Method and Apparatus for the Fast Detection of Connectivity Loss Between Devices in a Network

Abstract

A method and apparatus for quickly determining the status of a network device in a network communicates status inquiry messages over the forwarding plane of one network device to at least one neighboring network device. The status inquiry messages indicate the connectivity status with at least one neighboring network device.

Description

FIELD OF THE INVENTION

The present invention relates generally to communication systems, and more particularly, to propagating status inquiry messages in a network for determining the status of network devices.

BACKGROUND OF THE INVENTION

Many applications and services communicate information over computer networks, such as the Internet, in the form of data packets. Network devices (e.g., routers and switches) process and forward data packets to neighboring network devices in the network according to addressing information contained therein. Logic contained in the network devices is responsible for processing and forwarding received protocol data packets. This logic includes the “forwarding plane” and the “control plane”. The forwarding plane forwards protocol data packets relatively quickly because minimal data processing is performed. Conversely, the control plane forwards data packets (e.g., protocol data packets) much slower because the packets require additional processing (i.e., they must be queued or scheduled) before they are forwarded.

Network devices use routing protocols (e.g., OSPF or IS-IS Hello) to both determine network routes, and to store such routes in local routing tables. Accordingly, upon receipt of a packet, a receiving network device accesses the routing table to determine the route to use for forwarding the packets. During normal error free transmission on a computer network, packets are successfully forwarded and received between network devices. Undesirably, however, interruptions of data flow often occur causing network devices to lose connectivity with neighboring network devices in the network. In response to this problem, routing protocols commonly include “keep-alive” services to detect the loss of connectivity between neighboring network devices. Specifically, keep-alive services typically include “hello” messages that are communicated via the links of neighboring network devices to determine if such neighboring devices are properly operating.

Existing hello messages are typically processed by the control plane and thus, produce a significant communication delay (e.g., from a few seconds to tens of seconds) before a network device is alerted of a loss of connectivity with a neighbor. Such delay often inhibits a network device from recovering from a loss of connectivity, commonly causing data to be lost. For example, many applications and services, such as packetized voice, require notification of a loss of connectivity almost instantaneously (e.g., on the order of less than about 50-200 milliseconds). Accordingly, such a relatively long delay does not adequately provide the intended safeguards of the keep-alive services.

SUMMARY OF THE INVENTION

In accordance with one aspect of the invention, a method and apparatus for quickly determining the status of a network device in a network communicates status inquiry messages over the forwarding plane of one network device to at least one neighboring network device. The status inquiry messages indicate the connectivity status with at least one neighboring network device.

In various embodiments, the status inquiry messages are transmitted periodically to the at least one neighboring network device. In addition, the network device may communicate an advertisement message over the control plane to discover the at least one neighboring network device prior to the communication of the status inquiry messages over the forwarding plane. Following the identification of the at least one neighboring network device, the network devices may negotiate communication parameters for the exchange of status inquiry messages with the at least one network device. Among other things, the network device may be a switch or a router and the network may be a local area network (LAN), wide area network (WAN), or the Internet.

Preferred embodiments of the invention are implemented as a computer program having a computer usable medium with computer readable program code thereon. The computer readable code may be read and utilized by a computer system in accordance with conventional processes.

BRIEF DESCRIPTION OF THE DRAWINGS

The foregoing description and advantages of the invention will be appreciated more fully from the following further description thereof with reference to the accompanying drawings wherein:

FIG. 1 schematically shows a network arrangement that may be used with illustrative embodiments of the invention.

FIG. 2 shows an illustrative process of determining the connectivity status in the network shown in FIG. 1.

FIG. 3 shows an illustrative process describing in more detail the neighbor discovery function of the advertisement protocol shown in FIG. 2.

FIG. 4 shows an illustrative process describing in more detail how adjacencies are formed between neighboring network devices shown in FIG. 2.

FIG. 5 shows an illustrative process describing in more detail the negotiation of operating parameters for two adjacent network devices shown in FIG. 2.

FIG. 6 shows an illustrative process of checking the connectivity status of adjacent neighboring network devices.

FIG. 7 schematically shows a network devices that may implement preferred embodiments of the invention.

