Marketecture vs. Architecture
by Pete Gerr on December 10, 2009
For me, and I suspect an increasing number of customers, EMC’s announcements, like this week’s anti-climactic FAST blast, are increasingly “forward-looking” and provide little details or a convincing number of actual customer testimonies (EMC’s own IT department is an objective reference?!). That’s the “Marketecture” part (see Dilbert below), and points to a trend of EMC marketing getting even further ahead of EMC products. Caveat emptor!
(c) Scott Adams, Inc.
With V-Max and FAST, EMC is attempting to change the rules again (like it did with Centera) to create a market view that fits its own technology roadmap, not necessarily that of customers or users. That is Marketecture.
On the other hand is HDS, a company that analysts refer to as the “best kept secret in storage” because it consistently delivers products that solve customer’s problems but spends a lot less than its competitors do on marketing and awareness. HDS will fix that, I assure you, and the beneficiaries will be the customers, and of course HDS and Hitachi, Ltd who create and deliver some truly remarkable technology. I believe end-users and industry experts are less and less enamored and swayed by fancy slideware, press releases short on details, and fast-talking executives. They want to see and to understand what’s real. That’s where the architecture part comes in.
I came up with at least a dozen alternate, semi-clever names for this post, most of which included a play on “FAST” and “V-MAX”, EMC’s latest announcements, but it’s not really my style and besides, the press and industry types are doing a fine job on my behalf of focusing the harsh light of reality on paper tiger products and fluff press releases (especially on FAST).
So let’s get the facts out of the way first: EMC was able to claim “general availability” for FAST v1.0 before the end of CY2009 (an out-of-pitch party horn is heard in the distance). I’m a Mac user, but recall that Microsoft often claims “GA” on products that wouldn’t pass other vendor’s Beta QA requirements, and then charges for them as an “upgrade”. I’m sensing a pattern here.
EMC did state FAST would GA this year. It did. That’s about it. Put a bow on it; in fact, that’s an appropriate phrase given the impending holidays since from a customer perspective, FAST appears to be EMC “regifting” previously-sold products under a shiny new acronym and/or re-charging customers who have already purchased SymmOptimizer, Dynamic Cache Partitioning, and other products just to get an upgraded GUI. (Yes, Chuck and Barry, I know there are a few “new” bits here - let’s call them stocking stuffers, to this announcement for CX users).
I’d like to direct end-users and interested readers over to Hu Yoshida’s blog post in response to FAST, which clearly articulates what FAST is and isn’t, and encourages users not to be swayed by fancy graphics, clever marketing and cool acronyms. The reality here is, HDS delivered on the promises of FAST first, and “oh by the way”, did it about 10 years ago.
HDS delivered a common virtualized architecture and continues to expand upon it, but perhaps fell short on getting the word out broadly enough. EMC delivered, yet again, more marketecture, and the storage industry and customers are finding that its a transparent vision built upon a house of cards.
So, what do you make of this week’s announcements? I would also be very interested to hear your thoughts about how HDS can better express its technological innovation or if it’s doing enough already.
Comments (7 )
the storage anarchist on 10 Dec 2009 at 12:34 pm
Hu doesn’t seem willing to accept comments from me on his posts.
Will you? Let’s see.
Claiming to have been first with Fully Automated Storage Tiering by 10 years is about as misleading as it gets - perhaps it is intended to cover up the fact that Hitachi trailed EMC on both SATA and Flash drive support by more than a year each (and all the while you HDS folks were telling everyone who would listen that NEITHER belonged in enterprise storage arrays…how does THAT crow taste?)
QUESTION: Where’s the “Fully Automated” part of Tiered Storage Manager?
ANSWER: There isn’t one. A human has to analyze the data and decide which volumes to relocate.
Customers tend not to care about who was “first.” They want solutions to their problems, and the ones I talk to specifically demand AUTOMATION, not just tools. As explained in my recent post (http://thestorageanarchist.typepad.com/weblog/2009/10/2028-not-so-fast-hitachi-hp.html) FAST delivers, TSM does not.
