Method and system of session-aware load balancing

A network device coupled to a load balancing system, wherein the network device is to balance load of the load balancing system and is coupled to a plurality of servers, the network device comprising:

• a processor;
• and
• a non-transitory machine readable storage media coupled to the processor and storing load balancing software, which when executed by the processor, causes the processor to:
• assign a virtual Internet Protocol address (VIPA) to the network device, wherein the VIPA is shared with the plurality of servers,
• assign a set of next hop IP addresses in the network device, wherein each next hop IP address corresponds to a route to one of the plurality of servers,
• advertise the VIPA of the network device to network elements outside of the load balancing system,
• receive a packet with a destination IP address corresponds to the VIPA shared by the network device and the plurality of servers,
• determine that the packet with a source IP address is a first packet in a communication session with the source IP address,
• select one of the set of next hop IP addresses to forward the packet to reach a server of the plurality of servers based on a load balancing policy in response to the determination that the packet with the source IP address is the first packet in the communication session with the source IP address,
• forward the packet to the selected next hop IP address, and
• save the selection of the network device for forwarding subsequent packets of the communication session with the destination IP address.

A non-transitory machine-readable medium having instructions stored therein, which when executed by a processor, cause the processor to perform operations at a network device coupled to a load balancing system, wherein the network device is to balance load of the load balancing system and is coupled to a plurality of servers, the operations comprising:

• assigning a virtual Internet Protocol address (VIPA) to the network device, wherein the VIPA is shared with the plurality of servers;
• assigning a set of next hop IP addresses in the network device, wherein each next hop IP address corresponds to a route to one of the plurality of servers;
• advertising the VIPA of the network device to network elements outside of the load balancing system;
• receiving a packet with a destination IP address corresponds to the VIPA shared by the network device and the plurality of servers;
• determining that the packet with a source IP address is a first packet in a communication session with the source IP address;
• selecting, one of the set of next hop IP addresses to forward the packet to reach a server of the plurality of servers based on a load balancing policy in response to the determination that the packet with the source IP address is the first packet in the communication session with the source IP address;
• forwarding the packet to the selected next hop IP address;
• and
• saving the selection of the network device for forwarding subsequent packets of the communication session with the destination IP address.