Method and apparatus to control routing convergence during node reboot by entering a freeze state that inhibits routing convergence during the reboot process

• US 9,081,567 B1
• Filed: 07/21/2011
• Issued: 07/14/2015
• Est. Priority Date: 07/21/2011
• Status: Active Grant

• Current Assignee: Cisco Technology, Inc. (Cisco Systems, Inc.)
• Original Assignee: Cisco Technology, Inc. (Cisco Systems, Inc.)
• Inventors: Mahamuni, Atul B., Vasseur, Jean-Philippe
• Primary Examiner(s): Lee, Thomas
• Assistant Examiner(s): CASSITY, ROBERT A
• Application Number:
  US13/188,356
• Time in Patent Office: 1,454 Days
• Field of Search: 713/2
• US Class Current: 1/1
• CPC Class Codes: G06F 1/3203 Power management, i.e. event-based initiation of a power-saving mode
• CPC Class Codes: G06F 1/3206 Monitoring of events, devices or parameters that trigger a change in power modality

Abstract

In one embodiment, a technique may enable control of routing convergence in a computer network during reboot of a node of a routing topology represented by a directed acyclic graph (DAG). The rebooting node, e.g., a DAG root node, may transmit a reboot notification message to nodes of the routing topology indicating that the rebooting node is unavailable during a subsequent reboot process. In response to receiving and processing the reboot notification message, the nodes of the routing topology may enter a DAG freeze state that inhibits (prevents) routing convergence in the routing topology during the reboot process so as prevent adverse consequences on network stability, such as collapse of the DAG. Thereafter upon completion of the reboot process, the rebooting node may transmit a clear message to the nodes of the routing topology that instructs the nodes to exit the DAG freeze state.

Claims

1. A method, comprising:

transmitting, by a rebooting node before commencing a reboot process, a reboot notification message to specified nodes of a computer network having a routing topology represented by a directed acyclic graph (DAG), the reboot notification message indicating that the rebooting node of the DAG is unavailable during the reboot process, the reboot notification message further instructing the specified nodes to enter a DAG freeze state that inhibits routing convergence in the routing topology during the reboot process, the reboot notification message providing an identity of the rebooting node;

causing the specific nodes to receive traffic during the DAG freeze state, store the received traffic in a queue, monitor a length of the queue, and if the length of the queue exceeds a predetermined threshold, relay the reboot notification to other nodes of the routing topology; and

transmitting a clear message to the specified nodes upon completion of the reboot process, the clear message instructing the specified nodes to exit the DAG freeze state.

2. The method as in claim 1, wherein the reboot notification message identifies the rebooting node as a root node of the DAG.

3. The method as in claim 2, wherein the reboot notification message is a DAG Information Object (DIO) message and wherein transmitting the reboot notification message comprises:

generating the DIO message; and

configuring the DIO message with a reboot notification type-length-value (TLV) data set that transforms the DIO message to the reboot notification message.

4. The method as in claim 3, wherein configuring the DIO message comprises:

carrying a node and state attributes (NSA) object containing the reboot notification TLV data set within a metric container sub-option field of the DIO message, the NSA object providing information on node characteristics.

5. The method as in claim 1 further comprising:

configuring the reboot notification message with a reboot time type-length-value (TLV) data set that specifies an estimated time to complete the reboot process.

6. The method as in claim 1 further comprising:

configuring the reboot notification message with a propagation flag; and

asserting the propagation flag to specify propagation of the reboot notification message to a rank of nodes in the routing topology, wherein the rank of nodes comprises the specified nodes.

7. The method of claim 1 wherein transmitting the reboot notification message comprises:

setting a reboot notification timer configured with a value to enable the reboot notification message to traverse the routing topology so that the message is received and processed by the specified nodes.

8. The method of claim 2 further comprising:

generating a DAG Information Object (DIO) message; and

configuring the DIO message with a clear type-length-value (TLV) data set that transforms the DIO message to the clear message.