DESCRIPTION OF ILLUSTRATIVE EMBODIMENTS

In illustrative embodiments of the invention, status inquiry messages are forwarded via the forwarding plane of a network device to a neighboring network device to quickly determine the connectivity status between the devices. Details and examples of illustrative embodiments are discussed below.

It should also be noted that the term “network device” is used herein to broadly represent any type of device that functions to process and/or forward data packets to neighboring devices in a network. Such devices may include, among other things, hosts, routers, multiplexers, computer systems, and switches. In illustrative embodiments, network devices discussed herein have both a control plane and a forwarding plane.

FIG. 1 shows an IP subnet 10 that may be used with illustrative embodiments of the invention. Specifically, the subnet 10 includes a plurality of neighboring network devices 12. Each of the network devices 12 include, among other things, interfaces for communicating with each other. In illustrative embodiments, the network devices 12 are routers, such as the ACCELLAR™ Router, available from Nortel Networks Limited of Brampton, Ontario, Canada. The subnet 10 may be coupled with another network, such as the Internet. The network devices 12 communicate their operational status via messages in accordance with illustrative embodiments described herein.

FIG. 2 shows an illustrative process utilized by a network device 12 in the IF subnet 10 fox determining connectivity status with its neighboring network devices 12. To accomplish this task, a plurality of different types of messages are used. One type of message is an advertisement message, which is used for discovering neighboring network devices 12 and transmitting preliminary communication parameters to such network devices 12. A second type of message is a status inquiry message, which is a simple fixed format message exchanged between the network devices 12 for discovering the status of a network device 12. Status inquiry messages have the capability of operating at high frequencies for permitting network devices 12 to rapidily detect losses of connectivity.

The process in FIG. 2 begins at step 200, in which one of the network devices 12 first becomes initialed (i.e., when the network device becomes active). Specifically, when first initialized, the one network device 12 multicasts a solicitation message to neighboring network devices on the IP subnet 10.

The process continues to step 202, in which another network device 12 (the “inquiring network device 12”), after receiving the solicitation message from the noted network device, sends one or more advertisement messages (via its control plane) to both discover neighboring network devices 12 on the IP subnet 10, and to announce preliminary communication parameters for the status inquiry messages. The initial advertisement messages from the inquiring network device 12 are sent immediately upon receiving the solicitation message, while subsequent advertisement messages are sent periodically (discussed in further detail below). In illustrative embodiments, the inquiring network device 12 transmits one advertisement message from each of its interfaces.

The process continues to step 204, in which “adjacencies” are deemed to be formed between the inquiring network device 12 and each discovered neighboring network device 12. In illustrative embodiments, the inquiring network device 12 considers an adjacency to be formed when it reads its IP address in an advertisement message received from a neighboring network device 12. After an adjacency is established, the process continues to step 206, in which the inquiring network device 12 negotiates common communication parameters for status inquiry messages to be transmitted to and received from each located adjacent network device 12. This negotiation preferably is an agreement on the lowest common denominator (discussed below) of each preliminary communication parameter of the inquiring network device 12 and the adjacent network device 12. Each network device 12 that is adjacent to the inquiring network device 12 can have a different set of communication parameters with the inquiring network device 12.

Once operating parameters have been negotiated and agreed upon, the process continues to step 208, in which the connectivity status between the adjacent network devices 12 is determined. Specifically, status inquiry messages are exchanged (e.g., via a unicast) between the adjacent network devices 12 via their respective forwarding planes to determine connectivity status. In illustrative embodiments, the status inquiry messages are transmitted at a much higher frequency than the advertisement messages. Details of the various steps of this process are discussed below with reference to FIGS. 3-7.

FIG. 3 shows an illustrative process describing in more detail the neighbor discovery function of the advertisement messages shown in step 202 of FIG. 2. The process begins at step 300, in which the inquiring network device 12 multicasts advertisement messages on the IP subnet 10 over each of its IP interfaces. These advertisement messages advertise both the inquiring network device's 12 support of the process of FIG. 2, and its currently configured preliminary communication parameters.

As discussed with respect to step 202 above, an advertisement message from the inquiring network device 12 sent in response to a solicitation message is sent immediately to a neighbor on the subnet, while subsequent advertisement messages from the inquiring network device 12 are periodically transmitted at predefined time intervals (referred to herein as an “AdvertisementInterval”) at relatively low frequencies, and with a small amount of jitter. A random jitter factor is included to ensure that the advertisement messages are sent at different times, thus preventing the network devices 12 from becoming overloaded with receiving multiple messages transmitted at the same time. In illustrative embodiments, the advertisement interval is about 600 seconds. Of course, this interval value is an example and thus, not intended to limit various embodiments of the invention. This interval is not applicable when the network device is first initialized.

