Push-based hierarchical state propagation within a multi-chassis network device

• US 8,149,691 B1
• Filed: 03/25/2009
• Issued: 04/03/2012
• Est. Priority Date: 11/16/2005
• Status: Active Grant

• Current Assignee: Juniper Networks Incorporated
• Original Assignee: Juniper Networks Incorporated
• Inventors: Chadalavada, Bharani, Krishnaswamy, Umesh, Tuplur, Raj
• Primary Examiner(s): Rutkowski, Jeffrey M
• Application Number:
  US12/411,113
• Time in Patent Office: 1,105 Days
• Field of Search: 370216-220, 370386-389, 370/398, 370/503, 370/509, 370/512, 709/208, 709/220, 709/221, 709/248, 714 2- 46, 714/47
• US Class Current: 370/219
• CPC Class Codes: H04L 45/00 Routing or path finding of packets in data switching networks routing or path finding in wireless networks H04W40/00
• CPC Class Codes: H04L 45/58 Association of routers
• CPC Class Codes: H04L 45/60 Router architectures

Abstract

A multi-chassis network device sends state information to internal consumers within the multi-chassis device via a hierarchical distribution. As one example, a primary master routing engine within a control node of a multi-chassis router forwards state information to local routing engines within other chassis, which in turn distribute the state information to consumers on each chassis. Each local routing engine defers sending acknowledgement to the master routing engine until acknowledgements have been received from all consumers serviced by the local routing engine. Embodiments of the invention may reduce control plane data traffic and convergence times associated with distribution of state updates in the multi-chassis network device.

Claims

1. A multi-chassis network device comprising:

a routing engine for a first chassis configured to manage state information for the multi-chassis network device, wherein the routing engine for the first chassis is further configured to generate a state update indicating a change to the state information and send the state update to a second chassis;

a standby routing engine on the first chassis, wherein the routing engine of the first chassis sends the state update to the standby routing engine of the first chassis before sending the state update to the second chassis;

a routing engine for the second chassis configured to manage local state information for the second chassis, wherein the routing engine of the second chassis operates as an intermediate consumer of the state update from the routing engine of the first chassis; and

a consumer within the second chassis,

wherein the routing engine of the second chassis receives the state update from the routing engine of the first chassis and, in response, updates the local state information in accordance with the state update to synchronize the local state information with the state information of the first chassis and provides the state update to the consumer, and

wherein, in the event of failure of the routing engine of the first chassis, the standby routing engine of the first chassis continues to send state updates to the second chassis on behalf of the failed routing engine of the first chassis.

2. The multi-chassis network device of claim 1, wherein the routing engine of the second chassis receives a first acknowledgement regarding the state update from the consumer, and sends a second acknowledgement regarding the state update to the routing engine of the first chassis only after receiving the first acknowledgement.

3. The multi-chassis network device of claim 1, wherein the routing engine of the second chassis receives an acknowledgement request from the routing engine of the first chassis, and adds a marker within a temporally-linked data structure of state updates in response to the acknowledgement request.

4. The multi-chassis router of claim 3, wherein the marker is a ksync object.

5. The multi-chassis network device of claim 1, further comprising a standby routing engine of the second chassis, wherein the routing engine of the second chassis sends the state update to the standby routing engine of the second chassis.

6. The multi-chassis network device of claim 1, wherein the state update comprises one or more objects in a temporally-linked data structure.

7. The multi-chassis network device of claim 1, wherein the multi-chassis network device is a router.

8. A method for distributing a state update in a multi-chassis network device having a first chassis and a second chassis, the method comprising:

managing state information for the multi-chassis network device with a control unit of the second chassis, wherein the control unit of the second chassis operates as a master control unit for the multi-chassis network device;

generating, with the control unit of the second chassis, a state update that includes a change to the state information with the control unit of the second chassis;

sending the state update from the control unit of the second chassis to a standby control unit of the second chassis;

sending the state update from the control unit of the second chassis to a control unit of a first chassis;

receiving with the control unit of the first chassis the state update from the control unit of the second chassis of the multi-chassis network device;

updating, with the control unit of the first chassis, local state information in accordance with the state update to synchronize the local state information with the state information of the second chassis;

providing, with the control unit of the first chassis, the state update to a consumer within the first chassis; and

upon failure of the control unit of the second chassis, sending subsequent state updates from the standby control unit of the second chassis to the control unit of the first chassis.

9. The method of claim 8, further comprising:

receiving a first acknowledgement regarding the state update from the consumer with the control unit of the first chassis; and

sending a second acknowledgement regarding the state update with the control unit of the first chassis to the control unit of the second chassis only after receiving the first acknowledgement.

10. The method of claim 9, further comprising:

receiving an acknowledgement request from the control unit of the second chassis with the control unit of the first chassis, and

adding a marker within a temporally-linked data structure of state updates in response to the acknowledgement request.

11. The method of claim 10, wherein the marker is a ksync object.

12. The method of claim 8, wherein the control unit of the first chassis is a primary routing engine of the first chassis, the method further comprising:

sending the state update to a standby routing engine of the first chassis before providing the state update to the consumer; and

failing over to the standby routing engine of the first chassis in the event the primary routing engine of the first chassis fails.

13. The method of claim 8, wherein the state update comprises one or more objects in a temporally-linked data structure.

14. The method of claim 8, wherein the multi-chassis network device is a router.

15. A non-transitory computer-readable storage medium containing instructions that cause a programmable processor in a multi-chassis network device to:

manage state information for the multi-chassis network device with a control unit of a second chassis;

generate a state update that includes a change to the state information with the control unit of the second chassis;

send the state update from the control unit of the second chassis to a standby control unit of the second chassis;

send the state update from the control unit of the second chassis to a control unit of a first chassis;

receive with the control unit of the first chassis the state update from the control unit of the second chassis of the multi-chassis network device;

update, with the control unit of the first chassis, local state information in accordance with the state update to synchronize the local state information with the state information of the second chassis;

provide, with the control unit of the first chassis, the state update to a consumer within the first chassis;

receive a first acknowledgement regarding the state update from the consumer with the control unit of the first chassis; and

send a second acknowledgement regarding the state update with the control unit of the first chassis to the control unit of the second chassis only after receiving the first acknowledgement.

16. The non-transitory computer-readable medium of claim 15, containing instructions that cause the programmable processor to:

receive an acknowledgement request from the control unit of the second chassis with the control unit of the first chassis, and

add a marker within a temporally-linked data structure of state updates in response to the acknowledgement request.

17. The non-transitory computer-readable medium of claim 16, wherein the marker is a ksync object.