9. A method comprising:

receiving a reboot notification message transmitted by a rebooting node of a computer network before commencing a reboot process, the computer network having a routing topology represented by a directed acyclic graph (DAG), the reboot notification message indicating that the rebooting node is unavailable during the reboot process;

entering a DAG freeze state upon reception and processing of the reboot notification message, the DAG freeze state inhibiting routing convergence in the routing topology during the reboot process, the entering the DAG freeze state comprising receiving traffic during the DAG freeze state, storing the received traffic in a queue, monitoring a length of the queue, and if the length of the queue exceeds a predetermined threshold, relaying the reboot notification to other nodes of the routing topology;

receiving a clear message transmitted by the rebooting node upon completion of the reboot process; and

exiting the DAG freeze state in response to processing of the clear message.

10. The method as in claim 9 further comprising:

discarding low priority traffic of the received traffic from the queue when the length of the queue exceeds the predetermined threshold.

11. The method as in claim 9, wherein entering the DAG freeze state comprises:

setting a safety net timer having a value that is greater than a time estimated for completion of the reboot process;

exiting the DAG freeze state if the safety net timer expires prior to reception of the clear message; and

attempting routing convergence in the routing topology.

12. An apparatus comprising:

one or more network interfaces connected to a computer network having a routing topology represented by a directed acyclic graph (DAG);

a processor coupled to the network interfaces and adapted to execute one or more processes; and

a memory configured to store a process executable by the processor, the process when executed operable to;

transmit, by a rebooting node before commencing a reboot process, a reboot notification message to specified nodes of the routing topology, the reboot notification message indicating that the rebooting node of the DAG is unavailable during the reboot process, the reboot notification message including a propagation flag configured to identify the specified nodes, the reboot notification message instructing the specified nodes to enter a DAG freeze state that inhibits routing convergence in the routing topology during the reboot process, the reboot notification message further providing an identity of the rebooting node;

cause the specific nodes to receive traffic during the DAG freeze state, store the received traffic in a queue;

monitor a length of the queue, and if the length of the queue exceeds a predetermined threshold, relay the reboot notification to other nodes of the routing topology; and

transmit a clear message to the specified nodes upon completion of the reboot process, the clear message instructing the specified nodes to exit the DAG freeze state.

13. The apparatus as in claim 12, wherein the reboot notification message is a DAG Information Object (DIO) message and wherein the process when executed is operable to:

identify the rebooting node as a root node of the DAG within a field of the DIO message.

14. The apparatus as in claim 13 wherein the process when executed is operable to:

generate the DIO message; and

configure the DIO message with a reboot notification type-length-value (TLV) data set that transforms the DIO message to the reboot notification message.

15. The apparatus as in claim 12, wherein the process when executed is operable to:

assert the propagation flag to specify propagation of the reboot notification message to a rank of nodes in the routing topology, wherein the rank of nodes comprises the specified nodes.

16. The apparatus as in claim 12 wherein the process when executed is operable to:

non-assert the propagation flag to specify propagation of the reboot notification message to all nodes of the routing topology.

17. The apparatus as in claim 12, wherein the process when executed is operable to:

configure the reboot notification message with a reboot time type-length-value (TLV) data set that specifies an estimated time to complete the reboot process.

18. The apparatus as in claim 12, wherein the process when executed is operable to:

set a reboot notification timer configured with a value to enable the reboot notification message to traverse the routing topology so that the message is received and processed by the specified nodes.

19. The apparatus as in claim 13, wherein the process when executed is operable to:

generate the DIO message; and

configure the DIO message with a clear type-length-value (TLV) data set that transforms the DIO message to the clear message.

20. The apparatus as in claim 12, wherein the process when executed is operable to:

configure the reboot notification message with a propagation flag; and

assert the propagation flag to specify propagation of the reboot notification message to a rank of nodes in the routing topology, wherein the rank of nodes comprises the specified nodes.