There seems to be be more and more stuff appearing about Gluster; there’s a really nice article about Rolling Your Own Fail-Over SAN Cluster with Thin Provisioning, Deduplication and Compression using Ubuntu which just goes how far you can go with the DIY approach to building your own storage devices.
Please note that this article utilises iSCSI for it’s SAN connectivity but there’s no reason why you shouldn’t do a little more work and support FC as well and I daresay, that putting together FCoE is not beyond the realms of possibility.
I’d also suggest that people have a look at the 3.3beta stuff for reveal about what is coming down the line.
And I am certainly not suggesting that you should run your mission critical business applications on it but it really goes to show how far we’ve moved; premium features are beginning to turn-up in open-source systems.
A threat to the existing Storage Cabal? Not yet but for the more adventurous of you, there is a huge amount of potential.
Hi Martin
You don’t suggest people run this with their business critical application. Assuming you are referring to their GA code and not beta, what characteristics about Gluster trigger such a comment? Is it something Gluster specific or the usual disclaimer?
With multi core, lots of memory and SSDs, it has been all about the software for a while now. That said, I think the people running DIY storage, Gluster, GFS, Nexentastor, Falconstor, etc are going to be the same people who might use this in earnest. There seems to be an irrational (or just not quantified) fear of moving away from the “Storage Cabal” and their legions of support people. I would guess that the same people running Netapp and EMC today will be the same ones tomorrow no matter how much open storage platforms develop and evolve.
Frequently I think that I could do everything I do today with proprietary code, tomorrow with open source and save huge buckets of money and perhaps come out ahead functionality wise. OpenLDAP, Openstack, XEN Asterix etc etc.. but then I have the irrational fear noted above. I have the money to spend on the status quo, so why take the risk? Perhaps if I worked for a start-up the conversation will be different and Gluster could really be brought to bare under that scenario.
It is nice to see feature sets get pushed down to the lower tiers of the market, but I don’t think that alone will disrupt big name vendors… at least for a long while.
Just my 2c.
C
PS.. love the blog..