Oracle/Sun's HPC goes away …

You knew something like this could happen … but probably, like most others, you never thought it really would. Unfortunately, it appears to have happened.
The technologies in Oracle’s HPC quiver include SGE (aka GridEngine), Lustre, and several others.
Lustre really doesn’t have use cases outside of HPC storage. The SGE product largely doesn’t have use cases outside of HPC (though you could use it in some fairly creative ways).

We have customers with business dependencies on SGE. I need a read on this to see if SGE will continue, be abandoned, or revert to the community. Anyone who knows for certain, please, feel free to speak up.
Same with Lustre. This one is concerning me a bit. We have customers with business dependencies and requirements for it. I know that Whamcloud just started up, and we have been working with Clusterstor. We need a read on whether or not Lustre is going to survive this HPC stand-down. Possibly Oracle will invest in Whamcloud using Lustre IP? I dunno (BTW: thats a WAG on my part)

10 thoughts on “Oracle/Sun's HPC goes away …”

  1. @Chris
    Pretty sure Lustre has nothing to do with Exadata2. I think it might be SAMFS or similar in there. Likely Solaris based.
    Yeah, we have customers with dependencies upon SGE. Grab the tarballs now, and if you can, grab the latest cvs/svn dumps.

  2. @Chris
    Interesting on the use of OEL. Falconstor also uses OEL as their base. It is slightly different than RHEL, enough that you may need to rebuild some packages from the source RPMs.
    On SGE, yeah, going to grab it soon. Put it in either git or mercurial. I like the latter a little more than the former (easier for me to grok its usage/options). Will be hosting both on our sites.

  3. I just met with some Oracle reps and I asked them about this. They said that there were no layoffs in the HPC group. Who knows what is really going on?

  4. @Jess
    Some former Sun HPC folks I know were starting to look for new employment as of last week. I thought it was an isolated case, until I read the article.
    I wonder if its just sales though. On the product side, it doesn’t make much sense to keep spending money on developing something that you just removed the sales channel for, unless you are planning on bundling these in some manner and selling them to other companies. In this case, keeping the sales staff doesn’t make sense.

  5. @Chris
    Yeah … I have it on … er … pretty good authority as well … about the status.
    Now that OpenSolaris has shrugged off its (virtual) mortal coil, I wonder where the axe will fall next, and how quickly. I suspect it will be quick, and I am hard pressed to imagine some of my favorite tools surviving internally.
    I am guessing that SGE will be cut adrift (probably as an open source product), or Oracle is shopping it around (though its unlikely to get much for it). Lustre could be shopped, or Larry could use it as an investment in Clusterstor/Whamcloud.

  6. This morning I read an article on Channel Register about Oracle moving out of the HPC market. If this proves to be true, it is regrettable. I agree with the writer, Dan Olds, that HPC is the birthplace of IT innovation. Inevitably, by reducing the HPC market by one, we will witness a reduction in new technologies, processes and applications coming to the fore.
    However, the HPC market is now a mature market and the other players ??? vendors and integrators like OCF will fill the potential gap left.
    Today, organisations which invest in HPC systems invest in open systems. There is no vendor lock in and if an HPC vendor does lose its way, then migration to other platforms is more straightforward than it once was. We will continue to support our Sun customer base until the fat lady sings ???and if she does sing then our customers can be certain that OCF will make migration as painless as possible.

Comments are closed.