Posts Tagged ‘Texas Memory Systems’

Erik Eyberg Talks IBM Storage Strategy (Repost with corrections)

Sunday, June 14th, 2015

IBM_Edge_Upcoming_ImageThis is a repost of a previously posted interview with Erik Eyeberg, made as a correction to the previous blog that mislabeled Erik as Erik and Eyberg as Eyeberg.  He noted the discrepancy in an email on Friday and was so darned kind about it, my sense of guilt about my oversight was raised to an intolerable level.  We have corrected the incorrect reference both here and in the video clips below and regret the error.

The video interview was shot at Edge 2015.  This was my fourth IBM Edge event, if memory serves, and my third opportunity to get together with Mr. Eyberg, who came over to IBM with the acquisition of his former employer, Texas Memory Systems, and who has in a comparatively short period of time been promoted through the ranks. He has ditched the nerd glasses and hair cut he sported when I first met him for a more Benedict Cumberbatch look (well, that is what my daughters said when we were editing his video interview!), but he is still geek through and through.

And now, his business cards identify him as the Manager of World Wide Enterprise Storage Strategy and Business Development: a hefty title for a truly smart guy who has stepped almost effortlessly into his new and expanded role. It was a great pleasure to chat with him about IBM’s current technology and future directions. Here is part 1 of our interview, in which Eyberg sets the context, then ramps up a discussion of flash storage technology and the fit it is finding within business enterprises.

But wait, there’s more. We venture away from the evangelism of flash technology to discuss in Part 2 of the interview the lingering concerns that many folks (including me) have regarding the oversell of flash as a panacea and the problems created by memory wear and uneven performance. Erik’s point of view is interesting…

Eyberg makes a coherent case for IBM’s diversification in terms of storage offerings between “boxed” (conventional arrays) and “un-boxed” (software-defined storage) offerings. I liked his sensible discussion of unified management and REST near the end of the clip. It is good to know that IBM is still pursuing RESTful management tools for its kit. Look for RESTful management of the DS8000 series array with the release of version 7.5 of the array’s firmware.

On to the final part. Here is where Erik finishes his thoughts on RESTful management and what it will take for everything to be REST enabled for unified management (he seems dubious that this will happen). Then, he and I talk about tape technology, then about the future of storage from IBM’s perspective. Fascinating stuff.

Once again, special thanks to Erik Eyberg for agreeing to this interview (and to Lizbeth Ramirez Letechipia and company for helping me to round up Erik for this interview and for helping to get the clip approved by IBM).

For the record, this is one of several interviews I conducted at IBM Edge in exchange for room, board, transport, and free attendance at the event. I was also compensated for delivering five sessions at Tech Ed as part of the show.sessions at Tech Ed as part of the show.

Signal and Noise

Wednesday, March 6th, 2013

While I wasn’t in the room to hear the comment directly, IBM’s new storage chief, Ambuj Goyal, is reported by The Register to have stated that his objective was to move transaction storage away from disk to all-flash arrays.  Ultimately, he envisions an IBM that sells less storage.

Later in the article, he clarifies that he isn’t even suggesting flash-assisted disk or hybrid arrays, but all-flash only — probably leveraging Texas Memory Systems (recently acquired by Big Blue) RamSans.

I have my doubts about the readiness of Flash for anything like the heavy lifting of big transaction systems, especially given the memory wear problem that vendors choose (not) to deal with by simply adding a ton of additional memory to substitute in whenever a cell fails and a group of cells is marked as bad.  One credit card company told me recently that his CC processing systems, which handle over 1 million card swipe transactions per second, would burn out Flash SSDs within a few minutes of installation given the write limits that currently exist in SLC and MLC memories.

Of course, last time I checked, TMS was also building SSDs out of DRAM. DRAM SSDs don’t have the memory wear issues of Flash, but DRAM memory is volatile, while Flash is not.  And DRAM rigs tend to be substantially more expensive per GB, not that anyone ever accused IBM of being the low cost leader.

Goyal’s other talking points didn’t raise my hackles.  He thinks that storage virtualization (a la San Volume Controller or SVC) should be deployed to enable non-disruptive deployments of  storage and applications that use it.  I like storage virtualization too.

He also hangs his hat on IBM’s burgeoning Virtual Storage Center, a management console.  I have to agree that management is the missing link in efficient storage and I am delighted that IBM is developing yet another tool set for managing storage.  But is it a storage service management play or a storage resource management play or both?  And does it work with all storage gear, with all IBM storage gear, or only with select IBM rigs?   I hope to learn more about this at Edge 2013 to see whether it is ready for prime time.

I like IBM, having taken my earliest training in IT at IBM schools.  And I have known some top notch engineers from Big Blue over the last three decades.  I am not sure the Omni et Flash thing makes any more sense to me than Omni et Orbis (everything on disk) mantra that the array makers have been preaching for the last 20 years.

I would like to learn how they justify this direction.  Or are we all just chasing the goofy folks behind Evil Machine Corp’s XtremeIO announcement.

Guys, if you think going really fast for a really short time is cool, why not buy a top fuel drag racer?

 

Oh, that’s right.  Top fuel dragsters have a tendency to blow up.

Oh well, I’m starting to flash back to Ed “Big Daddy” Roth and his Rat Fink illustrations from my youth.  But I have long since set aside such childish things.

 

First it was the Mainframe, then Tape, now…

Tuesday, August 26th, 2008

eWeek is running an article arguing, in its teaser, that disk drives are dead — to be replaced by solid state memory.  The link turned out to be a pretty slide show without a lot of content, seemingly put there to provide a lot of advertising opportunities.

Sign of the times, I guess.  Kind of like political conventions that started last night and follow in September.  Not a lot of content, just a lot of commercials.

To be honest, I have always been a fan of SSD, and an admirer of folks like Texas Memory Systems’ Woody Hutsell and others who have been the most vociferous advocates and evangelists of the technology.  However, I have never heard Woody say that the technology would overtake and replace conventional disk — at least not without a significantly improved price/performance ratio.  At best, SSD can be deployed ahead of some very chatty databases to improve performance.  Usually that decision is predicated upon the notion that the work of the database is super mission critical.

Woody and I have also discussed over the past couple of years the opportunity to use memory in laptops as a means to handle battery power drain, particularly with Vista, which writes a trickle of data to the disk on an ongoing basis even when you aren’t doing anything.  Caching this data on less power consumptive media makes sense.

So, why all the hype about the death of disk?  I really wish we could get rid of spinning rust, just like I wish the national debt would go away.  But wishing ain’t doing.  The explanation might be linked to some of the advertisements around the piece.  Seems a lot of laptop makers are looking to flash memory to improve performance and increase cost (as in revenue, margins, etc.)  On the big iron end, interesting architectures are being forwarded by EMC and others that seem to aim at creating a tier 0 inside the array using memory chips.

This sounds rather complimentary to disk, not a replacement.  Ziff needs to find another way to sell ads, I think.