Around the Storage Block Blog
Find out about all things data storage from Around the Storage Block at HP Communities.

Are Fibre Channel disk drives dead?

  By Calvin Zito, @HPStorageGuy


It's time to get back to the topic of storage - as most of the regular readers of this blog know, I've been out for a few weeks dealing with my Mother's death.  Before I jump back into storage, I want to again thank everyone who was so kind to me.  I thank you very much - your kindness helped to get me through a difficult time.  Ok, so now onward....


There are a number of storage-focused groups on LinkedIn.  I noticed a conversation that was started on one of them discussing an article written by industry analyst George Crump.  The article itself was promoting a particular vendor but asked some good questions.  It covered a couple of different emerging technologies that could help reduce storage costs.  George makes the case that customers are buying more drives than they need from a capacity perspective to get the performance needed though I would argue that our EVA array minimizes this issue because of the built-in virtualization.  George suggests that SSD could help fill this gap but because it's still expensive compared to hard drive technology, it's not for everyone.  You can either read George's blog or read the conversation on LinkedIn to get the details of what he discussed.  Today, I wanted to highlight the answer that our own Tim Sheets gave.  Tim works in our Unified Storage Division managing portfolio strategy and marketing operations.  Here's what he had to say:


From a vantage point of a company that sells more HDD's than anyone else in the world, there are certainly changes on the horizon. SAS drives clearly offer an economic advantage and truth be told, the HDD manufacturers are trying to get away from FC drives because they are more expensive to produce. Now as for SSD's replacing FC drives... that is not going to happen wholesale at this time. The costs are too high, and there is still much to learn about the longer term reliability related to write cycles (SSD cells wear out at much lower cycles than traditional HDD's)  


While there is lots of work going on in the SSD market to manage write balancing, we cannot overlook the fact that there is much more work to be done to get to full enterprise reliability at volume economics. SLC is certainly more robust that MLC, but does not have the capacities of MLC. On the flip side, SLC has better write "wear-out" cycles than MLC. Either way, SSD costs are extremely high even when compared to FC HDDs. For now, I think we will primarily see SSD's in very high performance systems where customers are willing to pay a significant premium to get that performance. Texas Memory Systems and Fusion IO are good examples of this. However, we will not see mass adoption until the wear out rates become stable/proven and the costs drop dramatically. According to the SSD manufactures, that is still some time away for enterprise class drives.


Additionally, we will need to see the application developers modify their applications to take advantage of what SSD's can offer. Storage management/tuning tools will need to change in order to optimize storage environments using SSD's. And as Phillip from 3PAR notes above, storage tiering is critical here. (Editor's note: You'll have to go to the LinkedIn discussion to see what Phillip said). Today, solutions that are read intensive and need high IO rates are ideal for SSD. Write intensive applications are probably better off with rotating media for now if cost is a major factor.


All of this will take some time, as does adoption for most nascent technologies. SSD's will become more popular in certain applications, but the industry won't turn on a dime just yet. Longer term (3-5 years or more) they will begin to become more of a mainstream storage technology. For now, rotating media is here to say for a while.


Tim - I hope you don't mind me using your answer here for our readers.

Comments
Anonymous(anon) | ‎01-23-2010 04:49 PM

"George makes the case that customers are buying more drives than they need from a capacity perspective to get the performance needed though I would argue that our EVA array minimizes this issue because of the built-in virtualization."

Calvin, how is it that the built-in virtualization

"minimizes" the problem?

If anything, I would argue that the EVA's built-in virtualization takes away some of the IPOS, which means customers have to purchase a few extra drives to make up for that.

| ‎01-25-2010 06:50 AM

Hi Bit Racer,


We have a paper that talks about how the EVA utilizes all of the drives to improve performance - here it is: EVA Advantage: Eliminate Stranded Storage and Boost Performance white paper.  We even use spare drives for data which of course increases performance.


Lastly, there's not a "vritualization penalty" so customers don't need to buy extra drives.  Thanks for asking.

Anonymous(anon) | ‎02-14-2010 04:01 PM

Aren’t we mixing disk drive technology and array interconnect ?

For me the way I understand the question about the death of Fibre-Channel disk drives is primarily about the interconnect rather than the drive technology itself.

In the online mass-storage technologies, we do see an evolution between winchester technology and SSD technology and maybe in the near future developments like holographic storage, but isnt’t this is a different discussion than Fibre-Channel vs SCSI vs SAS which are interconnect technologies ? Some years ago there were marketing fights about what really made an enterprise storage array and FC drives were boasted as a key feature vs SCSI-based disk arrays. Now SAS and its fabric-like interconnect capabilities and device adressing way beyond the 16 of SCSI is definitely an alternative to FC for array back-end interconnect. That is also the promise of potential competitive cost advantage for systems manufacturers like HP, IBM and Dell who implement enterprise SAS drives in servers vs storage pure-play vendors like EMC and NetApp.

Fibre-Channel itself will not disappear be used for pure front-end fabric-to-array and server-to-fabric connectivity. Maybe with a changing hardware layer from fiber-optics to copper with developments around FCoE.

Or am I completely mistaken ?

Anonymous(anon) | ‎05-22-2010 04:21 AM

There is no difference between Enterprise SAS and FC harddisks. FC and SAS are both dual pathing capable and with SAS 6G you get a very fast connect. So there is no reason why storage manufacturer should use expensive FC drives. An EVA with SAS drives should have similar performance as with FC drives, because it's the same hardware. Only the connection to the backplane differs. I'm sure that lot's of manufactures changing there storage system to SAS hard drives. Maybe FC disks are dead... but FC himself not :smileyhappy: And that's the most important thing! I don't like iSCSI. :smileywink:

Kind regards...

Tschokko

Leave a Comment

We encourage you to share your comments on this post. Comments are moderated and will be reviewed
and posted as promptly as possible during regular business hours

To ensure your comment is published, be sure to follow the community guidelines.

Be sure to enter a unique name. You can't reuse a name that's already in use.
Be sure to enter a unique email address. You can't reuse an email address that's already in use.
Type the characters you see in the picture above.Type the words you hear.
Search
About the Author
25+ years experience around HP Storage. The go-to guy for news and views on all things storage..


Follow Us
The opinions expressed above are the personal opinions of the authors, not of HP. By using this site, you accept the Terms of Use and Rules of Participation