Memcached vs. HW Load Balancer with sticky sessions

That's it, I did some research on when (and when not) to use Memcached. I am familiar with distributed caching in terms of its main goals. Using Memcacached / like makes sense to me if you have multiple servers ... as this will give you a central virtual repository to retrieve your cached data.

However, if you have a hardware load balancer (say, F5 BigIP) that can do sticky sessions - does it have a distributed cache as profitable? AFAIK, it seems the only thing you do in this case is to ensure that you are not using your server server for cache. Are there any other benefits to using memcached in an environment where you already have HW-based load balancers that use sticky sessions?

As far as I know, having sticky sessions is not expensive in terms of performance. Obviously, I could be wrong.

+3
source share
3 answers

, memcached. , memcached .

, : . memcached , .

, , . , 20 . , , A. , memcached.

, , , , , B. B memcached, A, .

: , memcached , , , . , . , , , .

+3

cookie , , . .

memcached , - memcached , , .

+3

However, sticky sessions can potentially reduce the effectiveness of your load balancing. We saw this on our production site, where on one server (just by chance) there were many sessions with low activity, while the other server would have many sessions with high activity. Sessions balance, but not necessarily load or traffic.

This is not a complete answer, but it is an important factor to consider.

+2
source

Source: https://habr.com/ru/post/1769225/


All Articles