Choosing the optimum WCCP request distribution method

Influencing the number of requests redirected to each proxy-cache server

Part of a WCCP service group definition is defining the method the router will use to distribute requests across proxy-cache servers in the same service group that the router belongs to. (The request distribution method is also referred to as the hashing method.)

The WCCP request distribution method is designed to minimize duplication of objects between proxy-cache servers. Requests for objects from the same origin server are sent to the same proxy-cache server. When a proxy-cache server that was unavailable becomes available again, it takes over responsibility for handling requests for the same origin servers it handled previously.

Typically, you want to achieve an even distribution of requests across your appliances.

The following table shows the basic criteria available for request distribution and how they apply to the modes in which you can run an appliance.

For a NetCache running as...	Recommended distribution method (hashing)
A forward proxy	Destination IP address
An accelerator (also referred to as a reverse proxy)	One or both: Source IP address Source port

For a NetCache running as...

Recommended distribution method (hashing)

A forward proxy

Destination IP address

An accelerator (also referred to as a reverse proxy)

One or both:

Source IP address
Source port

Forward proxy recommendations

When users send requests to arbitrary Internet servers (that is, when users send requests to many destinations), distribution on destination IP address provides the most equal distribution over multiple appliances that are running as forward proxies.

Accelerator recommendations

If you are deploying the appliance as an accelerator, distributing requests based on source IP address provides the best distribution. If your appliance is accelerating a single Web server and you set the distribution method to destination IP address, most (or all) requests will be sent to one appliance because the destination IP address is the same for all requests. Even if your appliance is accelerating a few origin servers, the variation in destination IP addresses is limited.

For an accelerator, adding the criterion of hashing on source port, as well as source IP address, results in requests from the same client being distributed over multiple appliances. This distribution is created because the source port numbers are dynamically assigned by the operating system on the client and are in effect only for the duration of the connection.