US20140229587A1 - Networking as a service - Google Patents

Abstract

Networking as a Service (NaaS) delivers network services using remote appliances controlled by a hosted, multi-tenant management system. The system may include a heartbeating process for communication between a web-based server and appliances, in which the appliances periodically contact the management system on the server. The heartbeating process allows the appliances to maintain a completely up-to-date configuration. Furthermore, heartbeating allows for comprehensive monitoring of appliances and for software distribution. The system may also include means for authenticating appliances, without the need for pre-installed PSKs or certificates.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS
• This application is a continuation of U.S. Ser. No. 13/603,272, filed Sep. 4, 2012, entitled “NETWORKING AS A SERVICE,” which is a continuation of U.S. Ser. No. 12/357,390, filed Jan. 21, 2009, entitled “NETWORKING AS A SERVICE: DELIVERING NETWORK SERVICES USING REMOTE APPLIANCES CONTROLLED VIA A HOSTED, MULTI-TENANT MANAGEMENT SYSTEM,” which claims priority to U.S. Provisional Patent Application No. 61/021,883, filed Jan. 17, 2008, entitled “NETWORKING AS A SERVICE: DELIVERING NETWORK SERVICES USING REMOTE APPLIANCES CONTROLLED VIA A HOSTED, MULTI-TENANT MANAGEMENT SYSTEM,” all of which are incorporated by reference.
BACKGROUND
• Many organizations have locations where there are limited or no IT staff to maintain and operate network equipment. Current solutions often require expensive on-site staff; service contracts with local expert contractors; and/or investment in expensive remote network management and monitoring software, systems, or services. Additionally, current solutions allow an administrator to make changes to the network in multiple places like a CLI, Web-GUI, or network management system—for which there is no good way to synchronize changes from each place where changes occur. There is a need for lower cost, more efficient solutions for providing network services.
• In order to set-up and run a computer network, many organizations face the challenges of: building a server and installing and configuring vendor software to manage network devices; and installing, configuring, and maintaining a network management appliance. Furthermore, many organizations desire, but do not have access to solutions that allow: people from multiple domains, virtual networks, or companies to operate different portions of the network management software; and support of 10,000+ devices. There is a need for network management applications that provide solutions to these software challenges.
• Since network administrators can change a device's configuration settings both locally on the device and at the network management station, the actual configuration of a device can differ from that intended by the management station. Additionally, because of Network Address Translation (NAT), some devices cannot be reached when communication is initiated by the management station. There is a need for a solution to overcome these communication issues that affect network integrity.
• Most networking solutions today require the networking device to be reachable when communication is initiated from the management station. They also typically allow changes to be made both at the management station and at the networking device itself. Because of this, not all networking topologies are possible, and the actual networking device's configuration can “morph out from under” the management station's expectation of that device's settings. There is a need for solutions that provide greater flexibility in network topology and greater predictability.
• Network Address Translation (NAT) enables a LAN to use one set of IP addresses for internal traffic and a second set for external traffic. Because of NAT, some devices cannot be reached when communication is initiated by the management station, since the addresses for internal traffic may be hidden from the management station. Most networking solutions today require the networking device to be reachable when communication is initiated from the management station. There is a need to overcome this communication barrier between the management station and networking devices.
• Network Administrators spend a lot of time just keeping-up on all the possible versions of operational software available for a piece of networking gear. There is a need to provide network administrators with a simpler and more efficient software solution. Even worse, there is a continual flow of maintenance and security patches each month. There is a need to simplify this maintenance process.
• Home users often need networking devices issued by their employer in order to work from home, but their employer does not find it cost-effective to send a technician to the employees' home to configure the networking device. Employers also do not wish to spend time stocking those networking devices at the office and pre-configuring them on a per-unit basis before sending them home with their employees. Some networking vendors have implemented configuration servers which store one configuration for each networking device to be deployed, but even in such solutions, the configurations are associated with the networking device, rather than the end-user making use of that networking device. There is a need for a more efficient, cost effective method for employers to provide their employees with networking devices for home-use.
• Authenticating a networking device to a Web server often involves some key material being pre-installed on the device, such as a certificate, or pre-shared key (PSK). There is a need for efficient methods of performing authentication, authorization and accounting (AAA) on an insecure device—without a pre-installed PSK or similar.
