Server load balancer (SLB) provides traffic load distribution functionality for this release of the switch. Load distribution can be performed at layer 2 or layer 3 of the protocol layer.
SLB at layer 2.
When used at layer 2, load balancing groups (LBGs) consist of switch port members. Load distribution is performed through the link aggregation group (LAG) hash function of the switch. Load distribution policies are based on the contents of the L2/3/4 packet header. Failover is supported at the switch port members level. Depending on the fail-over method chosen, a failed member within an LBG is replaced by another switch port or by a group of switch ports within the group. Connectivity between the switch port and the server is monitored by a health check mechanism. The access control list (ACL) capability is extended such that a LBG can be set as a target when a switch port accepts a packet. Level 2 SLB is supported only in this NEM.
SLB at layer 3.
When used at layer 3, LBGs consist of server members identified by layer 3 protocol (IP) addresses. Server members are next hop targets of a targeted equal cost multiple path (ECMP) route. The LBG is identified by a Virtual IP address prefix (VIP) which is used to represent a virtual server. From the clients' point of view, the virtual server is a highly available server entity with expandable resource. Load distribution is performed through the routing hash function of the switch. Load distribution policies are based on the contents of the L3/4 packet header. Failover is supported at the server member's level. A failed member within an LBG is replaced by another server member that is part of the same group. Health check is performed by the switch through an ICMP ping.
Refer to the Sun Ethernet Fabric Operating System SLB Administration Guide for more information about this feature.