StorageMojo’s Global HQ is pulling up stakes and traveling to wilds of Palo Alto for the OpenStorage Summit 2010. I’m hoping to hear more about the future of ZFS and other storage stacks.
I’ll arrive Tuesday afternoon and will leave Thursday. If you’re in the neighborhood please stop by and say hello.
The StorageMojo take
The rapidly growing cloud storage market is slowly but surely changing the dynamics of the storage market. Private clouds will never be as flexible and cheap as public clouds, but the closer they get the more their economics will improve. It’s a game played at the margins.
That bodes well for enterprise adoption of open storage for private cloud infrastructure. Not everyone, but those with the scale and the moxie to make it work will find a significant competitive advantage.
Courteous comments welcome, of course.
Our experience with ZFS plus clueless drive array vendor is not good.
http://rockssandwater.blogspot.com/2010/10/massive-disk-failure-deja-vu.html
The future of ZFS very much depends on both good software support and good hardware support. Which might not be coming from Oracle.
We need good ZFS software + hardware support from a Tier 1 vendor. Nexenta, Coraid, and others are too small at this time, and might be sued out of existence by NetAPP.
@Rocky
This is a great blog entry, I read almost the whole blog! A lot of very useful and field information that made me think about some things when designing a datacenter solution for my customers.
Does anyone know what will happen to ZFS now when Oracle has got his hands on it?
BTW., any news regarding Oracle’s EMC purchase?
What is your take on this summit?
Anything new and interesting?
After so many top ZFS engineers left Oracle I guess some will land into new startup soon.
A variety of videos from the summit can be viewed here:
http://www.youtube.com/user/OpenStorageSummit
I would like to comment on a statement that you made, quoted “Private clouds will never be as flexible and cheap as public clouds, but the closer they get the more their economics will improve.”
On the surface, that might seems to be the case. Nevertheless, if you exam the pricing of AWS S3 for example, after a certain amount of data storage, the long term cost benefit starts to break down.
On the other hand, storage cloud federation technologies such as what Zettar has pioneered will make private storage clouds a very appealing proposition indeed. I outlined some further reasons in my blog. The similarity between SSDs and private storage clouds too.
Note: Dr. Fang is the founder of Zettar, which looks interesting because it allows enterprises to federate multiple cloud providers into one virtual file system, if I read it right. End note.
@Robin, I enjoyed our conversation the other day. While we are still at it, please let me share with you a bit of simple calculations:
The common perception among people is that AWS S3 (the most well-known and established public object cloud storage service) is inexpensive. Lets take a look of its most recent pricing;
A simple calculation for 50TB data stored over a year using high reliability tier is:
First TB:
$0.140/GB/month x 1024 GB/TB/month x 12 month/year = $1, 720/year
Next 49 TB:
$0.125/GB/month x 1024 GB/TB/month x 12 month/year = $75,264/year
Total: $1,720/year + $75,264/year = $76,984/year
The above just for storage. Remember to get data in and out of AWS S3, you have to carry out data transfer and send requests, and they add up *really* fast! So, double the price above to account for data transfers and requests, then add them together to get:
$230,952/year
I will leave your readers to ponder other cost/benefit factors 🙂
Best Regards,
Chin