incubator-directmemory-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Akash Ashok (Commented) (JIRA)" <>
Subject [jira] [Commented] (DIRECTMEMORY-16) Create a non Singleton MemoryStorage alternative
Date Fri, 21 Oct 2011 19:00:33 GMT


Akash Ashok commented on DIRECTMEMORY-16:

Just Skimming thru the patch on a Highlevel. There are a few unused imports in
Those could be removed. This is more on the lines of code indentation. we need to decide on
the code formatting soon before our codebase ends up huge and we have to do major code formatting.

I am going through the functionality and shall raise concerns, if any, soon. Thanks  
> Create a non Singleton MemoryStorage alternative
> ------------------------------------------------
>                 Key: DIRECTMEMORY-16
>                 URL:
>             Project: Apache DirectMemory
>          Issue Type: Improvement
>            Reporter: Raffaele P. Guidi
>         Attachments: DIRECTMEMORY-16-patch.txt
> Create a non Singleton MemoryStorage alternative (it also a foundation to build multiple
caches for the same VM).
> See the discussion in the ML with Ashish (and also Simone Tripodi brought up the subject
before) for details
> > Ashish wrote:
> While creating a DirectMemory sample, I realized that there is no way
> to create multiple caches within same JVM.
> This is a fairly common use case, and we might need multiple caches
> with different configurations.
> On Sat, Oct 15, 2011 at 10:48 PM, Raffaele P. Guidi
> <> wrote:
> > We can consider the current implementation as the shortest path to a working
> > and efficient off-heap cache, and I really would like to keep things this
> > simple - no configuration at at all for the default configuration - but the
> > subject of multiple buckets has been already brought up and give users a
> > choice will be great.
> >
> > Now, I like the current Cache facade as it is (it is straightforward for
> > simple use cases) but MemoryStorage should be refactored - it is probably
> > wrong to have it as a singleton (the underlying OffHeapMemoryBuffer, the
> > basic building block, is not a singleton) - and we could write a
> > non-singleton implementation for the Cache as well - and here it is the
> > implementation for multiple caches (buckets?) in the same VM.
> >
> > Then the Cache singleton facade could be rafactored to use the non singleton
> > one as its foundation.

This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:!default.jspa
For more information on JIRA, see:


View raw message