is memcached scalable and ...?
gwhalin at meetup.com
Wed Jul 26 12:03:18 UTC 2006
From the javadocs,
Enable storing compressed data, provided it meets the
This means it will compress the serialized data if the data size is
greater than the threshold size.
setCompressThreshold takes a long value which is a byte threshold. If
the serialized data is larger than this value and setCompressEnable is
set to true, the data will be gzip compressed as it goes into the store.
setWeights() applies a weighting to the list of servers in order to make
the chance of that server being chosen higher or lower compared to other
servers in the pool. There is an example of how to use this in the
In general, the defaults for most settings make sense in the client pool
settings. There are examples in the javadocs on how to set up a pool.
Not sure how to answer your second question. What does scalable mean?
The java client uses a connection pool and can be tuned to be as
scalable as you need I think. We use it in a pretty heavily loaded
environment, and it is far from being a bottleneck.
Also not sure how to answer your third question. Feel free to benchmark.
> I am a neebie. I am using the java version of memcached and I have some
> questions regarding it. Pls answer even though they are trivial :-
> 1. Where can i get proper information abt what the functions like
> MemcachedClient.compressionThreshold(), compressionEnable()... becoz
> javadocs dont seem to explain them. I am using them in my code but i
> dont know why i m using them?
> Similarly, for SocketIOPool.setWeights(), setNagle() , etc.. What are
> the basic parameters that need to be set for making a pool?
> 2. Is the java version scalable as the 'C' module one?
> 3. How fast is the 'C' module compared to the java version?
> I am still experimenting the more on it. Your response will be a great
> help. I would really appreciate it.
> Groups are talking. We´re listening. Check out the handy changes to
> Yahoo! Groups.
More information about the memcached