<   Previous Post  Information Infrastr...
Live Webcast : Incre...  Next Post:   >

Comments (3)
  • Add a Comment
  • Edit
  • More Actions v
  • Quarantine this Entry

1 localhost commented Trackback

BarryW-<div>&nbsp;</div> I try hard to stick to the facts in my blog, and I respect that you usually do as well. But this post is littlered with several falsehoods - hopefully you will correct your post to reflect reality:<div>&nbsp;</div> 1) EMC didn't "jump" at STEC in response to your blog. I don't know who your "good authroitiy within EMC" is, but the relationship with STEC began long before, in 2006 in fact. Given that close to 99% of EMC employees weren't aware of our work on flash drives, I'm not surprised that there might be an employee or two that doesn't know the actual timeline. But I do know it - intimately - since I was involved from the beginning (get a copy of last fall's EMC NOW for the details). This assertion is false.<div>&nbsp;</div> 2) EMC did in fact ship EFDs as GA product in March of 2008 (not May as you claim) - just as we had announced we would on Jan 14th. You clearly have no facts to back up your assertion here either, because I know exactly the first drives shipped.<div>&nbsp;</div> 3) CLARiiON did indeed start shipping EFDs in Q4. The fact you couldn't get any probably means you didn't order them early enough - customer demand was (and remains) extremely high.<div>&nbsp;</div> 4) All this talk about "exclusive" does not explain why it has taken IBM since September of 2007 (your blog post) to February of 2009 to ANNOUNCE the intent to ship STEC EFDs, nor the fact that you've announced you'll not start shipping them UNTIL APRIL 2009, and then on NEW DS8K systems only - no upgrades to existing purchases! <div>&nbsp;</div> Those are the facts, and I know them first-hand...I'd appreciate if you'd have the honor to correct (or retract) your misleading post.

2 localhost commented Trackback

BarryB,<div>&nbsp;</div> I also try hard to stick to the facts - I have corrected the dates as requested. The information I was given regarding the exclusivity deal with STEC was with regards to my blog - not that you only started working with them as a result. I have no doubt that you were working with them earlier. However, as this could be classed as heresay - I have retracted this statement as above.<div>&nbsp;</div> CX however I have clarified, we tried several times in Q4 to order these drives and were told there was no way to input the order - now this may well be a partner ordering issue or other, but this is fact as we saw it. I acknowledge however your statement that they shipped in Q4, however we could not order. <div>&nbsp;</div> As for the time between my post and now, I am the SVC dude remember, and I say what I see from my perspective - I'm not a marketing person like yourself, and I don't have access to all the product plans / development information regarding DS8000. With SVC we have been looking how best we can serve our customers, and as I said there is no prize for being first, only delivering integrated solutions that provide what our customers needs. I'm not knocking what EMC have done, they got them out earlier than us, no question.<div>&nbsp;</div> However, since we are correcting things, I would appreciate if you would remove the incorrect performance information you have on your post regarding DS8000 DA adapter pairs. The adapters are capable of A LOT more IO than you are quoting, and several members of the DA adapter team have asked I pick you up on this. Since there was no qualification of the testing performed, what drive types, firmware levels, generation of hardware, this is a misleading figure and I would appreciate if you would remove this from your post. Since you have no qualification or comparative data for this testing, its as useful as a Hitachi math statement. I think everyone out here would rather see EMC performance data.<div>&nbsp;</div>

3 localhost commented Trackback

Thanks for setting the record straight.

 
You still seem overly wrapped up with the notion of an exclusive, as if that makes any real difference. Are you insisting that this is the reason for IBM's long latency in bringing an EFD solution to market?
 
As to the DS8K back-end performance: after re-reading both IBM's competitive positioning papers and EMC's, and after conversations directly with EMC's performance engineers (folks that have your role here in EMC and who have actually run tests on the DS8K), I'm going to stick with my numbers.
 
All the data I can find evidences that the DS8K Turbo RRM back-end IOPS tops out at between 13.6 and 14.2 4KB block random IOPS for a single DA pair and scales fairly linearly until the 7th pair or so. I have also seen IBM-labelled non-confidential materials that say 8DA pairs on a DS8K turbo tops out at around 90K RRM 4KB IOPS.
 
Of course, these (non-confidential) IBM documents were published with many disclaimers, including:
 
<quote>Any performance data contained in this document was determined in a controlled environment. Actual results may vary significantly and are dependent on many factors including system hardware configuration and software design and configuration. Some measurements quoted in this document may have been made on development-level systems. There is no guarantee these measurements will be the same ongenerally-available systems. Users of this document should verify the applicable data for their specific environment.
 
<...>
 
Information about non-IBM products is obtained from the manufacturers of those products or their published announcements. IBM has not tested those products and cannot confirm the performance, compatibility, or any other claims</quote>
 
That last one is pretty funny (at least to me), especially given that the entire training presentation that I copied it out of was specifically comparing DS8K turbo performance against USP-V and DMX-3. Pretty much admits that the USP-V and DMX data was manufactured, not actually tested.