Once one of the advertisement messages is received by a neighboring network device 12, the process continues to step 302, in which the inquiring network device's IP interface (over which the received advertisement message was sent) is added to a list of IP interfaces contained in the neighboring network device's own advertisement to be transmitted to the inquiring network device 12. Thus, the inquiring network device 12 is deemed to have been “discovered” by the neighboring network device 12. It should be noted that the list of neighbor IP interfaces is for the subnet that transmitted the advertisement only and not for any other subnet.

FIG. 4 shows an illustrative process describing in more detail how an adjacency may be formed between the inquiring network device 12 and a neighboring network device 12 as shown in step 204 of FIG. 2. The process begins at step 400, in which it is determined if an advertisement having the IP interface of the inquiring network device 12 is received by the inquiring network device 12 (from the neighboring network device 12) before expiration of an AdvertisementDeadInterval timer. More specifically, the AdvertisementDeadInterval timer times a predefined time period (referred to as “AdvenisementDeadInterval period”) to check the status of an adjacency. In particular, as shown below, if the requiting network device 12 does not receive an advertisement (with the IP interface of the inquiring network device 12) by the end of the AdvertisementDeadInterval period, then an adjacency is deemed to be terminated. Accordingly, in such, case, connectivity is considered, to be lost between the two network devices 12. Status inquiry messages (see FIG. 6) thus no longer are transmitted to the neighboring network device 12. The inquiring network device 12 nevertheless continues to transmit advertisements over the interface for the neighboring network device 12 in an effort to discover and/or establish another adjacency. These new advertisements, however, do not include the IP interface of the neighboring network device 12.

Returning to the query in step 400, if no such advertisement is received before the end of the AdvertisementDeadInterval period, then the process ends. Conversely, if such advertisement is received in that brae period, then the process continues to step 402, in which the timer is started (if this is the first such advertisement), or restarted (if this is not the first such advertisement). The neighbor network device 12 IP interface then is added to a reply advertisement (step 404), which then is transmitted to the neighboring network device 12 (step 406). After receipt of the reply advertisement, the neighboring network device 12 also executes the process shown in FIG. 4.

In illustrative embodiments, the value of the AdvertisementDeadInterval is larger than the AdvertisementInterval (this interval being noted above). For example, the AdvertisementDeadInterval may be about twice the value of the AdvertisementInterval.

As noted above, in addition to neighbor discovery and establishing adjacency, advertisement messages are also used to communicate various preliminary communication parameters. For example, the following communication parameters for determining connectivity status may be used: the StatusInquiryInterval and the PeerDeadInterval. The StatusInquiryInterval and the PeerDeadInterval define the communication parameters utilized by adjacent network devices 12 when transmitting status inquiry messages. Specifically, the StatusInquiryInterval indicates how often (in terms of milliseconds) status inquiry messages will be sent by a network device. For example, if the StatusInquiryInterval is three milliseconds, then a given network device should send status inquiry messages at least every three milliseconds. The PeerDeadInterval indicates how long a network device should wait (in terms of milliseconds) from the last received status inquiry message before declaring a loss of connectivity. The value of the PeerDeadInterval illustratively is larger than the StatusInquiryInterval. For example, the PeerDeadInterval may be three times the value of the StatusInquiryInterval.

Of course, it is possible for two adjacent network devices to be configured with different preliminary values for their communication parameters. These values must be agreed upon before the transmission of status inquiry messages can begin. FIG. 5 shows an illustrative process describing in more detail the negotiation of one communication parameter (e.g., the StatusInquiryInterval or the PeerDeadInterval) for two adjacent network devices 12 on IP subnet 10 as shown in step 206 of FIG. 2. Continuing with the above discussion, the process is described in terms of the inquiring network device 12 and the neighboring network device 12.

The process begins at step 500, in which the value of a parameter (the StatusInquiryInterval in succeeding examples) received from the inquiring network device 12 is compared to the value of the same parameter in the neighboring network device 12. If the parameter values are determined to be the same at step 502, then such value is used as an agreed on communication parameter, thus ending the process.

