Another couple orders of magnitude performance: RDMA with Infiniband or 10Gbe

[enealDC]

The managed flash software seems very interesting. I think that SSDs are the next evolution for storage. From a physics perspective, I'm not sure how much more can be done to improve mechanical disks. But right now, it's hardly affordable and because of the life time of the more affordable MLC variety, I have questions about reliability and MTBF.

I think that RDMA would be awesome, but RDMA is going to be for IB and right now, there are some innovative things that can be done to improve iSCSI performance over native IP/Ethernet.

http://www.ele.uri.edu/Research/hpcl/STICS/iCache.pdf

If you have access to this white paper, it talks about a different caching strategy to improve performance of iSCSI by 58 to 70ish percent. I'd like to see some innovation along these lines...

[To-M]

Thanks for this info - I passed it on to our development and management team as we are looking for ideas.

[beng]

G'day,
Thought I would "bump" this topic and pop Tiered storage for SSD/SAS/Sata Previous post
I agree with enealDC that SSD are the next evolution. The performance we are seeing with Databases on SLC in RAID1 is outstanding. But lets face it, it will be a long time before 1.5TB SATA are displaced by SSD.
But the amount of data that benefits from really high IOPS or transfer rates is actually quite small (in my experience). Most data on our clients networks is taken up by images and email(PST's).

So the ability to seamlessly shift frequently accessed data to faster storage and older/archived data to cheaper storage would be fantastic.

Cheers Ben

[enealDC]

Hey Beng. You are spot on! Understanding the IO requirements is a critical component of virtualization and storage management. I think not too many folks really spend time trying to understand the best blend of mix of the various types of storage when approaching these two topics. With the economy the way it is, there is less money to spend (unless you are a bank with bail out funds). Knowing where to put the high IO, but high cost storage and exactly how much to put is very important. Being able to leverage the different tiers and quality of storage is going to be critical at this junction while we wait for paradigm to shift and costs to drop. And you are right, 1.5 TB SATA drives are not going to be displaced anytime soon. But there is a danger with drive sizes getting so large. We have to constantly remind ourselves (and our customers) that delivering IO is the name of the game when it comes to SATA volumes and that larger is not necessarily *better*.