BRIEF DESCRIPTION OF THE FIGURES
• FIG. 1 illustrates networking as a service.
• FIGS. 2 through 6 illustrate various aspects of delivering network management as a multi-tenant hosted software as a service.
• FIG. 7 shows networking components used in a heartbeating process.
• FIG. 8 illustrates authentication of client premises equipment to a Web service.
DETAILED DESCRIPTION
• Embodiments will now be described in detail with reference to the drawings, which are provided as illustrative examples so as to enable those skilled in the art to practice the invention. Notably, the figures and examples below are not meant to limit the scope of the invention to a single embodiment; other embodiments are possible by way of interchange of some or all of the described or illustrated elements. Moreover, where certain elements of an embodiment can be partially or fully implemented using known components, only those portions of such known components that are necessary for an understanding of the invention might be described, and detailed descriptions of other portions of such known components may be omitted so as not to obscure the described embodiment. In the present specification, an embodiment showing a singular component should not be considered limiting; rather, the invention is intended to encompass other embodiments including a plurality of the same component, and vice-versa, unless explicitly stated otherwise herein. Applicants do not intend for any term in the specification or claims to be ascribed an uncommon or special meaning unless explicitly set forth as such. Embodiments encompass present and future known equivalents to the known components referred to herein by way of illustration.
• A technique includes building network appliances (routers, firewalls, VPN, WAN Optimization, Switching, WLAN Access point, etc.) that may have no command line interpreter (CLI)—a text-based input-output method implemented on most networking devices—or Web-GUI and may only be configured, managed, and updated via the appliance calling “home” via a secured Web-service over the Internet on a periodic basis, such as every minute. The model of appliances described herein using web services for remote management via a hosted multi-tenant system could apply to any internet connected special purpose appliances where automated management is important. Anything from internet connected search appliances to internet connected refrigerators.
• Networking as a Service (NaaS) allows the device to be remotely controlled over the Internet without the need for a person on site to configure or manage the appliance. In addition the only way a person can make moves, adds, or changes to the network is through the hosted web services—making the web service a “system of record” improves security, control, and standardization of configuration of large numbers of devices. This is a management system that can scale to manage 10K, 100K, or 1M+ remote appliances.
• NaaS virtualizes the network, enabling a customer to accelerate information technology innovation using an agile infrastructure, while reducing networking costs. Important components of a NaaS solution are: a subscription service business model that provides a lower cost solution than traditional network solutions; an application development platform that links network settings to business requirements; a configurations database with automated and virtualized support for each appliance function, configuration file and setting; virtualized SaaS-based management to speed deployment and increase control and standardization of branch networks; and virtualized network appliances that arc dynamically configured to be integrated routing, security, VPN and WAN optimization appliances, with the option to include LAN, WLAN, and AFE.
• NaaS allows delivery of network equipment to customers as a self-managed service, instead of an equipment purchase that requires significant software integration and operational labor or software. By offering NaaS, the service provider can reduce the customer's non-bandwidth networking expense (equipment, software, & labor) by a factor of 5×-10×, depending on the size and complexity of the network, while generating gross margins comparable to current network equipment margins.
• Key aspects of the NaaS business model may include the following. NaaS may be implemented as a subscription based service. For example, customers “subscribe” to a network where they pay a monthly fee (invoiced annually, perhaps) based on the number of networking devices and functionality used. Customer business benefits include: pay as you go (or use); no hardware depreciation costs; and no annual software maintenance fees.
• NaaS bypasses traditional networking distribution. With NaaS, the networking solution is sold directly via inside sales to small/medium sized customers; and direct sales (pre-qualified via inside sales) are made to large customers. The product is drop shipped to the customer via FedEx or UPS, for example, with inventory management and logistics outsourced to retail companies such as Amazon. After some initial customer traction, there may also be a potential opportunity to leverage non-networking IT channels, such as MSFT channel, application developers and SIs, etc.
