Around the Storage Block - All Comments

re: Top Ten Reasons Why DAS Will Grow!

SEYAR — Sun, 07 Sep 2008 20:11:30 GMT

i disagree... DAS tends to die... and look at latest ISS servers releases: DL185 G5 = 12To intern storage ....

re: Green Storage #5 – Drive type and speed choices affect your power requirements

DGarrels — Wed, 03 Sep 2008 19:38:06 GMT

Can you provide more info? I've looked in the ProLiant Servers section of the configurator site mentioned above and all the products you mention are listed.

They're also in the Power Calculators section linked on the left.

Thanks, David Garrels

re: Green Storage #5 – Drive type and speed choices affect your power requirements

DINGOME Félix — Wed, 03 Sep 2008 13:10:54 GMT

I live and work in FRANCE.

In the new configurator site, according to the proliant servers section, i don't find ML350 G5 , ML370 G5, DL360 G5, DL380 G5.

Thank you for your answer.

re: Top Ten Reasons Why DAS Will Grow!

CalvinZ — Tue, 02 Sep 2008 15:13:24 GMT

Hi Mark - Thank you for your comment. You are correct that support for VMFS and therefore having shared LUN support and the ability to run VMotion is a limiter to SAS adoption in VMware environments. You will be pleased to know we are working on a fix for a late issue we caught with VMware that caused us to limit this support. You can expect to have support for the MSA 2000sa for Shared VMFS via an MSA 2000sa firmware update in about 30 Days."

Thanks again!

re: Top Ten Reasons Why DAS Will Grow!

Mark Wibaux — Mon, 01 Sep 2008 02:14:09 GMT

Certainly the "mini-SAN" functionality of the newer SAS based products is a great advantage and I can see it making more headway that the older SCSI based products that did the same thing (ie MSA500) but unfortunately until full support for the virtual environment is available in products like the MSA2000sa, adoption will be impossible.

By this I mean the fact that the MSA2000sa does not currently support shared VMFS which severely limits its usage with VMware.

re: Top Ten Reasons Why DAS Will Grow!

Leonel Fonseca — Fri, 29 Aug 2008 23:57:39 GMT

I agree with you: DAS adoption will soar in behalf price and virtualization.

re: EMC distortion about capacity efficiency

CalvinZ — Fri, 29 Aug 2008 23:25:15 GMT

Hi Stephen,

I'll try to find someone to make that comparison and explain it (I'm not technical enough to do it myself) but one of my pre-sales EVA experts did send me a screen shot of the calculation he did with our EVA capacity calculator tool that we have. He configued a 240 drive EVA using VRAID5 and had around 74% efficiency. Its not an apples to apples comparison but probably better reflects real-world than what Chuck did.

I also got an email from an EVA customer that saw what Chuck was saying and while I don't want to quote all of it, he said "who in their right mind would ever configure their array like this ... (guy)"? I'll omit the actual word he used as my intent isn't to offend Chuck. But this customer has 15 EVA's and couldn't believe what he was reading from Chuck.

I also got an email from another pre-sales storage rep; he mentioned that EMC is passing around a white paper to prospects giving an example of an EVA with 16 disk groups of 8 drives each with double sparing turned on. Good luck finding any EVA customer that has more than 5 disk groups and most probably have 1-3.

One last EVA point - regardless of the efficiency, every drive in an EVA contributes to array performance. Even if someone configured some huge capacity as spare, since the EVA stipes across the entire array (including spare space), every drive is busy doing I/O's. I wouldn't expect Chuck or EMC to make this point but I think I'll do a series of blogs in the coming weeks about virtualization in the EVA and what it does for our customers.

In the end, I'm glad that I work for a company like HP that takes the high road and doesn't use tactics like I continually hear coming from Hopkinton.

Thanks for posting Stephen!

re: EMC distortion about capacity efficiency

Stephen Foskett — Fri, 29 Aug 2008 21:57:42 GMT

Calvin,

Thanks so much for posing this detailed response. I am learning more about EVA every day!

