.e., referring to a packet comprising data which requires no ACK packet response), BROADCAST (i.e., referring to a packet directed to all nodes in a given subnetwork), MULTICAST (i.e., referring to a packet directed to some nodes in a given subnetwork), among others. Depending on the value of this field, the structure of the packet may vary, as described below. In one implementation, this field consists of one byte, allowing 255 possible packet type values.[0057]
• Packet ID (PI): a number which serves as a unique identifier for said packet. In one implementation, this field consists of three bytes.[0058]
• Quality of Service (QOS): a set of bits which specify several network parameters that modify how a package is handled by network routers. The most important of these parameters is packet priority. In one implementation, this field consists of one byte.[0059]
• Network Path Length (NPL): a number indicating the number of nodes that said packet must traverse in order to reach its destination, including the source and destination nodes. In one implementation, this field consists of one byte.[0060]
• Data Pointer (DP): a number that points to the beginning of the data section below. Specifically, it contains the offset in bytes from the beginning of the packet to the beginning of the data section.[0061]
• Network Path Pointer Table (NPPT): a set of fields that contain pointers to entries in the Network Path section described below. Specifically, said pointers specify the offset from the beginning of the packet to the start of every entry in the Network Path section below. The order in which network nodes are found on this table determines the order in which network nodes will be traversed by the packet. The table has as many entries as specified in the Network Path Length above. Note that to alter the network path that a message must follow you can modify the order in which network address are pointed at by NPPT entries. For instance, to reverse the order of the nodes that a packet must traverse, reverse the order of the entries in the NPPT. In one implementation, each table entry consists of one byte.[0062]
• Network Path Destination Index (NPDI): a number specifying the next entry in the NPPT to which the packet should be transmitted. For instance, when the packet is at[0063]source node34, this field has a value of two (“2”), referring to the second entry in the NPPT, which points to the address ofrouter32. After the packet reachesrouter32, this field will have a value three (“3”), referring to the third entry in NPPT, which points to the address ofrouter12, and so on. In one implementation, this field consists of one byte.
• Multicast Pointer (MP): (only included if Packet Type equals “40”, in multicast packets.) This field contains a number that points to the last router entry in the Network Path Pointer Table, which connects to the subnetwork that contains all nodes receiving the multicast packet. This number must be equal to or lesser than the Network Path Length specified above. In one implementation, this field consists of one byte.[0064]
• Network Path section consists of a list of at least two network address entries comprising all network routers a packet must visit to reach its destination. Each entry consists of two members: a network type identifier and a network address.[0065]
• Network Type (NT): a number specifying the network type of the router network address that follows. In one implementation, this field consists of one byte.[0066]
• Network Address (NA): a variable-length number specifying a network address to which said packet must be sent. In one implementation, this field consists of four bytes.[0067]
• Finally, the Data section contains two fields: • Data Length (DL): a number specifying the length of the data segment contained in said packet. In one implementation, this field consists of two bytes.[0068]
• Data Segment (DS): a data segment of the above data length. This segment contains the actual data for which this message is sent. In the example of FIG. 2B, the data of the packet traveling from[0069]nodes212 to214 comprises a control command which activates the alarm siren controlled bynode214.
OPERATION OF INVENTION
• Let there be a packet X, structured as described above, which originates at a source node and travels across a network to reach a destination node. Packet X includes a network path section with a complete list of network routers it must traverse to reach its destination. Although packet X comprises the exact network route it must traverse to reach its destination, it does not contain any protocol-specific information. Instead, packet X is a message unit that must be encapsulated in a protocol-specific packet depending on the network where it may circulate. Hence, each node that sends or receives packet X must encapsulate or decapsulate packet X accordingly.[0070]
• FIG. 4 is an exemplary illustration of the following description of the method presented in this invention. It is a simplified version of FIG. 1, following the same numeral pattern. A packet X is created at a[0071]source node34 which must traverse a network to reachdestination node42.Source node34 encapsulates packet X within a packet X0, a protocol-specific packet that can be processed by subnetwork37 (i.e., Type II Network). Next,source node34 sends packet X0 out across subnetwork37 towardsrouter32.
• [0072]Router32 receives packet X0 and decapsulates it to retrieve packet X. Next,router32 updates packet X″s NPDI to point to the next destination address in the NPPT (i.e., router12). Next,router32 encapsulates packet X into a packet X1, following the protocol used by subnetwork20 (i.e., Type III Network), and sends packet X1 out acrosssubnetwork20 towardsrouter12.
• [0073]Router12 receives packet X1 and decapsulates it to retrieve packet X. Next,router12 updates packet X″s NPDI to point to the next destination address in the NPPT (i.e., router16). Next,router12 encapsulates packet X into a packet X2, following the protocol used by aninternet10, and sends packet X2 out acrossinternet10 towardsrouter16.
• [0074]Router16 receives packet X2 and decapsulates it to retrieve packet X. Next,router16 updates packet X″s NPDI to point to the next destination address in the NPPT (i.e., node42). Next,router16 encapsulates packet X into a packet X3, following the protocol used by subnetwork16 (i.e., Type I Network), and sends packet X3 out acrosssubnetwork24 towardsnode42.
• When packet X[0075]3 is received atdestination node42, it is decapsulated and packet X is retrieved from it.Node42 processes the data section of packet X.
• To summarize, a packet originates at a source node and comprises an exact network path that the packet must traverse. Such packet is sent through all nodes and routers comprised in the packet″s network path section. The packet is encapsulated, and the destination address is included in the encapsulation headers. Then, the encapsulated packet is decapsulated. This process is repeated until the packet passes through all required subnetworks of dissimilar communications protocols and media. At each network router, the packet is modified accordingly so that its NPDI points to the next destination that the packet must reach. The rest of the contents of the packet remains untouched. Once it reaches its final destination node, the packet″s data section is processed.[0076]