• NaaS provides a hosted management infrastructure. NaaS hosts the management infrastructure at authorized data centers, where each customer's data is encrypted and secured from other customers. A customer benefit is the option to delegate management for all or part of the NaaS network to third parties. Two examples of this are provided. Car Company has the option to manage and control a minimal level of security required for dealer networks to connect to Car Company directly. Car Company dealers can manage the rest of the devices (or manage remaining devices on their own). Global Insurance Company can outsource network configuration options to a virtual network operator (VNO) while retaining security configuration responsibilities internally. This capability will enable new types of partners and applications to interact with the network to automate workflows and business processes—potentially creating new distribution models.
• NaaS as a self-service model. Instead of requiring networking experts to design, install, monitor, and manage the network, organizations can self-manage their network with non-expert personnel who can access their NaaS-managed devices with a web browser anywhere in the world. This is made possible by the simplified network configuration options and automated installation and configuration process of NaaS.
• NaaS allows the use of commodity hardware, multi-core software, and virtual machines. Standard off-the shelf appliances with multi-core capable software and virtual machines may be used to deliver broad network functionality with acceptable levels of reliability.
• NaaS may be enhanced with inclusive product design. A NaaS provider may build and control the hardware, software, configuration options, and management infrastructure to ensure compatibility, reliability, and configuration.
• NaaS permits customer to control delegation of operation. Customers control who can access their NaaS configuration, creating opportunity for new customers to enable new channels to operate and manage the devices.
• NaaS may be tightly integrated with Web 2.0 Applications to drive additional value. As a first example, enable XML based API to share data and automate NaaS management. As a second example, integrate with Google Domains and Applications to enable customers to manage internal and external DNS within a single application. Other examples could include: social networking (find people who can manage a NaaS); compare best practices for NaaS deployment; and integrate with SaaS network monitoring applications like Splunk or open source network management applications such as IT Groundworks to integrate with existing network infrastructures. Note that since embodiments include control and monitoring of networking devices via software running on Internet servers, (“in the cloud”), data from other Web servers can easily be integrated with monitoring data to add new value. This holds equally true for data on servers not resident on the general Internet, but in a company's private intranet. For example, a customer that uses Splunk to collect log data from a variety of their other networking devices and servers could combine Web-accessible logs and statistics with those logs to deduce new insights that otherwise would not be possible.
• Some benefits of embodiments over the state of the art may include: increased control of the network, network appliances, and network services; increased security via standardization; increased business flexibility—the customer can repurpose, that is re-use the networking device for a different purpose or task the appliances based on business need with a click of a web service interface); increase speed, ease of deployment, and execution; increased simplification, for example 1 way vs. 3 to control a device, which may also be limited to pre-tested configurations; increased reliability; and 5× to 10× cost reduction for the customer.
• FIG. 1 shows an implementation of NaaS for a branch-office type solution.Human facilities110 are connected to adata center120 via any wide-area networking technology. The human facilities comprisenetworking devices112 which allowend users111 to access the network. Thehuman facilities110 may be telecommuters, a temporary business location, or a branch office. Thedata center120 comprisesdevices122 andservers124. Thedata center120 andhuman facilities110 are connected via the Internet to a virtualized software as a service (SaaS)management center130. The virtualizedSaaS management center130 comprisesservers134. The virtualizedSaaS management center130 may be integrated withbusiness applications140 for automated management of network devices.
• A NaaS solution includes: (1) discover local devices, network services and an uplink carrier; (2) configure a network using a business wizard and a library of proven and tested network configurations; (3) form and maintain a secure network; and (4) automatically monitor networking devices using a heartbeat process, including auto-upgrades of software. These are discussed in detail below.
• NaaS Simplifies Networking
• One reason that traditional solutions are complex is that there are many versions of code that can run on each networking device making for thousands of potential combinations. NaaS greatly simplifies this by ensuring each device is running one and only one version of software, thus making only one combination to test. Having only one version of software on the hub, one version on the CPE, and one version of the management service provides an on-going sustainable advantage. To facilitate this single version of code solution, the following may be used: databases of tested working solutions for business problems—see co-pending application entitled Virtualization of Networking Services, inventors Carl Steven Mower and Matthew Alan Palmer, filed Jan. 21, 2009, incorporated by reference herein; and expand application integration—for example, use XML API to integrate the network with business applications.
• NaaS Solutions are Easily Adapted
• NaaS solutions are very easily modified at several different levels of the system, shown here:

• TABLE I
  Modification of NaaS Solutions

  			Development Type
  Type of Change	Example:	Change made by:	& Schedule
  Customer	Change IP Sec or	Customer's IT staff for	On - demand
  Configuration	WEP Security Keys	customer specific instance	1-5 minutes
  Settings
  Security	New virus signature	Network service provider's	On - demand
  Content	available	operations staff, all	1-8 hours
  		customers receive at once
  Configuration	New network	Network service provider's	On - demand
  Templates	configuration	operations staff, all	1-4 days
  	template	customers receive at once
  Web Service	Surface to	Network service provider's	Scheduled Update
  Functionality	customer's IT staff a	operations staff, all	1-2 weeks
  	template setting that	customers receive at once
  	previously had been
  	hard-coded
  CPE & Hub	Add IPv6 or WAN	Network service provider's	Scheduled Update
  Functionality	Optimization	operations staff, all	4-6 weeks
  		customers receive at once

• Delivering Network Management Applications as a Multi-Tenant Hosted Software as a Service
• Embodiments provide for developing, designing, and delivering network management software applications from a multi-tenant software as a service model using Web services that are distributed across the internet; to provide network management applications to multiple organizations or companies. Network management functions are provided via a series of Web pages which allow users to configure and monitor networking devices. The actual re-configuration of the networking devices and collection of performance statistics occurs via a separate Web-based services, as described below.
• The process for delivering network management applications is shown inFIGS. 2-6. InFIG. 2 the NaaS user views, and may edit, configuration and/or monitoring data. Theuser205 logs in to the network, via theInternet210, and is authenticated using a standard protocol, such as: a username and password; the user's company's AD/LDAP, provided by thehub240 at their DMZ (a computer host or small network inserted as a “neutral zone” between a company's private network and the outside public network); and a secure-ID token. Web pages are served from theWeb UI server215 in a stateless fashion. All states are maintained in thedatabase server220.
• Note that the network shown inFIG. 1 and the networks shown inFIGS. 2-6 are equivalent. For example, all of the servers shown inFIGS. 2-6 are equivalent to pieces of theserver134 inFIG. 1.
• InFIG. 3, the heartbeat process is illustrated. The heartbeat process is a method of communication between anetworking device235 and anetwork management server250, otherwise known as the heartbeat server, via theInternet245. Every minute, each networking device (CPE)235 sends one packet to theheartbeat server250, the packet containing—all logs since the last, acknowledged, heartbeat (communication). Theheartbeat server250 accepts the data and responds with one packet containing a UID of a configuration file that thenetworking device235 should be running; and build number(s) of software image(s) that thedevice235 should be running. The response packet from theheartbeat server250 is implicitly an acknowledgement that the logs and performance data have been accepted, so that theCPE235 purges those values from its RAM. If theCPE235 is not running the latest configuration or software image(s), it fetches them via a secure transport method, such as sFTP or SSL, from theserver255. Note that thenetworking devices235 are remotely upgraded with no action taken by any user(s) connected to that device. Furthermore, note that thenetworking devices235 are conveying their log entries and monitored statistics to theheartbeat server250 on each heartbeat. Centralizing the collection and storage of this data aids the customer in meeting government compliance and regulatory requirements, and frees them from implementing infrastructure to save this data on their own.
• InFIG. 4, the networking device image upgrade process is illustrated. If the last heartbeat returns a newer build number than it is currently running, then thenetworking device235 registers that it is running “stale” code. Thenetworking device235 slowly fetches, via a secure transport method, a newer image. A delay in fetching the image is intentionally added between each block fetched, so as to not saturate the limited WAN link. Thedevice235 burns the image into one of two non-volatile ram or hard-drive partitions on the device for active software, (the partition not holding the currently running image). On the next reboot, thedevice235 reboots into the new partition. If the image does not successfully receive a heartbeat reply in the first five minutes, it reverts to the older image. If that backup/older image likewise does not receive a reply, a third “emergency” partition is booted which contains a “safe”/simple boot image that merely copies a “safe” image from theserver255, stored in a devicesoftware data base270, and boots it. This process may be used similarly to upgrade thehub240.
• FIG. 5 illustrates the process of theuser205 publishing a modified configuration, stored in theconfiguration database225, to the networking device(s)235. Theuser205 instructs the system to “publish” a modified configuration through a web page on the user's browser. TheWeb UI server215 generates the actual unique configuration file for each affected device. TheWeb UI server215 copies those new configurations to each of theservers255, where the configurations are stored in the short term monitoring database. TheWeb UI server215 updates the heartbeat server's shortterm monitoring database260 with the UID of those new configurations for each affected device. On the next heartbeat, the affected devices will receive that new UID for their configuration file, and recognizing they have a “stale” configuration, will fetch the new configuration file via a secure transport method, such as sFTP or SSL. Within approximately five minutes, each affected device may be running the newly published configuration.
• InFIG. 6, the process of collecting network device statistics is illustrated. Theheartbeat servers250 distill and push data to the long-term database230 from the short-term databases260, every minute or so.
• Using the method of delivering network management applications described above allows: multiple organizations to leverage the same infrastructure; applications to be hosted and maintained at multiple locations on the internet for redundancy; multiple parties to access the network management application to operate/control specific groups of appliances across multiple organizations and security boundaries; access to Web Services API via the internet to develop addition network management applications. This method of network management is readily scalable to large numbers of parties and devices and is more scalable than prior art methods of network management.