Like you, I immediately saw some holes in the comparison, as documented on my blog:

blog.fosketts.net/.../grapples-and-tangelos-why-its-impossible-to-compare-fairly

I suggest, though, that you include a simple correct answer to Chuck's scenario: How much is the overhead of a similarly-configured EVA given the rest of his specs?

I'd love it if each vendor would go through the exercise themselves as a way to come up with something closer to apples-to-apples than EMC can do themselves.

Thanks,

Stephen

re: Enterprise Solid State Technology Today: Hype or Reality (Part 1)

CalvinZ — Wed, 27 Aug 2008 14:38:00 GMT

Great input Mike... thanks for jumping in!

Calvin

re: Enterprise Solid State Technology Today: Hype or Reality (Part 1)

MikeB — Wed, 27 Aug 2008 01:21:51 GMT

My take on FCoE - wait and see. Sure many industry "experts" say we will see everything converge tomorrow and that FC will be replaced in the future. As you noted the same was said of iSCSI and the market still favors FC today (based on market #’s). The reality with iSCSI was/is that there is a place and time for it. My money says we are years (3+) before FCoE will be a major player. There are three key things that must occur for FC to be repalced.

1) FCoE products are readily available from a variety of vendors

2) The equipment and inteconnects are cost effective (10GbE is not real cheap today, but many are working to make it so)

3) Storage Admins grow comfortable with the technology after it has proven itself to be reliable and stable.

Even if all that proves out you still have the 800lb elephant sitting in the corner....politics.

In many organizations there are two networks today; 1)The FC network infrastructure typically run by the Storage Admins and 2) The IP network run by a separate group. The big draw with FCoE is that you can converge the two networks and the associated cabling. My gut feeling is that many SA's are not going to give that up to the IP guys just yet. Maybe i'm wrong, only time will tell. Some say that the potential cost savings from the cabling will cause many business to make this a reality, but again that was a big draw with iSCSI. The cost difference of copper vs. Fiber has not proven to be a real deciding factor,except in the SMB space. I’m sure someone will come along and say “my xyz company of 10trillion servers all run on iSCSI”, bu the sales number’s say different. In the end despite the large cost difference you have two networks today and very well may have two networks tomorrow.

On the subject of SSD drives - fantastic technology.....for the right purpose. The cost per GB is very high today and will likely remain so for the next several years, I am guessing. On top of that it, it is too new of a technology that will need to prove itself to in the industry. Many have some exposure to the early desktop SSD's which are not all that impressive. Enterprise SSD is different, but the perception of high failure rates from the desktop SSD is already out there.

Personally I would like to see an array with a hybrid approach. In my dream world the array controllers would be smart enough to watch i/o patterns on LUN’s and dynamically transition those with higher iop requirements over to a SSD based disk pool. Kind of like a layer of secondary controller cache if you will. This should not only be dynamic, but perhaps even allow the admin to set some sort of Quality of Service policy on a given LUN. If no QoS policy is set in the array then it dynamically manages the LUN’s. If there is a policy set then it will apply some sort of preference based on those settings to the defined LUN’s and run them from the SSD drive pool. Perhaps the array management software could be smart enough to even make recommendations on which LUN’s should have a higher priority set on them. This recommendation to the admin would be based on past performance data that it would collect and analyze on a continual basis.

re: Solid State Disks – Sorry EMC, Fibre Channel Disks aren’t dead yet! (Part 2)

CalvinZ — Tue, 26 Aug 2008 19:46:48 GMT

Hi Sergio,

Jim Hankins is on vacation this week so I wanted to respond to your comment.

If you set a Fibre Channel (FC) disk drive on a table next to a solid state drive (SSD), the spec sheets will have performance like what you stated. However, that isn't real world. Drives work in an environment that have other factors that will impact the "real-world" performance.

