Inside System Storage -- by Tony Pearson

Tony Pearson Tony Pearson is a Master Inventor and Senior IT Specialist for the IBM System Storage product line at the IBM Executive Briefing Center in Tucson Arizona, and featured contributor to IBM's developerWorks. In 2011, Tony celebrated his 25th year anniversary with IBM Storage on the same day as the IBM's Centennial. He is author of the Inside System Storage series of books. This blog is for the open exchange of ideas relating to storage and storage networking hardware, software and services. You can also follow him on Twitter @az990tony.
(Short URL for this blog: )
  • Add a Comment
  • Edit
  • More Actions v
  • Quarantine this Entry

Comments (6)

1 localhost commented Trackback

Tony-<div>&nbsp;</div> If IBM in fact "gets it" about Flash, then:<div>&nbsp;</div> * Why have IBM still not corrected the technical inaccuracies in the referenced white paper from my original SplashDance post - after more than 2 months?<div>&nbsp;</div> * Why do you attempt to mislead customers that Dataset Flash Copy is a non-disruptive, array-based means to relocate LUNs into flash, when in fact DFC doesn't even support open systems hosts?<div>&nbsp;</div> * Why do you try to FUD people that the 200/400GB drives are not for write-intensive workloads? No secret that they have less spare capacity. But if you actually knew what you were talking about, you'd know that the default format is massive overkill that does nothing to really extend the real-world life of the drive while it increases the cost per usable GB significantly! <div>&nbsp;</div> Of course, you undoubtedly don't have enough SSDs installed in DS8Ks to understand that yet. It's hard to "get it" until you actually "do it."<div>&nbsp;</div> And BTW, the Fear you try to spread about frequent replacement probably crosses over your "no blog wars" policy: if an EMC EFD has to be replaced, EMC does so under warranty - EMC stands behind every product they sell, and the customer need not worry about the frequency of replacement.<div>&nbsp;</div> Oh, and I'll let you in on a little secret - Symmetrix actually MONITORS the actual wearing of the flash, and will pro-actively phone home for a replacement LONG before the drive would ever exhibit a failure scenario. Does the DS8K?<div>&nbsp;</div> I'll highlight a few more observations from your REVISED (May 7) version of the RedPaper - not much has really changed:<div>&nbsp;</div> * Since you cannot mix SSDs and HDDs on the same DA pair, the ONLY way to spread SSDs across all DA pairs is if you have no disk drives.<div>&nbsp;</div> * SSDs are only available on brand-new DS8Ks, and not as upgrades to existing systems.<div>&nbsp;</div> * DS8K maximum drive configurations are severely reduced if you use SSDs<div>&nbsp;</div> And in fact, it seems that SSDs are apparently a non-standard IBM offering, requiring a special-case RPQ. If you actually "got" flash drives (and the DS8K was able to use them efficiently), I hardly think IBM would be so conservative as to add the overhead and complexity of RPQ to the ordering process.<div>&nbsp;</div> For the record, the only announced V-Max future feature is one that IBM probably hadn't even thought of until it was announced by EMC - FAST. So while your storage engineers are trying to figure out flash, ours are far ahead in optimizing performance, utilization and automated tiering with EFDs.

2 localhost commented Trackback

BarryB,<div>&nbsp;</div> With reference to your final comment - if you read my post you will see we patented the sub-lun moving of extents based on performance criteria back in 2004. There was never enough of an advantage (between FC and SATA) to implement said function - so to make such comments about us 'trying to figure it out' has no proof nor substance.

3 localhost commented Trackback

BarryW - sorry, but I dont think anyone is reading your blog anymore...are you writing blogs again instead of caching algorithms?<div>&nbsp;</div> ;-)<div>&nbsp;</div> Seriously, understanding the concept requires more than a patent to prove that you "get it" - or that anyone else at IBM *other* than you "gets it." In reality, it is the actual implementation that demonstrates the necessary skill and comprehension.<div>&nbsp;</div> But hey, why argue? <div>&nbsp;</div> It has taken over a year to get Flash drives into the DS8K, and you are starting off behind where EMC was back in March 2008 with DMX4. And despite the claims to have been working with flash longer than anyone owing to IBM's server-based support for flash, it is obvious to those of us who don't wear IBM-issued Sunglasses that you continue to under-invest in storage. Where's flash for XIV? iSeries? Non-disruptive migration in/out of Flash on DS8K?<div>&nbsp;</div> And we all know that IBM is a huge company - I thus seriously doubt your that storage engineers have even discussed flash with your server engineers, for that matter.<div>&nbsp;</div> Heck, EMC probably invests more in Symmetrix alone than IBM does across ALL of its home-grown storage products (DS8K, XIV and SVC).<div>&nbsp;</div> Customers clearly recognize this, as evidenced by the May 7, 2009 issue of Goldman Sachs "Data Center Techtonics" survey, which highlights that IBM is falling further and further behind - in both storage and in server share. In fact, Goldman notes that horizontal integration across a broad storage portfolio like EMC (and NetApp) handily outweighs any value of the vertical integration that IBM (and HP) are trying to use to leverage captive account control.<div>&nbsp;</div> But in the end, it really isn't about "getting it" or even "thinking of it first" - it is all about "delivering it" so that CUSTOMERS can get value from the new technology.<div>&nbsp;</div> And in that context, EMC is leading, IBM is following.

4 localhost commented Trackback

BarryB,<div>&nbsp;</div> Some of us have a rewarding day job, and blogging is 'for fun' - as opposed to being full time marketing folks such as yourself.<div>&nbsp;</div> Flash in XIV goes against the ethos of the product as it stands today - other future plans on storage and server platforms are available for interested customers.<div>&nbsp;</div> I had a very fruitful trip to Austin 18 months ago where Team IBM had a stack integrated SSD summit, software, services, servers, storage - you name it - so again making wild statements without any knowledge is just absurd. Just because EMC doesn't have that influence outside the storage world...<div>&nbsp;</div> When you've only really got one product, I guess you have to spend all your money there...<div>&nbsp;</div> I agree its about customer value at the end of the day, in that context EMC is promising things in the future, IBM promises on the day it delivers.

5 localhost commented Trackback

"IBM promises on the day it delivers"<div>&nbsp;</div> That would mean something if you were actually keeping pace with the competition. Not only has IBM been slow to get to SSDs into IBM storage, but the DS8K is pretty much the only storage platform that STILL doesn't support thin provisioning, more than 2 cache partitions, or non-disruptive LUN migrations within the array.<div>&nbsp;</div> As Goldman Sachs observed, customers aren't holding their breath waiting for your storage or server platforms to catch up - they're voting with their feet.

Add a Comment Add a Comment