Oh, and my advice in response to your closing ask for ways to better express yourselves, you might start by re-reading that cartoon you embedded in your post, inserting the word “SPC” before “benchmark test”.
Scott Adams was clearly talking about HDS, not EMC.
Pete Gerr on 10 Dec 2009 at 1:38 pm
Hi Barry,
I’m happy to respond.
However, I’m not going to resort to the schoolyard fist-a-cuffs tone or approach, I’ll simply state the facts as I see them, and my opinion. Let the readers, market, and customers decide for themselves.
I do find numerous inconsistencies in your comment and logic, and I just want to be clear.
1. Please re-read the statement in my post. I wrote, “HDS delivered on the promises of FAST first”, I did not claim HDS delivered Fully Automated Storage Tiering first. I don’t believe the majority of enterprise customers want to relinquish full control over data movement and migration operations. Please see #3 below.
2. In his post, Hu described key features in HDS’ Tiered Storage Manager (TSM) product which I contend provide a more complete solution which addresses customer’s most critical issues today, while not forcing operational changes, additional complexity, or by reducing existing functionality. In contrast, my understanding is FAST does not currently support the previously-introduced Virtual Provisioning feature, just to name a single case, so again the customer is forced to choose to change. I’m sure you’ll re-post if my understanding is inaccurate here.
3. Finally, I disagree with your statement that customers “demand automation” now. I’ll concede the point that I’m sure there are a small number of customers that do “demand automation”, as you put it. I believe, and research and experience tells me, that the majority of enterprise-class customers that operate large-scale data centers want migration and data mobility without complexity, or with less complexity. That doesn’t mean they want automation or additional products. I believe (the majority of) customers want to retain control over these migrations, still and want more intuitive, user-friendly tools to help them accomplish their jobs. This is the real goal and the approach HDS takes and has always taken - do right by the customer first, all else will follow.
Best regards,
Pete
the storage anarchist on 10 Dec 2009 at 2:24 pm
Fair enough - we can agree that we see different requirements from the market. Perhaps we are even talking about the needs of very different customers.
1) The promise of FAST is to automate a process which is fast becoming too complex for human intervention - balancing available resources against dynamic workloads. Hitachi has historically taken the tools approach to such challenges, while EMC has tended towards policy-driven automation - in addition to FAST, this is true for features such as cache partitioning, priority controls and even multi-site replication recovery. This difference belies a fundamental difference in the way we view things, I guess…but I contend that TSM doesn’t deliver on the same promises as FAST.
2) FAST v1 is BY DEFINITION automation for full-LUN storage tiering, and as such is immediately applicable to the predominant way users deploy high-end storage today. While FAST v1 indeed does not support Virtual Provisioning, FAST v2 is specifically designed for the VP environment, where it will leverage the foundation of VP to distribute sub-LUN chunks/extents across multiple different tiers - again, by automated policy. Thus, FAST v1 and v2 are actually two complementary components of the automated solution; asserting that the fork cannot cut like the knife isn’t really an insult to the fork.
3) Again, I concede we are probably talking about different customers. The ones I meet operate under flat-to-shrinking budgets and are challenged to support data growth in excess of 60% CAGR annually with flat head count. For them the V-Max offers an array with up to 2PB of usable capacity today (3.6PB soon with 2TB SATA drives); by itself this simplifies things to fewer managed components. The important thing to realize is that in such large environments, MANUAL optimization is simply not an option - there simply aren’t enough people with enough time on their hands to be analyzing reports and hand-optimizing the data layouts. Hence EMC embarked to deliver FAST with a focus on intelligent AUTOMATION, so as to maximizing the value of exponentially-growing storage assets without requiring an exponential growth in people to manage them.
Both your and Hu’s posts seem to overlook not only the value of automation, but also the promise of reduced storage costs FAST can deliver. Optimizing storage around the lowest $/GB of SATA and using the I/O performance of flash (the loest $/IOP), customers can leverage FAST to reduce both their acquisition and operational costs. I won’t quite claim that the death of the 15K disk drive looms, but you can’t argue that they are today the most expensive solution to both IOPS and Capacity.