Another aspect that is often ignored is reliability. The customers that use FC drives are slow to adopt new technologies. Even when SSD drives are at price parity with FC drives (which is years away, not days), customers will move slowly to SSD's only when they are satisfied that they won't impact the availability and performance of their applications.

HP is not an HDD vendor and has nothing to gain by trying to impede the ramp of SSD . I think we're trying to stem the artificial exurbance created by some silly predictions about how quickly SSD will replace Fibre Channel drives. As our customers start to move to solid state technology, HP will be there not with hype but with tested, trusted solutions that cut across our servers and storage.

Finally, my colleague Jieming Zhu from our StorageWorks CTO team just posted a blog on the topic of Solid State Technology that addresses some of your comments. Take a look at it here: http://www.communities.hp.com/online/blogs/datastorage/archive/2008/08/26/enterprise-solid-state-technology-today-hype-or-reality-part-1.aspx

Thanks for your thoughts,

Calvin

re: Where is EMC's Maui? Did it get whacked?

CalvinZ — Tue, 26 Aug 2008 19:08:34 GMT

Hi Mike - what you've said about Hulk and Maui is spot on to what we know as well. We've actually heard from customers who are testing Hulk/Maui and have come to HP to discuss ExDS (after testing).

As to where else you might see the dense drive shelves in ExDS, we can't share those details now but rest assured, our team knows those details and we'll talk about it more in the near future. Customers that need to know can talk to their HP representative to have those details shared under non-disclosure.

Thanks for your insight!

Calvin

re: Deduplication, online storage, and cannibals

Larry Freeman — Tue, 26 Aug 2008 00:26:40 GMT

Hi Warren, here is a little more background on NetApp dedulpication: We've delivered dedupe on over 13,000 NetApp FAS systems in a little over a year. It has surprised us all to see just how fast this feature has taken off. Maybe even more surprising is how fast our customers have gravitated towards primary storage apps. VMware VM's and File Services (aka /home dirs) are leading this charge. We estimate over 50% of our users are currently dedupe-ing primary storage.

You are correct when you say that we caution users to proceed slowly with high performance apps. Deduplication, like any other system process, does consume resources. But what our customers repeatedly tell us is they see no difference in system performance before, during, and after deduplication.

With the recent release of deduplication on V-Series, we look forward to hearing more dedupe success stories from our HP users.

Larry Freeman

re: Where is EMC's Maui? Did it get whacked?

MikeB — Mon, 25 Aug 2008 02:12:24 GMT

Hi Calvin here is what I've heard/read over the last 10 months or so.

Rumors are that Hulk is the piece they are developing with help from Ibrix and is now shipping very very quietly as the InfiniFlex 10000. (see: blocksandfiles.com/.../5245)

Regarding Maui, several online mentions of it indicate that it is part of or based on the OceanStore project (see: www.wikibon.org/The_Times_they_are_A-Changin%E2%80%99:_Architecting_cloud_computing). Here is another link for info on the OceanStore project (bnrg.eecs.berkeley.edu/.../OceanStore-endeavour012000.ppt)

Now here is a question for you and HP. Any chance we will see the disk shelf used in the XDS in other arrays, say the EVA or the XP? Granted this is SAS technology, but if you could utilize dual port SAS drives in that shelf it might be viable.

just a thought.

MikeB.

re: Solid State Disks – Sorry EMC, Fibre Channel Disks aren’t dead yet! (Part 2)

Sergio Soccal — Sun, 24 Aug 2008 14:21:21 GMT

I´m disagree with you.

I think that the cost os SSD is less that standar disk in some case.

A standar SSD give you 10.000 IOPS(and up to 20.000 IOPS), a standar 15Krpm 180 IOPS.

So, you will need more that 50 HDD 15krpm to compare IOPS performance .

A SSD 64 GB cost U$S 1400 ; 50x72GB(15Krpm) is more expensive. and probably you need a storage system to work on it.

So, I believe that there is a mentality problem (to accept the new tecnology) and a market of HDD that manufacturer don't like to change.

Regards,

Sergio soccal

ssoccal@gmail.com