Forum Discussion

Ethan_McG_23292's avatar
Ethan_McG_23292
Icon for Nimbostratus rankNimbostratus
May 11, 2016

LTM Web Acceleration Profile Configuration

My organization and I are new to BIG-IP. Everything has been working quite well, but I do have a question about Web Acceleration profiles. For LTM, the iApps I have used so far create one by default. This seemed to be working satisfactorily, until the programmers began complaining about it. The previous load balancer my company used apparently did not do anything like F5's Web Acceleration profiles do, so the old ADC wasn't caching any items in that regard.

 

I got with the programmers and told them we can clear the cache when we make changes, we can reduce the maximum age, and we can limit what items are cached, but they don't like that either. They want the LTM to recognize a document has changed without any cache clearing or anything like that. They want to be able to make changes to web pages (including to items the Web Acceleration profiles caches) and have all machines see those changes instantly and automatically. Of course, this obviously needs to happen, and the web pages we serve up change often, but the only way I see of doing that is clearing the cache on the Web Acceleration profile.

 

Is there a provision for the LTM to check the web server for items to get their last modified dates so that if an item is changed the new version will be served up automagically instead of the old version? Do Web Acceleration profiles provide enough improvement to justify their use in a medium sized environment? What do other companies regularly do?

 

application delivery
BIG-IP
caching
devops
iApps
LTM
web acceleration
web acceleration profiles
web pages

3 Replies

Sort By
  • Domai_23823's avatar
    Domai_23823
    Icon for Nimbostratus rankNimbostratus
    May 11, 2016
    Hello Ethan I work for a public utility company and we do use web acceleration...we enable caching pdf doc images and static pages. When ever we update the website I manually go and clear the ram cache. I am very curious on how ltm can detect the change and refreshes the cache automatically. Keep us posted pls.
  • Yann_Desmarest's avatar
    Yann_Desmarest
    Icon for Cirrus rankCirrus
    May 12, 2016

    Hello,

     

    The caching feature doesn't allow to autonomously update the cache when there is a change on the backend server. BIG-IP works with respect of rfc2616.

     

    What you can do is to create a shell script that regularly check if there is some updates on the backend server and if there is, the script can clear the ramcache for the webacceleration profile used in the Virtual Server.

     

    or this can be triggered from the backend server or the developer directly to instruct the bigip to clear the ramcache. You can use ssh or maybe iControl API.

     

  • jaikumar_f5's avatar
    jaikumar_f5
    Icon for MVP rankMVP
    May 13, 2016

    Well you have to help them understand that the sole purpose of Ram-cache is for static contents. Not for dynamic objects. If there is a dynamic object you dont insert Ram-Cache profile.

     

    Nevertheless, try modifying ramcache-aging-rate/ramcache-max-age to like 1800sec or less if you require. The default value is 3600 seconds.