Conversely, if the parameter values are determined to be different at step 502, then the process continues to step 504, in which a common value must be selected. In illustrative embodiments, the common value selected is the larger of the two parameter values, thus ending the process. Specifically, it is preferred that the larger parameter value be selected (instead of the smaller one) in illustrative embodiments to accommodate environments where one of the two network devices 12 cannot operate at the same speed as the other device. Thus, the faster device operates at a slower rate to accommodate the slower device. This process was referred to above as using the lowest common denominator. Of course, the different parameters values (StatusInquiryInterval or PeerDeadInterval) may be selected in a different manner. Accordingly, selecting the larger value is illustrative only and is not intended to limit the scope of the invention.

In a manner to other timing intervals discussed above, each network device 12 may apply a jitter factor to the StatusInquiryInterval and the PeerDeadInterval. For example, the actual parameters used may between 75% and 100% of the selected values. This ensures that the status inquiry messages are not sent at the same time, thus avoiding overloading a receiving network device 12.

FIG. 6 shows an illustrative process of checking the connectivity status of adjacent neighboring network devices 12, referenced at least in part in step 208 of FIG. 2. In illustrative embodiments, this entire process is executed in the forwarding plane of each of the network devices 12 executing this process. The connectivity status is checked every set time interval (discussed below) to ensure that a neighboring network device has been in communication within a reasonable time interval. Details are discussed below.

As discussed above with regard to FIG. 2, the process of FIG. 6 uses status inquiry messages. In illustrative embodiments, each status inquiry message is carried in an IP packet having a header with the address of a network device to receive the message, and a body having a four bit field. The first bit is the StatusInqiuryHeard bit, which indicates if the network device 12 sending the message has received a status inquiry message from the network device 12 receiving the message. In some cases, the StatusInquiryHeard bit is set only if the message is received before expiration of PeerDeadInterval timer (discussed below). The StatusInquiryHeard bit has two states, namely “on” or “off.” A status inquiry message is considered to be validated when the StatusInquiryHeard bit is set to be “on.”

In a manner similar to other processes discussed herein, this process is described from the point of view of the inquiring network device 12 and its neighboring network device 12. Accordingly, an adjacency is deemed to have been previously established between such network devices 12. In such case, prior to execution of this process, the inquiring network device 12 and neighboring network device 12 (via their respective forwarding planes) have each transmitted validated status inquiry messages to each other. This may have occurred by the inquiring network device 12 transmitting an unvalidated status inquiry message to the neighboring network device 12, and the neighboring device 12 responding with a validated status inquiry message. Execution of the process of FIG. 6 is begun by the inquiring network device 12 after receipt of the validated status inquiry message from the neighboring network device 12. Receipt of a validated status inquiry message means that successful bi-directional communication has been established and/or maintained.

The process begins at step 600, in which the inquiring network device 12 periodically transmits validated status inquiry messages to the neighboring network device 12 via the BP interface logically connected with the neighboring network device 12. In illustrative embodiments, status inquiry messages are transmitted via such IP interface once every time interval equal to the StatusInquiryInterval. For example, this time interval may be three milliseconds. Of course, the inquiring network device 12 may send status inquiry messages through other of its IP interfaces to other network devices 12 in the subnet 10. As noted above, if an adjacency is deemed to have been terminated, then no more statue inquiry messages are transmitted.

The process then continues to step 602, in which it is determined it a validated status inquiry message has been received before expiration of a given timer (referred to herein as a PeerDeadInterval timer). More particularly, this timer is started upon receipt of the last validated status inquiry message received from the neighbor network device 12. If a subsequent validated status inquiry message is not received, from the neighboring network device 12 within the time period specified by the timer, then (as shown in this figure) connectivity is deemed to have been lost.

Accordingly, if no validated status inquiry message has bean received before the expiration of the timer, then the process continues to step 604, in which loss of connectivity is considered to have occurred. In such case, a loss of connectivity is signaled, thus ending the process. For example, the inquiring network device 12 may signal the loss of connectivity to an application such as a routing protocol, which can respond appropriately. No additional status inquiry messages thus are to be transmitted until an adjacency is established. Connectivity may be reestablished, if desired, by the methods already discussed. In particular, connectivity may be reestablished by the discovery/adjacency process, and then through the exchange of status inquiry messages.