• The above is a description of the basic communication method, assuming the simplest type of communication service, namely, forward transmission demanding no reply. The method varies according to the type of communication service specified in the packet″s Packet Type field. These variations are described next.[0077]
• If the type of communications service requires a response from the destination node, then a return or backward path to source[0078]node34 must be known bydestination node42. Note that in the basic method above,destination node42 cannot retrieve the return path towardssource node34 from packet Ax″s headers. For instance, packet X only containsrouter16's address as seen byinternet10, and does not containrouter16's address as seen bysubnetwork24.
• The type of communications in which a source node requires a response from a destination node is described next. This is referred to as Response communications service. When a router receives a packet from one of its two interconnecting subnetworks, it modifies the packet″s network address path list (NT and NA fields) substituting the router address as seen by its sending subnetwork for the router address as seen by its receiving subnetwork. For instance, in FIG. 4A, when packet X leaves[0079]node34 and reachesrouter32, it containsrouter32's address as seen by subnetwork37 (i.e., receiving subnetwork).Router32, then, replaces said address byrouter32's address as seen by subnetwork20 (i.e., sending subnetwork), and transmits packet X torouter12. When packet X reachesrouter12, its network address path containsrouter12's address as seen bysubnetwork20.Router12, then, replaces said address by router12 s address as seen byinternet10, and transmits packet X torouter16. This substitution recurs at every router found on packet X″s network path. Once packet X reachesdestination node42, its network path section contains all router addresses as seen fromnode42's perspective, thus, allowing it to reachsource node34, when a response is necessary.
• When said destination node needs to send a response packet out to said source node, it may copy the structure of the received packet X, invert the entries of the NPPT (so that the network path is reversed), set the appropriate header fields (e.g., PT, NPDI, QOS, etc.), and enclose the corresponding data segment. Hence, this show that processing required for generating response packets is minimal.[0080]
• In a second type of communications service, the Acknowledge (ACK) communications service, it is desired that responses be sent back from a destination node to a source node to acknowledge receipt of a packet. In these circumstances, if acknowledgement messages are not received at the source node within a time frame, the source node may resend the packet. An instance of this type of communication services is illustrated in FIG. 5A (for Packet Type equal to “50”).[0081]Source node34 generates packet X of said type and sends it to destination node42 (step1). Upon receiving packet X,router32 sends an ACK packet (Packet Type equals “51”) to source node34 (step2, ACK0). Then, packet X travels tonodes12 and16 (steps3,4 and5). Oncedestination node42 receives packet X, another ACK packet is sent back to source node34 (step6, ACK1). Thus, only upon receipt of both ACK0 and ACK1 packets does sourcenode34 consider packet X transmission successful. Otherwise,source node34 may resend packet X.
• In the Broadcast communications service, the purpose is to send a packet that will be broadcast to all nodes connected to a given subnetwork. A broadcast packet is structured as described above. The Packet Type field is set to “45”, and the destination node network address is set to a broadcast indicator (i.e., all address bytes are set to “255”). After said broadcast packet leaves a source node, it traverses a set of network routers, and then is broadcast into a subnetwork by the last router in the network path. FIG. 5B illustrates an instance of broadcast. In FIG. 5B,[0082]node34 generates a broadcast packet destined forsubnetwork24. Said broadcast packet leavesnode34 and traverses all intermediate subnetworks (i.e.,32,12 and16) as described above. Whenrouter16 checks the network path section of the broadcast packet and finds an address consisting of “255” numerals, it encapsulates it into a protocol-specific broadcast packet and sends it out tosubnetwork24.
• In the Multicast communications service (i.e., Packet Type equals “40”), the purpose is to send a packet that will reach many but not all nodes contained in a specific subnetwork. FIG. 6C illustrates an instance of this service, where a multicast packet X leaves[0083]node34 and reachesnodes42 and44, but does not reachnode43. The header section of multicast packets comprises an extra Multicast Pointer field, as described above, that specifies the NPPT entry that points to the address of the end router in the network path which must send the multicast packet (e.g.,router16 of FIG. 6C). Multicast packet X traverses the network path exactly as any other packet until it reaches the router specified by Multicast Pointer (e.g., router16). If the subnetwork where multicast packet X must be sent out (e.g., subnetwork24) supports multicasting, said router encapsulates said packet X inside a protocol-specific multicast packet and sends it out. If said subnetwork does not support multicasting, said router will retrieve all multicast target addresses from the network address section of packet X (e.g.,nodes42 and44) and transmit a protocol-specific encapsulated packet to each specified destination. In the instance of FIG. 6C, incase subnetwork24 does not supportmulticasting router16 will send copies of packet X tonodes42 and44, as specified in the network address section of packet X.
Conclusion, Ramifications and Scope of Invention
• Thus, the reader will see that the present method of device-to-device communications provides a flexible method of source routed communications designed to fulfill requirements of interdevice communication which solves all limitations of existing methods. This comprehensive method can be used in any network independently of the associated communication protocol and/or media, reduces the complexity and cost of system implementation of existing communications methods, and supports one-to-many interdevice communication.[0084]
• While our above description contains many specificities, these should not be construed as limitations to the scope of the invention, but rather as an exemplification of one preferred embodiment thereof. Obviously, modifications and alterations will occur to others upon a reading and understanding of this specification such as, for example, several possible variations to the presented packet structure to include other network routing parameters, several variations to the presented communications services, especially to different modes of the Acknowledge communications service. The description above is intended, however, to include all such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.[0085]

