Autonomic traffic load balancing in link aggregation groups
Mechanisms are provided for performing traffic load balancing on ingress traffic directed to a Link Aggregation Group (LAG). The mechanisms monitor a ingress traffic load across a plurality of links of the Link Aggregation Group (LAG). The mechanisms determine if the ingress traffic load across the plurality of links is unbalanced. Moreover, the mechanisms, in response to determining that the ingress traffic load across the plurality of links is unbalanced, send a message to a switch associated with the LAG requesting the switch to modify routing of ingress traffic to the LAG to perform ingress traffic load balancing.
1. A method, in a device comprising a processor, for performing traffic load balancing on ingress traffic directed to a Link Aggregation Group (LAG), the method comprising:
monitoring, by the device, an ingress traffic load across a plurality of links of the Link Aggregation Group (LAG);
determining, by the device, if the ingress traffic load across the plurality of links is unbalanced; and
in response to determining that the ingress traffic load across the plurality of links is unbalanced, sending, by the device, a message to a switch associated with the LAG requesting the switch to modify routing of ingress traffic to the LAG to perform ingress traffic load balancing, wherein the message is an extended Link Aggregation Control Protocol (LACP) message extended to include one or more fields for specifying parameters for the ingress traffic load balancing to the switch, and wherein the one or more fields comprises a first field in a LACP header of the extended LACP message that is reserved for an actor, and a second field in the LACP header of the extended LACP message that is reserved for a partner.
2. The method of claim 1 , wherein the parameters for the ingress traffic load balancing comprises a first parameter identifying a first link in the LAG, that is an overused link in the LAG, and a second parameter identifying a second link to which ingress traffic directed to the overused link should be redirected.
3. The method of claim 2 , wherein the switch, in response to receiving the extended LACP message, updates a routing table in the switch to redirect ingress traffic from the first link to the second link.
4. The method of claim 1 , wherein the extended LACP message is a heartbeat message sent from the device to the switch.
5. The method of claim 1 , wherein monitoring an ingress traffic load across a plurality of links of the LAG comprises:
generating a statistical measure of an amount of data flowing through each link of the LAG over a specified time quanta; and
comparing the statistical measures of each of the links in the LAG to each other link in the LAG to determine if one link has a statistical measure that is equal to or greater than a threshold amount greater than the statistical measure of the other links.
6. The method of claim 1 , wherein the switch, in response to receiving the message, updates a routing table of the switch to redirect ingress traffic directed to an overused link in the LAG to another link in the LAG.
7. The method of claim 6 , wherein the routing table of the switch is extended to include a directive field for each entry in the routing table, wherein the directive field stores a pointer to a directive database that indicates a directive for routing purposes, and wherein updating the routing table comprises updating a directive field of an entry in the routing table associated with the overused link to point to a directive in the directive database to route ingress traffic to the another link.
8. The method of claim 1 , wherein the device is a network adapter coupled to a host system.
9. The method of claim 1 , wherein the device is a network adapter of a host computing system, the method is performed by the network adapter, and wherein the ingress traffic load is an ingress traffic load for traffic flowing to the network adapter from the switch.