[infinispan-dev] Compressing Marshaller Wrapper

["Trustin Lee (이희승)"]

Well, here's my first post hoping it's not so nonsense. :)

I wonder if the cache entries are stored in a compressed form.  For
example, a client outside the cloud will prefer the compressed form
while a client inside the cloud will prefer the uncompressed form.  For
maximum throughput, the cache store could store both compressed and
uncompressed versions?

Manik Surtani wrote:
> I think it makes more sense in the marshaller, simply because the compressed stream is useful elsewhere as well - such as the network, etc.
> 
> On 25 Feb 2010, at 13:40, Philippe Van Dyck wrote:
> 
>> Sounds like a better idea, closer to the real purpose since the Marshaller is used in other places... but how do I configure this ?
>>
>> Anyway, the marshaller is done, I am testing it, but the code is very easy to adapt to a cachestore...
>>
>> phil
>>
>>
>> On Thu, Feb 25, 2010 at 2:33 PM, Mircea Markus <mircea.markus at jboss.com> wrote:
>> Can't we handle this through a delegating CacheStore, that would compress the data and pass it on to the actual store for further processing (the actual store will take then the binary data and marshall it (ni marshalling I assume as it receives byte[]).?
>>
>> On 25 Feb 2010, at 14:06, Philippe Van Dyck wrote:
>>
>>> Done
>>>
>>> https://jira.jboss.org/jira/browse/ISPN-357
>>>
>>> Phil
>>>
>>> On Thu, Feb 25, 2010 at 12:22 PM, Manik Surtani <manik at jboss.org> wrote:
>>>
>>> On 25 Feb 2010, at 11:17, Philippe Van Dyck wrote:
>>>
>>>> Thanks Manik but it is actually related to the compression of the stored and (https) transferred ones on the S3 side.
>>>> The S3 bill is 95% lower here... and https transfer time is lower too (you pay for booth).
>>> Makes sense.  Create a JIRA for this, I'm sure such a feature would be generically useful to others as well - if you're happy to contribute your work back in.  :)
>>>
>>>> The Marshaller interface is exposing a lot of methods and it looks like the call each other... 
>>> The original Marshaller interface is legacy, from JBoss Cache, which was built as an extension to JGroups' Marshaller interface.  It may look a little confusing as a result, but have a look at the AbstractMarshaller which should help simplify things.  Further, look at the VersionAwareMarshaller, which is the entry point into the marshalling system.
>>>
>>> HTH
>>> Manik
>>>
>>>
>>>
>>>> phil 
>>>>
>>>> On Thu, Feb 25, 2010 at 12:11 PM, Manik Surtani <manik at jboss.org> wrote:
>>>> If it is compression for transmission you are concerned about, you can do this by adding the JGroups COMPRESS protocol to your JGroups cfg.
>>>>
>>>> On 25 Feb 2010, at 11:02, philippe van dyck wrote:
>>>>
>>>>> Hi All,
>>>>>
>>>>> Currently, I compress all data before sending it to the cache. Once compressed, I gain 95% of the JSonized qi4j objects.
>>>>>
>>>>> I did some profiling during the load tests and compression is taking roughly 80% of the cpu time.
>>>>> So I would like to compress only the data sent to the store, not in memory.
>>>>>
>>>>> Looks like the Marshaller is my friend here, and I plan to write a compressing wrapper around it.
>>>>>
>>>>> Now, when I look at it, I see two ways to wrap the compression process.
>>>>>
>>>>> One way is with the ObjectInput / ObjectOutput but I am bothered by the reentrant flag.
>>>>> The other is the ByteBuffer stuff, no concurrency problem here, but it looks like more work.
>>>>>
>>>>> WDYT ?
>>>>>
>>>>> Cheers,
>>>>>
>>>>> phil
>>>>> _______________________________________________
>>>>> infinispan-dev mailing list
>>>>> infinispan-dev at lists.jboss.org
>>>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>> --
>>>> Manik Surtani
>>>> manik at jboss.org
>>>> Lead, Infinispan
>>>> Lead, JBoss Cache
>>>> http://www.infinispan.org
>>>> http://www.jbosscache.org
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> infinispan-dev mailing list
>>>> infinispan-dev at lists.jboss.org
>>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>>
>>>> _______________________________________________
>>>> infinispan-dev mailing list
>>>> infinispan-dev at lists.jboss.org
>>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>
>>> --
>>> Manik Surtani
>>> manik at jboss.org
>>> Lead, Infinispan
>>> Lead, JBoss Cache
>>> http://www.infinispan.org
>>> http://www.jbosscache.org
>>>
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> infinispan-dev mailing list
>>> infinispan-dev at lists.jboss.org
>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>
>>> _______________________________________________
>>> infinispan-dev mailing list
>>> infinispan-dev at lists.jboss.org
>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>
>> _______________________________________________
>> infinispan-dev mailing list
>> infinispan-dev at lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>
>> _______________________________________________
>> infinispan-dev mailing list
>> infinispan-dev at lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
> 
> --
> Manik Surtani
> manik at jboss.org
> Lead, Infinispan
> Lead, JBoss Cache
> http://www.infinispan.org
> http://www.jbosscache.org
> 
> 
> 
> 
> 
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> infinispan-dev mailing list
> infinispan-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/infinispan-dev

-- 
what we call human nature in actuality is human habit
http://gleamynode.net/


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 260 bytes
Desc: OpenPGP digital signature
Url : http://lists.jboss.org/pipermail/infinispan-dev/attachments/20100226/b37eae4f/attachment-0002.bin