• Heartbeating
• A key implementation aspect of a NaaS solution is the general process of networking devices “heartbeating” to the Web-based servers.FIG. 7 shows components involved in the heartbeat process. Theheartbeat servers710, such as Pareto heartbeat servers, may be generic CPUs running proprietary code. The hardware for these servers is not proprietary—any type of hardware or operating system can be used. Theheartbeat servers710 are connected to theInternet720. A branch/homeInternet access device730 is used to connect the branch/home location to theInternet720. For example, theinternet access device730 may be a DSL or cable modem. Upstream branch/home networking devices740 are optional networking or computing devices positioned between theInternet access device730 and thenetworking device750. Thenetworking device750, such as an Edge networking device, is located at the branch/home location. Downstream branch/home networking devices760 are optional networking or computing devices connected directly or indirectly to thenetworking device750, excluding theInternet access device730 and anyupstream networking devices740.
• In embodiments, the heartbeating process, also referred to herein as the “phone home” process, may work as follows. At the factory, thenetworking device750 is programmed with a list of DNS names for one ormore heartbeat servers710 on theInternet720. Thedevice750 is also programmed with a unique PKI certificate which contains the unique serial number of thedevice750. This certificate is signed by a private key only known to thedevice750. The corresponding public key is stored on theheartbeat servers710. Thenetworking device750 powers-on and gains an IP address via DHCP. In the DHCP response, thedevice750 receives the address of one or more DNS servers. Thenetworking device750 automatically discovers alldownstream devices760. This may be accomplished by numerous methods, including scans of IP address ranges to determine which addresses respond, capturing packets promiscuously, inferring device type by Ethernet MAC address, port scans of devices found, etc. Thenetworking device750 simultaneously discovers allupstream devices740 using methods similar to those for discoveringdownstream devices760.
• Thenetworking device750 packages all information discovered about the upstream anddownstream devices740 and760. Thedevice750 attempts to open a SSL connection to thefirst heartbeat server710 on the list programmed at the factory using the unique PKI certificate programmed on thatparticular device750 at the factory. If a connection cannot be established with thefirst heartbeat server710 on the list, thedevice750 moves on to try thesecond heartbeat server710 on the list. The process is repeated until a connection is established or the list is exhausted. In the event the list is exhausted without a connection successfully being formed, thedevice750 starts again at the top of the list ofheartbeat servers710.
• Using the connection established toheartbeat server710, thenetworking device750 sends the information it has collected about its upstream and downstream neighboring devices,740 and760, to theheartbeat server710. Additionally, it sends any log entries and performance measurements collected since the last successful heartbeat. Theheartbeat server710 acknowledges receipt of this information, thus constituting a “successful heartbeat” and causing thedevice 750 to clear the record of the log entries and performance measurements just uploaded.
• Using that same connection, thenetworking device750 queries theheartbeat server710 to determine the proper configuration thedevice750 should be using. Each and every configuration is given a unique identifier which is returned as the result of this query. Thedevice750 compares the unique identifier with the identifier for the configuration it is actually using. If they differ, thedevice750 separately initiates a secure download of the appropriate configuration from theheartbeat server710. The filename for the configuration is the same as the unique identifier, for this implementation. Once downloaded, thedevice750 switches to use that new configuration immediately. Using the same connection, thedevice750 queries theheartbeat server710 to determine the proper software version that thedevice750 should be using. Each and every software image is given a unique identifier which is returned as the result of this query along with an effective date/time. Thedevice750 compares the unique identifier with the identifier for the software image it is actually using. If they differ, thedevice750 separately initiates a slow, secure download of the appropriate configuration from theheartbeat server710. The filename for the configuration is the same as the unique identifier. The transfer is measured and slow, with an intentional delay between the fetch of each block on the order of 1 second, so as to avoid saturating the limited link to the Internet. Once downloaded, the image is burned to the device's flash memory. Once the effective date/time is reached, thedevice750 reboots into and runs the new software image.
• The heartbeat process repeats, for example once every 1 minute. The same process takes place for all networking devices on the network. Note that the networking devices are remotely upgraded with no action taken by any user(s) connected to that device. On each heartbeat, each device compares its current software version to that advertised by the heartbeat server and if the software that the device is running is not the most recent advertised, it fetches newer software from the heartbeat server, stores it permanently in the device, and starts running it. Furthermore, note that the networking devices are conveying their log entries and monitored statistics to the Web service on each heartbeat. Centralizing the collection and storage of this data aids the customer in meeting government compliance and regulatory requirements, and frees them from implementing infrastructure to save this data on their own.
• Phone Home for Device Configuration
• Each networking device “phones home” on a periodic basis. Each such contact, initiated by the networking device is called a “heartbeat”. When a device heartbeats to a heartbeat server on the Internet, it requests the unique identifier of the configuration that device should be running If the configuration the device is actually running differs, the device fetches that newer configuration and starts using it immediately.
• Several “heartbeat servers” are implemented to ensure availability of one such server at all times. Each networking device is configured with a list of such heartbeat servers as part of their configuration. When a device is first installed on the network, it is programmed with a starting set of heartbeat servers to use in fetching its first configuration.
