Loading patent…
AssigneeArista Networks, Inc.
InventorsYashvir Singh, Satish Kumar Eerpini, Digvijay Singh Gahlot, Rajnish Gupta, Francois Labonte
Estimating the term…
The present disclosure is directed to detection of downed leaf devices in a distributed schedule fabric (DSF). In accordance with some embodiments, keep alive messages are exchanged among the leaf devices in the spine-leaf network. In some embodiments, the keep alive messages are sent periodically so that when a first leaf device does not see a keep alive message from another second leaf device after a certain period of time, the first leaf device can consider that the second leaf device is down or otherwise not reachable. Each leaf device can include a reachability table to keep track of which leaf devices are reachable and which are not reachable. If a leaf device is not reachable, the packet can be dropped, otherwise the packet can be further processed and forwarded to the next hop.
A method in a first network device among a plurality of network devices, the method comprising the first network device:
The method of claim 6, wherein the aggregation group is an ECMP (Equal Cost Multipath) group.
The method of claim 6, wherein the aggregation group is a LAG (Link Aggregation Group) group.
A first network device comprising:
The first network device of claim 9, wherein the computer-readable storage device further comprises instructions for controlling the one or more computer processors to receive a message from a device to which the plurality of network devices send heartbeat messages, wherein the received message indicates that the second network device is not reachable.
The first network device of claim 9, wherein the computer-readable storage device further comprises instructions for controlling the one or more computer processors to periodically send and receive heartbeat messages to and from the plurality of network devices, wherein the second network device is deemed not reachable when the first network device does not receive a heartbeat message from the second network device within a predetermined period of time.
The first network device of claim 9, wherein the second network device is in an aggregation group, the method further comprising, in response to determining that the second network device not reachable, removing ports on the second network device from the aggregation group.
The first network device of claim 13, wherein the aggregation group is an ECMP group.
The first network device of claim 13, wherein the aggregation group is a LAG group.
A non-transitory computer-readable storage device in a first network device, the non-transitory computer-readable storage device having stored thereon computer executable instructions, which when executed, cause the first network device to:
The non-transitory computer-readable storage device of claim 16, wherein the second network device is in an aggregation group, the method further comprising, in response to determining that the second network device not reachable, removing ports on the second network device from the aggregation group.
The non-transitory computer-readable storage device of claim 18, wherein the aggregation group is an ECMP group.
The non-transitory computer-readable storage device of claim 18, wherein the aggregation group is a LAG group.
The non-transitory computer-readable storage device of claim 16, wherein the data table associates the egress port with a validity parameter, wherein updating the data table includes setting the validity parameter to a first data state that indicates the second network device is not reachable, wherein packets whose corresponding egress ports are on the second network device are dropped by virtue of the validity parameters associated with the egress ports being set to the first data state.
Dependencies are inferred from the claim text.