1.001: this is like déjà vu all over again
Is it just me, or have IBM's storage execs suddenly started sounding like their mainframe execs used to a few decades ago?
Seems that every day someone else over there at Big Blue wants to go on record insisting that the best IT solution is one that is supplied by a single vendor - end-to-end integration, replace the "stickiness" of unique solutions with the "stickiness" of IBM software. And all that.
In Andy Monshaw's interview with eWeek.com, he makes it pretty clear that he believes that the best IT solution comes from a single vendor, with end-to-end integration. As in: buy your servers, your storage, your networking, your applications and your services all from IBM. Don't worry about vendor lock-in, because you'll be able to put IBM SVC kit in front of your old third party gear and do things the Blue Way with little or no hassle.
Oh, and really fast flash-based storage simply isn't good enough (go figure, since they don't have that in their portfolio). No, apparently (according to Andy) you're going to need to integrate flash into every aspect of your compute platform before you can gain any measurable value.
Bullship, I say to that: Bullship!
Andy goes on to say he no longer needs to "sweep the floor" - that an unintended benefit of the SVC is that customers no longer have to be locked into their storage platforms (as if they ever where - really).
But don't be fooled - despite the eloquent words, make no mistake: in Andy's vision, you'll still be handcuffed. He'll just paint whatever you have on the floor IBM Blue with his SVC spray paint.
It's what's best for the customer, he asserts.
In Andy's world, apparantely, there's nothing wrong with vendor lock-in, so long as you're wrapped from head to toe in IBM Blue.
lock-in by another name?
And then fellow blogger TonyP chimes in (from an unexpected angle) with his another round of his inevitable misleading FUD against EMC, challenging certain unnamed "proprietary interfaces" as EMC's vehicle for lock-in.
Care to back that FUD up with some specifics, TonyP?
From where I sit, this unprovoked (and unsubstantiated) attack is coming from a decidedly Black Pot. I'm not exactly sure about the motivation behind the allegations, though, since they are coming from a company that still generates about 1,000 new patents a year to protect their sole-source mainframe cash cow, whose licensing policies effectively prohibit small companies from providing alternative solutions for their proprietary server platforms and who have yet to fully implement available industry standards such as SMI-S across their entire portfolio.
- zSeries/just-plain- "Z"? Proprietary
- System 38/AS400/iSeries/Series i/just-plain-"i" - Proprietary
- FlashCopyTM? Proprietary
- Global MirrorTM? Proprietary
- GDPS? Proprietary (and expensive)
You get the picture - IBM's portfolio bulges with "interfaces" that are barricaded behind patents, or available only through expensive cross-licensing, ensuring customer lock-in and competitor lock-out for all but the most wealthy (or daring) of challengers. That's why there are only three storage suppliers for zOS disk storage, and two for the native "I" platform.
I imagine that if IBM could patent "vendor lock-in" for their own exclusive use, they would!
TonyP and Monopoly's Mr. Pennybags
Separated at Birth?
But isn't this pretty much the same sort of behavior what got IBM in anti-trust trouble a couple of decades ago? (For those of you younger than 30 or so, you might want to Google "IBM anti-trust" for a little background on the topic).
But more important than that - didn't the whole Open Systems revolution emerge specifically to ensure customers wouldn't have to choose a single vendor in order to get a highly integrated and efficient IT platform for his/her business?
In fact, wasn't the whole idea to break the vendor dependency for interopable components and allow customers to choose what they thought was "best of breed" for their data center, and NOT have to exhaust their own valuable resources to make everything work well together?
And isn't an open, level playing field supposed to foster effective integration, continued innovation and (dare I say it) honest competition?
So why the heck are IBM executives starting to sound like they're back on the old roadway to sole-source provider through proprietary vertical integration?
Is this really what's best for the customer?
Those who don't learn from history...are destined to repeat it.
oh, one last thing...
Allow me to set the record straight on TonyP's FUD: the thing that makes it hard to move from EMC storage to IBM storage isn't some undefined EMC-proprietary interface.
Nope. It's the simple fact that IBM storage doesn't provide the same comprehensive functionality, distributed protection, continuous availability and cost-effective scalability that EMC's customers enjoy.
That seems a lot of sacrifices to make in return for IBM's end-to-end
vendor lock-in "integration."
Case in point: EMC recently delivered array-based equivalent to IBM's FlashCopy in Enginuity 5773 (using licensed interface specifications). As you might imagine, this was done by leveraging the existing local replication implementation that makes up the TimeFinder suite on Symmetrix. The integration was pretty straight-forward, except that the Enginuity microcode engineers literally had to disable dozens of features that TimeFinder supports in order to be operationally compatible with IBM's feature-limited equivalent.
There was more than one surprised beta tester of the new "Compatible Flash" feature in 5773 who complained about the regression of functionality in the new product...I guess that they had hoped we'd have been a little less compatible.
Who wants to switch to get less?