Akka Clustering and Distributed Cache

1,723 views
Skip to first unread message

Vikas Hazrati

unread,
Oct 21, 2013, 9:54:55 AM10/21/13
to akka...@googlegroups.com

Hello,

We have a situation in which we would end up holding large amounts of data in the memory for fast querying. This data would be huge enough to mandate it rolling over to a few more JVMs on different nodes.

Would it be a good idea to

- Use a distributed cache such as JCS/EhCache/Memcache to store data
- Front end this with Akka cluster so that the request is made to a cluster
- Manage cache node addition / deletion with listener over Akka cluster join / leave


The idea is that the request would be made to the Akka cluster which would inturn compute / extract data from the distributed cache and service it back. Thoughts / Questions appreciated.

Regards | Vikas
www.knoldus.com

Ngoc Dao

unread,
Oct 21, 2013, 11:04:43 PM10/21/13
to akka...@googlegroups.com
You can have some ideas from project Cleakka:

I have not published it yet because the consistent hash cache client is still incomplete.

Vikas Hazrati

unread,
Oct 22, 2013, 3:27:58 AM10/22/13
to akka...@googlegroups.com
Thanks Ngoc, I would take a look there. It does seem interesting. Based on your research / work do you think it might be a good idea to bundle in something like a Hazelcast for caching behind an Akka cluster than spinning up your own cache? I am assuming that there would be certain pros with writing your own JSR107 compliant solution. Just checking if there could be a big miss in going with something like Hazelcast / Ehcache in the background.

Regards | Vikas
www.knoldus.com

Ngoc Dao

unread,
Oct 22, 2013, 3:55:45 AM10/22/13
to akka...@googlegroups.com
Actually in my Xitrum web framework, I'm including both Akka and Hazelcast (for cache):
http://ngocdaothanh.github.io/xitrum/

Hazelcast is very good.
If you want to use a distributed cache right now, you should use Hazelcast.
Hazelcast itself is distributed, it can run side by side with Akka, not behind Akka.

I'm creating Cleakka because Hazelcast has many other additional features like distributed pubsub, lock etc.
Most of the time I don't need them, so I think it's worth a try to create a small distributed cache library based on Akka.

Vikas Hazrati

unread,
Oct 22, 2013, 5:32:27 AM10/22/13
to akka...@googlegroups.com
Perfect, that sounds good. I checked out Xitrium. It looks impressive ! (thumbs up!)

Regards | Vikas
www.knoldus.com

Roland Kuhn

unread,
Oct 22, 2013, 5:46:37 AM10/22/13
to akka...@googlegroups.com
Hi Ngoc,

this sounds pretty interesting, fun stuff :-) I bookmarked it for diving into the code later, still wishing for that 36h-day-extension …

Regards,

Roland

--
>>>>>>>>>> Read the docs: http://akka.io/docs/
>>>>>>>>>> Check the FAQ: http://akka.io/faq/
>>>>>>>>>> Search the archives: https://groups.google.com/group/akka-user
---
You received this message because you are subscribed to the Google Groups "Akka User List" group.
To unsubscribe from this group and stop receiving emails from it, send an email to akka-user+...@googlegroups.com.
To post to this group, send email to akka...@googlegroups.com.
Visit this group at http://groups.google.com/group/akka-user.
For more options, visit https://groups.google.com/groups/opt_out.



Dr. Roland Kuhn
Akka Tech Lead
Typesafe – Reactive apps on the JVM.
twitter: @rolandkuhn


Reply all
Reply to author
Forward
0 new messages