On Fri, Jan 23, 2009 at 7:52 AM, SimonT <[email protected]> wrote:
> it seems that
> there are arguments for file-based solution over memcached in terms of
> speed when the caching done is per node (as we do currently). For
> example:

File-based makes sense per-node if the majority of per-node cache is
not redundant.

In general, file-based makes sense if:
  * memory is at a premium
  * latency to other nodes is high
  * shared access to specific keys is easily partitioned to nodes
  * disk bandwidth dwarfs cache bandwidth

...
> Anybody care to comment? In a high concurrency situation, does
> memcached perform comparitively better? Are there any other factors we
> should be considering?

If you have a lot of writes, disk is going to bottleneck before memory/network.

Reply via email to