• Networking devices have no local user interface to make configuration changes—all changes are made at the management station. Such changes made at the management station ultimately are propagated to the heartbeat servers.
• Significant benefits of using techniques described in this paper over alternative approaches include the following. Since devices have no local user interfaces, their configuration settings do not diverge from those intended by the management station. Since each networking device is programmed at the factory with a starting list of heartbeat servers, no configuration of the networking device is required for it to get its first configuration settings.
• Phone Home for Device Monitoring
• Each networking device “phones home” on a periodic basis. Each such contact, initiated by the networking device is called a “heartbeat”. When a device heartbeats to a heartbeat server on the Internet, it pushes all performance statistics and monitoring status to the management station as part of that heartbeat. See the related “Phone Home for Device Configuration” for more details of the heartbeat process.
• Significant benefits of techniques described in this paper over alternative approaches include the following. Since the networking device initiates communication with the management station, this solution can traverse NAT. Note that the networking devices are conveying their log entries and monitored statistics to the Web service on each heartbeat. Centralizing the collection and storage of this data aids the customer in meeting government compliance and regulatory requirements, and frees them from implementing infrastructure to save this data on their own.
• Phone Home for Software Distribution
• Each networking device “phones home” on a periodic basis. Each such contact, initiated by the networking device is called a “heartbeat”. When a device heartbeats to a heartbeat server on the Internet, it requests the unique identifier of the operational software that device should be running If the operational software the device is actually running differs, the device fetches that software and starts using it.
• Several “heartbeat servers” are implemented to ensure availability of one such server at all times. Each networking device is configured with a list of such heartbeat servers as part of their configuration. When a device is first installed on the network, it is programmed with a starting set of heartbeat servers to use in updating its operational software the first time.
• Significant benefits of techniques described in this paper over alternative approaches include the following. For each type of networking device there is exactly one and only one current version of its operational software. Network Administrators no longer need to be concerned about the multitude of possible software versions available for today's devices. This solution also has the advantage that even a device that has been held in a warehouse for many months after manufacture will be running the latest version of operational software upon the first heartbeat after being installed on the network, without the need for operator intervention. Note that the networking devices are remotely upgraded with no action taken by any user(s) connected to that device. On each heartbeat, each device compares its current software version to that advertised by the heartbeat server and if the software that the device is running is not the most recent advertised, it fetches newer software from the heartbeat server, stores it permanently in the device, and starts running it.
• Zero-Touch Deployment via One-Time Password
• Another component to a NaaS solution is the ability to deploy devices with no other configuration tasks expected of the end-user except to enter a one-time password. Networking device configuration is a function not of the serial number of the particular networking device per se, but rather is a function of the human that wishes to use the networking device to accomplish productive work. Unfortunately, most networking device configuration tools only configure devices as a function of the networking device's identity, not the end-user behind that device.
• In embodiments, the networking device is shipped from the manufacturer directly to the end-user's location, without being “pre-configured” by the employee's IT staff. The end-user performs the following steps. The end-user unpacks the networking device. He connects the upstream port to the cable/DSL modem. He connects his PC/laptop to a downstream port. He powers-on the networking device and opens a Web browser on their PC/laptop. The networking device “captures” the end-user's HTML session and presents a screen asking the end-user for his one-time-password. The end-user enters the onetime password given to them by his IT staff. When the end-user was initially approved to receive a home networking device, the IT staff did the following: added the end-user's username to the network management system as an authorized user; put the end-user in the appropriate group of authorized end-users with similar business requirements; e-mailed the one-time password generated by the network management system to the end-user; when the networking device receives the one-time password, the device provides the password to the network management system, and receives a device configuration suitable to address the needs of the particular end-user.
• In embodiments, the process for deployment of a networking device in a home office are as follows. The CPE is shipped from the manufacturer directly to the end-user. While it may often be the case that the CPE is actually sent to the customers' IT department for distribution to the end-users, this need not be assumed. Even when the device is sent to the customer's IT department, the initial deployment model does not require the IT department to perform any operation(s) on each and every individual CPE, as that would be far too onerous in large-scale deployments. The CPE is connected to up-stream devices and is powered-on. It is assumed that the end-user is savvy enough to connect the CPE to the up-stream device and power it on. LEDs may be used on the CPE to indicate a link for all Ethernet ports. All Ethernet ports should have auto-crossover circuitry in order to detect and connect the use of incorrect Ethernet cable types.