Conversely, if at step 602 a validated status inquiry message has been received, then the process continues to step 604, in which the timer is reset. In illustrative embodiments, the time period that the timer is reset is greater than the StatusInquiryInterval. For example, if the StatusInquiryInterval is three milliseconds, the time period of this timer may be twelve milliseconds.

The process continues to step 608 in which successful bi-directional communication is deemed to have been maintained. The process then loops back to step 600, in which validated status inquiry bits are periodically transmitted to the neighboring network device 12.

To summarize, the connectivity status between neighboring network devices 12 can be in any one of three states. Namely, the connectivity may be in an up state, a down state, or an unknown state (i.e., the connectivity status is not known). When in the up state, status inquiry messages are transmitted from the receiving network device 12 with the StatusInquiryHeard bit set to “on.” When in the down state, no status inquiry messages are transmitted. When in the unknown state, status inquiry messages are transmitted from the receiving network device with the StatusInquiryHeard bit set to “off.” This unknown state is deemed to occur between the time that an adjacency is first established, and the time that a validated status inquiry message is received from a neighboring network device 12.

In another embodiment, instead of a StatusInquiryHeard bit, status inquiry messages may contain an identification field that contains a unique, non-decreasing, and unsigned 32 bit number. An example of how such a number would be generated is from a 32 bit time of day (TOD) clock. Similar to the process in the previous embodiment discussed above, the connectivity status between neighboring network devices 12 can be in any one of three states. Namely, the connectivity is in an up state when status inquiry messages with the same 32 bit number found in the sending network device 12 is transmitted from the receiving network device 12. When in the down state, no status inquiry messages are transmitted. When in the unknown state, status inquiry messages are transmitted from the receiving network device with a different 32 bit numbers than that found in the sending network device 12.

In an illustrative embodiment, the aforementioned method for determining connectivity status is implemented as hardware (e.g., line cards, application, specific integrated circuits, FPGAs, and/or digital signal processors), or other related components in the network device 12. Moreover, as noted above, the process shown to FIG. 6 is executed substantially entirely in the forwarding plane of each network device 12, while the processes shown in FIGS. 3-5 illustratively are executed in the control plane of each network device 12. Accordingly, the status inquiry messages can be transmitted at much higher frequencies than other types of messages, thus supporting applications requiring rapid status information, such as packetized voice.

FIG. 7 schematically shows one of the network devices 12 (as shown in FIG. 1), which contains logic for processing messages in accordance with FIGS. 2-6. The network device 12 includes a status module 72 for executing the process shown in FIG. 2, a control plane 74, and a forwarding plane 76. Logic within the status mod ale 72 transmits messages via advertisement messages 77 or status inquiry messages 78. The status module 72 forwards advertisement messages 77 via the control plane 74, and forwards the status inquiry messages 26 to the neighboring network devices 12 (see FIG. 1) via the forwarding plane 76. As previously noted, by enabling the status module 72 to forward the status inquiry messages 78 over the forwarding plane 76, high frequency rates required to send and receive the status inquiry messages 78 can be achieved. Of course, this embodiment is illustrative only and is not intended to limit the scope of the invention. Other illustrative embodiments may be implemented without a status module. In such embodiments, the communication or the advertisement messages and the status inquiry messages is done directly between the control plane and the forwarding plane.

The aforementioned embodiments may also be implemented at least in part in any conventional computer programming language. For example, some embodiments may be implemented to a procedural programming language (e.g., “C”) or an object oriented programming language (e.g., “C++”).

