On Thu, May 5, 2011 at 1:02 AM, Josh Berkus <josh@agliodbs.com> wrote:
>
>> FWIW, EnterpriseDB's "InfiniCache" provides the same caching benefit. The way that works is when PG goes to evict a
pagefrom shared buffers that page gets compressed and stuffed into a memcache cluster. When PG determines that a given
pageisn't in shared buffers it will then check that memcache cluster before reading the page from disk. This allows you
tocache amounts of data that far exceed the amount of memory you could put in a physical server.
>
> So memcached basically replaces the filesystem?
No, it sits in between shared buffers and the filesystem, effectively
providing an additional layer of extremely large, compressed cache.
Even on a single server there can be benefits over larger shared
buffers due to the compression.
> That sounds cool, but I'm wondering if it's actually a performance
> speedup. Seems like it would only be a benefit for single-row lookups;
> any large reads would be a mess.
Depends on the database and the workload - if you can fit your entire
100GB database in cache, and your workload is read intensive then the
speedups are potentially huge (I've seen benchmarks showing 20x+).
Write intensive workloads, less so, similarly if the working set is
far larger than your cache size.
--
Dave Page
Blog: http://pgsnake.blogspot.com
Twitter: @pgsnake
EnterpriseDB UK: http://www.enterprisedb.com
The Enterprise PostgreSQL Company