• Once the CPE is connected and powered on, the CPE gets an IP address from an up-stream DHCP server. In the event there is no DHCP server, the CPE will default to 192.168.0.1 and intercept any HTML packets to a page served locally from the CPE that explains that the CPE was unable to get an address via DHCP along with some suggested steps for the end-user to try. The CPE collects information about its surroundings. The CPE will note which down-stream ports have links, and the information given it by the DHCP server, such as the address or addresses of the DNS server(s). The CPE will also note its firmware version(s) and hardware configuration information. The CPE will also start discovering other attributes of the local network including, for example: a) using a security scanner, such as nMap, to discover what types of devices and operating systems are connected to the CPE; b) a port scan of the discovered devices to identify which ports are operational; c) an attempt to attach to basic services of the discovered devices such as http, SMPT, printing, basic routing protocols (OSPF, BGP, etc.); and d) scanning for existing wireless networks to identify SSID's, Security levels, etc. This data will be uploaded to the Web service and used to assist with site specific configuration settings and requirements.
• The CPE contacts the NaaS management service. The CPE will have a certificate installed in its non-volatile RAM at time of manufacturing which contains the CPE's serial number. The CPE sends this certificate and the information about its surroundings to the NaaS management service. Assuming the certificate is found to be valid, all further communication from the CPE to the network management service will be in an encrypted SSL session. Network management service validates the certificate and checks the CPE firmware version. If the CPE's firmware is out of date, the response to the CPE indicates this fact which causes the CPE to securely download, burn and reboot into this new firmware version.
• At this point the end-user is challenged for one-time credentials. Although, if the particular CPE had been pre-registered by the customer's IT department, this is skipped. At this point, the CPE does not have its policy. Once the end-user opens a Web browser on their downstream PC, any HTTP request is intercepted and redirected to the network management service. The end-user or authorized third party installer (i.e. FedEx delivery person) enters their one-time credentials, which may have been issued to them from an IT department. Note that using such one-time credentials allows the IT department to issue any CPE to the end-user without having to note which serial number was issued to which employee. It also allows the CPE to be “drop-shipped” directly from manufacturing to the end-user. Assuming the network management service validates the one-time credentials, the policy the IT department assigned to that end-user is associated with the CPE's serial number. The CPE requests a policy from the network management service. The CPE securely pulls its configuration/policy from the network management service, and reconfigures itself in accordance with that configuration/policy. Typically, creating one or more IPsec or SSL tunnels to the customer's hub device. On any successive power-ups, the CPE would already have an associated configuration/policy, and therefore would not require one-time credentials from the end-user.
• The end user(s) must authenticate to the network. Just like a user would be prompted for a user credential to access the LAN or an SSL/IPSec Remote Access VPN. When the user attempts to connect to the network they will need to enter their security credentials before network access is granted.
• Using this process, the end-user need not perform any configuration operations other than to simply enter a one-time-password. Furthermore, the employer need not stock and pre-configure the networking devices, but can simply instruct the vendor to ship units directly from their factory to the end-user's location.
• CPE Authentication with no PSK on the CPE
• Authenticating a device, a customer premises equipment (CPE), to a Web service typically involves some key material being pre-installed on the CPE, perhaps a certificate, or sometimes a pre-shared key (PSK). However, embodiments include a process to authenticate a CPE, to a Web service without having any PSK material on the CPE. This process permits secure authentication, authorization and accounting (AAA) of CPE using a secure Web service without storing a PSK on the CPE.
• FIG. 8 illustrates the process of authentication of aCPE810 to aWeb service820. Initially, aCPE810 is totally locked down as a walled garden with access to only DNS and theWeb service820. The end user opens hisbrowser830 and makes any web request. See request1. Captive portal software fields the request and sends a splashscreen HTML response832 that has a redirect to the HTTPS link to the Webservice login server826, with information in the redirect URL about theCPE810, including information such as the CPE Mac Address, etc.
• The end user'sbrowser830 performs the HTTPS redirect in thesplash screen832 that sends a request to the Web service'slogin Web server826. Seerequest2. Thelogin Web server826 generates a short lived NONCE, and stores it associated with the remote CPE's MAC address in thedatabase824. A NONCE is a number used once, such as a random or pseudorandom number issues in an authentication protocol.
• The login Web server responds with thelogin form834. See request3. The end user enters the device password, which in this example is a PSK that has been stored in the database entry associated with the device previously sent securely to the end user by an administrator. Thelogin Web server826 validates this password. If this process is successful then the process continues, otherwise the user is presented again with the login screen including an error message.
• The login Web server responds to a correct password with aconfirm page836, placing a JavaScript stub in the confirm page with the URL of thecaptive portal812 and the generated NONCE. Seerequest4. The end user'sweb browser830 displays theconfirm836, executes the JavaScript that issues an HTTP (insecure) POST of the generated NONCE back to thecaptive portal812 which stores it in non-volatile RAM (NVRAM)814. Seerequest5.
• Heart beat daemon816 andheartbeat server822 handshake over HTTPS (secure) connection. Seerequest6.Heartbeat daemon816 sends the NONCE back to theheartbeat server822 in the next scheduled heart beat cycle, before the NONCE expires,heartbeat server822 validates the NONCE, looking it up by device MAC address in thedatabase824, and if valid it sends the MD5 hash of the user entered device password, or an equivalent cryptographic hash function. Theheartbeat daemon816 uses the MD5 hash of the device password as an authentication token for subsequent heartbeats.
• Note that the network inFIG. 8 is equivalent to part of the network inFIG. 1, except thatFIG. 8 provides more detail of the components involved in authentication of a networking device.
• Although embodiment have been particularly described, it should be readily apparent to those of ordinary skill in the relevant art that changes and modifications in form and details may be made without departing from the spirit and scope of the disclosure. It is intended that the appended claims encompass such changes and modifications.