In an alternative embodiment, the disclosed apparatus, and method for may be implemented as a computer program product for use with a computer system. Such implementation may include a series of computer instructions fixed either on a tangible medium, such as a computer readable medium (e.g., a diskette, CD-ROM, ROM, or fixed disk) or transmittable to a computer system, via a modem or other interface device, such as a communications adapter connected to a network over a medium. The medium maybe either a tangible medium (e.g., optical or analog communications lines) or a medium implemented with wireless techniques (e.g., microwave, infrared or other transmission techniques). The series of computer instructions embodies all or part of the functionality previously described herein with respect to the system. Those skilled in the art should appreciate that such computer instruction can be written in a number of programming languages for use with many computer architectures or operating systems. Furthermore, such instructions may be stored in any memory device, such as semiconductor, magnetic, optical or other memory devices, and may be transmitted using any communications technology, such as optical, infrared, microwave, or other transmission technologies. It is expected that such a computer program product may be distributed as a removable medium with accompanying printed or electronic documentation (e.g., shrink wrapped software), preloaded with a computer system (e.g., on system ROM or fixed disk), or distributed from a server or electronic bulletin board over the network (e.g., the Internet or World Wide Web). Of course, some embodiments of the invention may be implemented as a combination of both software (e.g., a computer program product) and hardware. Still other embodiments of the invention are implemented as entirely hardware, or entirely software (e.g., a computer program product).

Although various exemplary embodiments of the invention have been disclosed, it should be apparent to those skilled in the art that various changes and modifications can be made that will achieve some of the advantages of the invention without departing from the true scope of the invention. These and other obvious modifications are intended to be covered by the appended claims.

Claims

1. A method of determining a time interval to be used by a first node in a communication network in detecting loss of connectivity to a second node of die communication network according to a connectivity loss detection protocol, the method comprising: receiving, at the first node, a communication from the second node, the communication indicating a setting for the time interval preferred by the second node;comparing, at the first node, the setting of die date interval preferred by the second node to a setting of the time interval preferred by the first node;when the setting of the time interval preferred by the first node equals the setting of the time interval preferred by the second node, setting at the first node a value of the time interval to be used in the connectivity loss detection protocol to the setting of the time interval, preferred by the first node and the second node;when the setting of the time interval preferred by the first node does not or equal to the setting of the time interval preferred by the second node, setting at the first node the value of the time interval to he used in the connectivity loss detection protocol to a larger of the setting of the time interval preferred by the second node and the setting of the time interval preferred by the second node.

2. The method of claim 1, wherein the value of the time interval is used to determine time intervals between sending of connectivity loss detection messages.

3. The method of claim 1, wherein the value of the time interval is used to detect loss of connectivity due to absence of connectivity loss detection messages received from the second node within the time interval.

4. The method of claim 1, performed to determine a value of a first time interval used to determine lime intervals between sending of connectivity loss detection messages and performed again to determine a value of a second time interval used to detect loss of connectivity due to absence of connectivity loss detection messages.

5. The method of claim 4, wherein the value of the second time interval exceeds the value of the first time interval.

6. The method of claim 5, wherein the value of the second time interval is up to three times the value of the first time interval.

7. The method of claim 2, comprising transmitting connectivity loss detection messages from the first node to the second node at times determined from the value of the time interval.

8. The method of claim 7, wherein the connectivity loss detection messages are transmitted via a forwarding plane for detection of connectivity loss in the forwarding plane.

9. The method of claim 8, wherein the first node transmits connectivity loss detection messages to the second node at successive times separated by the value of the time interval minus applied jitter.

10. The method of claim 9, wherein the applied jitter has a randomized value within a predetermined range.

11. The method of claim 10, wherein the value of the time interval minus applied jitter is at least 75% of the value of the time interval.

12. The method of claim 1, comprising transmitting a communication from first node to second node indicating the setting of the time interval preferred by the first node.

13. The method of claim 12, wherein the communication indicating the setting of the time interval preferred by the first node is transmitted to the second node responsive to initiation of the first node.

14. The method of claim 12, wherein communications indicating dm setting of the time interval preferred by the first node are transmitted in the second node periodically while the first node and the second node are active in the communication network.

15. The method of claim 14, wherein the communications indicating the setting of the time interval preferred by the first node are transmitted to the second node periodically at times determined according to a time interval negotiated between the first node and the second node.

16. The method of claim 15, wherein the time interval is negotiated by the first node according to the method of claim 1.

17. The method of claim 1, performed successively each time a communication indicating a setting for the time interval preferred by the second node is received by the first node.

18. The method of claim 17, wherein the communications indicating a setting for the time interval preferred by the second node are received periodically by the first node while the first node and the second node are active in the communication network.

19. The method of claim 18, wherein the communications indicating the setting of the time interval preferred by the second node ace received by the first node periodically at times determined according to a time interval negotiated between the first node and the second node.

20. The method of claim 19, wherein the time interval is negotiated by the first node according to the method of claim 1.