topic Re: CSS loadbalancing based on parts of the URL in Application Networking

CSS loadbalancing based on parts of the URL

jfoerster — Thu, 07 Apr 2005 08:26:28 GMT

HI,

I've to do loadbalancing based on parts of the URL.

The URL will look like this:

www.myserver.com/a/b/c/d/x=12345/bdc/*.htm

If the x=<digits> is includes I've to use server1-3 1 if the x is not included I've to use server 4-6

I can right now think of two (2) methods:

1) url ="/x=*"

2) using a header-field group with field "request-line" and the operator contains.

Now my two questions:

- Is it enough to define in method 1 only "/x=*" or do I have to write down the whole path (/a/b/c/d/x=*) or is it possible to do url="*/x=*".

- Is there any other way to have this sort balancing done?

- Which method is best practise?

Thanks in advance for answering.

Kind Regards,

Joerg

Re: CSS loadbalancing based on parts of the URL

Gilles Dufour — Thu, 07 Apr 2005 08:56:26 GMT

Joerg,

1.you will need to specify the full path up to x=.

2.can't think of any other solution

3.no particular recommendation

Gilles.

Re: CSS loadbalancing based on parts of the URL

jfoerster — Thu, 07 Apr 2005 09:09:18 GMT

Hi Gilles,

thanks for your reply.

So both methods are possible. From my point of view method two is more flexible even if less specific. Is there any other draw back you can think of using method two, the one with the header group.

Kind regards,

Joerg

Re: CSS loadbalancing based on parts of the URL

drex_72 — Thu, 21 Apr 2005 16:32:18 GMT

We are using Header Field Groups and have not run into any unusual problems. It seems to work flawlessly.

We are using the "request-line contain" method to key off a certain word. Then if that word doesn't exist, it just defaults to the regular HTTP content rule.