Network switch load balance optimization
View Patent ↗A network load balancing method in a blade grouping chassis can be provided. The method can include detecting system conditions among blades coupled to a single blade grouping chassis. The method also can include maintaining a rotation of selected ones of the blades among which individual blades are selected for use in network load balancing inbound workloads. Finally, the method can include removing an individual blade from the rotation responsive to detecting a system condition in the individual blade indicative of an inability of the individual blade to support high availability in network load balancing. As such, the method further can include re-adding the individual blade to the rotation responsive to detecting a system condition in the individual blade indicative of an ability of the individual blade to support high availability in network load balancing.
1. A highly available network data processing system comprising:
a blade grouping chassis comprising a unified power supply and a switch module configured for network load balancing;
a plurality of server blades disposed in the chassis, each of the blades comprising a service processor and corresponding service processor device driver; and,
a management module coupled to the chassis over a communications link, the management module comprising program code enabled to detect system conditions among the plurality of server blades, to maintain a rotation of selected ones of the blades among which individual blades are selected for use in network load balancing inbound workloads, and to remove a blade from a network load balancing rotation responsive to detecting a system condition in the blade to support high availability in network load balancing.
2. The system of claim 1 , wherein the switch module is an Ethernet switch module.
3. The system of claim 1 , wherein the communications link is an interface to core logic (12C) bus.
4. The system of claim 1 , wherein the communications link is an interface to Ethernet.
5. The system of claim 1 , wherein the system condition comprises intense central processing unit (CPU) utilization in the blade.
6. The system of claim 1 , wherein the system condition comprises an executing application known to consume excessive central processing unit (CPU) cycles.
7. The system of claim 1 , wherein the system condition is a notification of an impending failure in a hardware component of the blade.
8. The system of claim 1 , further comprising an additional blade grouping chassis comprising a plurality of additional server blades, the management module comprising program code further enabled to remove a blade among the additional server blades from a network load balancing rotation responsive to detecting a system condition in the blade.