Predictive and Observed Load Balancing Methods - How long does it observe traffic?
Just got a quick question. We're having a discussion at work regarding the Predictive and Observed load balancing methods and how long they actually analyze the traffic flow.
According to SOL6406 it observes traffic "every second". It states the following:
" Predictive mode
The Predictive mode dynamic load balancing algorithm ranks server performance over time and prefers pool members that exhibit an improvement in performance over those that exhibit a decline. The BIG-IP system calculates a dynamic ratio value based on the number of L4 connections to each pool member over time. Every second, the BIG-IP system observes the number of L4 connections to each pool member, compares it to the previous connection counts to determine the relative performance trend, and assigns a ratio value to each pool member. If the connection count is unchanged, the system does not change the ratio value. If the connection count is increasing for a pool member, the system reduces the ratio value for that pool member. If the connection counts are decreasing on a particular pool member, the system increases its ratio value. When a new connection is requested, Predictive mode load balances the connections based on the ratio values that the system has assigned to each pool member, preferring the pool member with the greatest ratio value."
However, F5 university states that the BIG-IP device analyze traffic a couple of seconds back, not just the previous second.
Which one is correct?