Claims (8)

What is claimed is:

1. A method of source routing to implement device-to-device communications across a hybrid distributed device control network, said method comprising: a) orginating a packet at a source node; b) having the packet consist of sections including;

i) a header;

ii) a network path; and

iii) data.

c) encapsulating the packet in a protocol-specific packet used by the subnetwork of the source node; d) passing said protocol-specific packet to the first destination router in the network path;

e) having the router decapsulate the protocol-specific packet;

f) increment the next path destination index counter by one;

g) using the next path destination index counter to point to the next path destination address;

h) encapsulating the packet in a protocol-specific packet used by the next destination subnetwork; i) passing said protocol-specific packet to the next destination router in the network path; and

j) repeating the previous five steps until the packet reaches the final destination node.

2. The method according toclaim 1 wherein said encapsulation packet has an encapsulation header that contains a destination address.

3. The method according toclaim 1 wherein an acknowledgement is requested comprising the additional steps of:

k) orginating an acknowledgement packet at the destination node; l) having the acknowledgement packet consist of sections including;

i) a header;

ii) a network path; and

iii) data.

m) creating the acknowledgement network path by inversing the network path of the packet;

n) encapsulating the acknowledgement packet in a protocol-specific packet used by the subnetwork of the destination node; o) passing said protocol-specific packet to the first destination router in the acknowledgement network path;

p) having the router decapsulate the protocol-specific packet;

q) increment the next path destination index counter by one;

r) using the next path destination index counter to point to the next path destination address;

s) encapsulating the acknowledgement packet in a protocol-specific packet used by the destination address″s subnetwork; t) passing said protocol-specific packet to the next destination router in the network path; and

u) repeating the previous five steps until the acknowledgement packet reaches the source node.

4. The method according toclaim 1 wherein a packet is being broadcast to all the nodes in a subnetwork in which the additional steps of:

k) identifying the packet as a broadcast packet; and l) encapsulating the broadcast packet in a protocol-specific packet used by the destination subnetwork; and m) passing said broadcast packet to the nodes on the destination subnetwork.

5. The method according toclaim 1 wherein a packet is a multicast to the nodes in a specific subnetwork in which the additional steps of:

k) identifying the packet as a multicast packet; l) checking to see if the subnetwork supports multicasting;

m) encapsulating the multicast packet in a protocol-specific packet used by the subnetwork and pass said multicast packet to all of the nodes in the subnetwork if the subnetwork supports multicasting; and n) encapsulating the multicast packet in a protocol-specific packet used by the subnetwork and pass said multicast packet to the the nodes specified in the network address section of the packet, if the subnetwork does not support multicasting.

6. The method according toclaim 1 wherein said packet helder section contains the following fields:

a) Packet Type;

b) Packet ID;

c) Quality of Service;

d) Network Path Length;

e) Data Pointer;

f) Network Path Pointer Table;

g) Network Path Destination Index; and

h) Multicast Pointer.

6. The method according toclaim 1 wherein said Network Path section contains the following fields:

a) Network Type; and

b) Network Address.

7. The method according toclaim 1 wherein said Data section contains the following fields:

a) Data Length; and

b) Data Segment.

Images