For more information regarding the security incident at F5, the actions we are taking to address it, and our ongoing efforts to protect our customers, click here.

Forum Discussion

Patrik_Jonsson's avatar
Patrik_Jonsson
Icon for MVP rankMVP
Apr 02, 2014

Tricky (?) questions about the LTM Web Acceleration

Hello! This is going to be a large body of text, so please have patience. Our developers has asked some excellent detailed questions about the LTM caching which I have been unable to fully answer...
application delivery
design
LTM
performance
ram cache
TMSH
IheartF5_45022's avatar
IheartF5_45022
Icon for Nacreous rankNacreous
Apr 02, 2014

Here's some information I have;-

 

In 10.1 and beyond, ramcache has a "striped" cache, meaning that rather than each TMM having an independent cache, the TMMs collaborate to provide a unified cache with each TMM having a piece or "stripe" of the cache. The default split is 25% inter-stripe, 75% own.

 

http://support.f5.com/kb/en-us/solutions/public/11000/900/sol11990.html The RAM Cache feature introduces the new Ramcache.InterStripeCacheMemoryPercent database variable. This variable, set at 25 percent by default, defines what percent of the RAM Cache per TMM is reserved for local versions of cached objects.

 

The "remote hits" and "remote misses" statistics are used when a tmm receives a request for a cacheable object "owned" by another tmm. The first tmm may have a local copy of the object in its inter-strip cache - if it does not, it has to ask the other- or "remote" (in RAM Cache lingo) tmm. Hits and misses are self-explanatory.

 

The only use for this information I have found was to understand why eviction counts start when you reach 75% of your profile-assigned RAM Cache (as other 25% inter-stripe), although I'm sure F5 support have a use for it.