Coyote Point Systems Equalizer GX Series, Administration Manual

Page: 388 / 866

Working with Clusters and Match Rules
Match Rule Order
When you add more than one match rule to a cluster, the order in which the match rules are
processed is important to system performance. Since processing a match rule requires system
CPU and memory, the most efficient way of ordering match rules is from the most common case
to the least common case. In this way, you ensure that the greatest number of client connections
possible will process the first match rule and, if it matches the request, stop processing match
rules for that request.
In other words, the goal is to load balance the highest possible number of requests according to
the settings in the first match rule, which has the effect of reducing to a minimum the amount of
match rule processing required for requests to that cluster.
This is best illustrated by an example. Let’s say you want to construct a set of match rules that
achieves these goals:
l
Direct all requests whose URL contains one of two specific directories to specific server
pools. Assume these two directories are.../support and.../engineering.
l
Of the two directories above, we expect more requests to contain.../support.
l
Load balance requests whose URL does not contain a directory across all servers.
l
We want to process requests that do not contain a directory the fastest, since we expect
that 75% of requests to this cluster will NOT contain a directory in the URL.
The set of match rules that achieves this, their order, and how the match rules are evaluated, is
described in the following figure.
388 Copyright © 2014 Coyote Point Systems, A Subsidiary of Fortinet, Inc.