Forum Discussion

MOHIT_125417's avatar
MOHIT_125417
Icon for Altostratus rankAltostratus
Jun 02, 2014

Difference between Load balancing (Node) & Load Balancing (Member)

Hello Experts,

 

Please share the differences between the Load balancing (Node) & Load Balancing (Member)

 

9 Replies

  • Hi Mohit, When you balance against the node the LTM looks at all connections on the node even active connections that are not driven through the VIP-Pool connections. When you balancing against a member of pool it only looks at connections that are created and maintained through the VIP-Pool, ignoring other active connections.

     

    I hope this helps,

     

    -=Bhattman=-

     

  • So which one is best one & mostly used....

     

    i think it is more about what you want to achieve.

     

  • So which one is best one & mostly used....

     

    i think it is more about what you want to achieve.

     

  • As Nitass mentioned it's about which one you think you are going to need. If you are unsure then try one and then move to another and observe. See which one works for you.

     

    I hope this helps, -=Bhattman=-

     

  • giltjr's avatar
    giltjr
    Icon for Nimbostratus rankNimbostratus

    I would think choosing member vs. node would also depend on which load balancing method you are using.

     

  • So basically, Node load balancing is based on actual server's resource utilisation; Member load balancing is based on actual VS utilisation - incoming connection for that VS.

     

    Is my thinking right here?

     

  • wlopez's avatar
    wlopez
    Icon for Cirrocumulus rankCirrocumulus

    Generally speaking, for most of the load balancing methods like ratio, least connections, observed, and predictive, the difference between member versus node come down to how many pool members are associated to each node. The member options make their decision based on traffic to the specific pool members (server-IP:port). The node options do the same but they are based on overall traffic to the node (server-IP). If a given node (server) only has one pool member associated to it, then you shouldn't see any difference between them. However, if a node has different pool members associated to it, either within the same pool or different ones, then you will get different results between them. On this type of scenario, the decision between the member or node variants of the load balancing methods really depends on what you want to accomplish.

     

    Resource utilization on the servers (CPU, memory, etc.) is only used on certain load balancing methods like Dynamic Ratio.

     

    Hope this helps.