Claims (21)

What is claimed is:

1. A method comprising:

monitoring for one or more heartbeat packets from a network device of a network;

receiving from the network device a first heartbeat packet from the network device, the heartbeat packet including one or more logs of the network device, the logs comprising new log information since a last acknowledged heartbeat packet but not comprising old log information before the last acknowledged heartbeat packet.

response to the first heartbeat packet, sending a first response network packet to the network device, the first response network packet including a configuration identifier of a latest configuration for the network device, and the first response network packet further including a building identifier of a latest software image for the network device.

2. The method ofclaim 1, further comprising providing the latest configuration to the network device if the network device is not running the latest configuration.

3. The method ofclaim 1, where the providing the latest configuration to the network device comprises providing one or more new configuration files associated with the latest configuration.

4. The method ofclaim 1, further comprising providing the latest software image to the network device if the network device is not running the latest software image.

5. The method ofclaim 1, wherein the latest configuration is one that the network device should be running.

6. The method ofclaim 1, wherein the latest software image is one that the network device should be running.

7. The method ofclaim 1, wherein the first response network packet is an implicit acknowledgement to the network device that the logs included in the first response network packet have been accepted

8. The method ofclaim 1, wherein the network device is configured to purge log information upon receiving the first response network packet.

9. The method ofclaim 1, wherein the network device is customer premises equipment (CPE).

10. The method ofclaim 1, further comprising:

receiving an instruction to publish a modified configuration;

in response to the instruction, generating the latest configuration for the network device;

updating a monitoring database with the configuration identifier of the latest configuration, the monitoring database providing the configuration identifier of the latest configuration for use in the first response network packet.

11. A system comprising:

means for monitoring for one or more heartbeat packets from a network device of a network;

means for receiving from the network device a first heartbeat packet from the network device, the heartbeat packet including one or more logs of the network device, the logs comprising new log information since a last acknowledged heartbeat packet but not comprising old log information before the last acknowledged heartbeat packet.

means for sending a first response network packet to the network device in response to the first heartbeat packet, the first response network packet including a configuration identifier of a latest configuration for the network device, and the first response network packet further including a building identifier of a latest software image for the network device.

12. A system comprising:

a processor;

a computer readable medium including instructions configured to cause the processor to perform the operations of:

monitoring for one or more heartbeat packets from a network device of a network;

receiving from the network device a first heartbeat packet from the network device, the heartbeat packet including one or more logs of the network device, the logs comprising new log information since a last acknowledged heartbeat packet but not comprising old log information before the last acknowledged heartbeat packet;

in response to the first heartbeat packet, sending a first response network packet to the network device, the first response network packet including a configuration identifier of a latest configuration for the network device, and the first response network packet further including a building identifier of a latest software image for the network device.

13. The system ofclaim 12, further comprising providing the latest configuration to the network device if the network device is not running the latest configuration.

14. The system ofclaim 12, where the providing the latest configuration to the network device comprises providing one or more new configuration files associated with the latest configuration.

15. The system ofclaim 12, further comprising providing the latest software image to the network device if the network device is not running the latest software image.

16. The system ofclaim 12, wherein the latest configuration is one that the network device should be running.

17. The system ofclaim 12, wherein the latest software image is one that the network device should be running.

18. The system ofclaim 12, wherein the first response network packet is an implicit acknowledgement to the network device that the logs included in the first response network packet have been accepted

19. The system ofclaim 12, wherein the network device is configured to purge log information upon receiving the first response network packet.

20. The system ofclaim 12, wherein the network device is customer premises equipment (CPE).

21. The system ofclaim 12, further comprising:

receiving an instruction to publish a modified configuration;

in response to the instruction, generating the latest configuration for the network device;

updating a monitoring database with the configuration identifier of the latest configuration, the monitoring database providing the configuration identifier of the latest configuration for use in the first response network packet.

Images