First days at Sun, Honeycomb opensourced

So now that MySQL is part of Sun, I can be real Sun technologies fanboy :)
So besides all the ‘it is done’ news, I noticed on Sun’s site that they’ve (oh, we’ve) opensourced Honeycomb. Thats the technology I’ve been drooling about quite a few years ago, and would’ve always wanted to see open. Now it is.
Honeycomb (aka StorageTek 5800) is distributed content-addressable object storage system (CAS), that allows putting in objects, and their metadata, have it indexed, and represent virtual directories based on that metadata. Accessible via WebDAV, API, CLI, etc… Previous solutions used to be expensive, now put it on top on Thumpers, and it is easy to manage storage archive. Now all I need is to find if it really is suitable for our needs, and start playing more with it.
Oh, and MySQL storage engine access to it would be quite cool too, it would mean Brian’s archiving engine is history :)

Update: Damn, apparently only emulators and clients released so far, and according to Sun folks, the HADB bits might be complicated to release – though the other stack might appear public and open soonish.

6 thoughts on “First days at Sun, Honeycomb opensourced”

  1. Hi!

    They solve different problems. I would like one of these object stores to have a SQL front end though…

    Looked at Hypertable?

    Cheers,
    -Brian

  2. Hi!

    Another thought… it does not matter if they open source this or not. If no community forms around it to push it forward… who cares.

    Cheers,
    -Brian

  3. Yup, looked a bit at HT, trying to figure out what is the efficient storage backend for it, that would be long-lasting and maintainable.

  4. Are you a fanboy because Sun gave all MySQL employees their own Thumper? That might not be such a good thing in the long run if they don’t pay your electricity bill. :)

  5. I happen to work at the similar field (content management systems). Here we use homegrown repository which is integrated with Lucene and database of your choice.

    Interestingly though, there’s JSR 170 (Content Repository for JavaTM technology API) which addresses similar problem (http://jcp.org/en/jsr/detail?id=170) and is supposed to be the standard. If I understand correctly, Honeycomb only stores data which doesn’t change; whereas JSR 170 works with versioned data. Can anyone with a better knowledge of Honeycomb comment on any more ideological differences between these two projects?

Comments are closed.