Finally, your assertions that people don’t want to give up control to intelligent automation are in direct contradiction to everything we (collectively) have accomplished over the past 20 decades on our journey to server (and storage) virtualization. Where once we manually optimized memory managers and hand-crafted application placement on dedicated physical resources, today we leave it to the operating systems and virtualization platforms to manage. IT has (gladly) passed that responsibility over to software; why should storage be any different?
No, although trust must indeed be earned before control is relinquished, I assert (and I think Tony Asaro for one agrees with me): it is inevitable that automation will replace human control of storage resources. Fighter Pilots no longer control their planes - the machines are far too complicated for for simple hands-on flight. Instead, pilots merely make the decisions about where to go and what to shoot down, and the plane handles the rest.
And indeed, I am perfectly willing and able to have a non-combative discussion with you (and Hu). It’ll be easier without the “regifting” insults, Dilbert cartoons, and “gee I’m not sure but…” FUD. Although I’ve probably earned it over the past couple of years, I’m willing to bring this back to an honest and open discourse based on facts if you guys will do the same.
And my offer to Hu to explain Symmetrix, V-Max and FAST to him goes for you and Claus and whoever else is interested. I cringe every time I see you (HDS collectively, not you personally) misrepresenting EMC’s products based on outdated knowledge from the Mr. T days.
Respectfully,
Barry
Pete Gerr on 10 Dec 2009 at 4:54 pm
Hi Barry,
I think our differences come down to a matter of degree and a matter of perspective; I certainly support automation, couldn’t and wouldn’t refute the advances made in this area by our industry and others, as you point out.
It would intellectually dishonest for me to state that “automation doesn’t matter”, or that “automation cannot help reduce user’s operational burden”; HDS believes this, our products, solutions, and services support this notion and deliver that value today and have for some time.
Again, it’s a matter of degree - I think HDS believes customers do want and need some degree of automation; I think the difference is, again, the perspective with which we view the market, our customers, and the “right” time to introduce innovation.
HDS chooses not to spend as much to market our innovation and waits, perhaps too long, to toot our own horn. Geoffrey Moore would likely describe HDS as “pragmatic” and aligned with the “early majority” of technology adopters - sort of in the middle of the bell curve. That’s a conscious choice on HDS’ part. That does not mean HDS doesn’t innovate, or innovate at as rapid a pace or as much as other vendors, an attribute usually linked to “early adopters”; I could write a book on HDS’ and Hitachi Ltd’s innovations in a variety of industries.
Maybe we are speaking with different customers, though I doubt that. Our respective employers have been battling each other for the same market share points for decades. What I think is that HDS might hear what those customers are saying in a slightly different way than EMC does, and certainly would say that HDS reacts differently, not worse or better, not slower or faster, to what we hear. We occupy different roles in the marketplace.
That said, I reserve the right to call EMC out, or any other vendor out, when I think they’re doing something dishonest, or when what they’re saying in marketing doesn’t match the real story. That’s what I was doing with my post and comments here.
I’ll grant you the right to do the same, and I suppose we can both continue to beat each other up about failed or futile marketing campaigns….let’s call it the “Mr. T. Doctrine”. Just don’t sick the “EMC Orb of Death” on me from the marketing campaigns of 2000 or 2001! 
the storage anarchist on 11 Dec 2009 at 5:15 am
Fair enough. You can keep “calling out” your perceived dishonesties - it is indeed your right. But I don’t need your permission to do the same, thank you just the same.
And if you and your fellow HDS bloggers would like to understand the reality of EMC’s products, my offer stands.
One little observation before I go though. You (and others) attribute much (maybe all) of EMC’s success to its marketing, implying that said marketing is covering up or inflating the value of EMC’s products. While the engineers may see that as an insult, and marketing might consider that a compliment, I contend that no amount of marketing can make up for products that do not meet customer needs - especially over the long haul.
Note that Symmetrix has maintained its #1 market share position year in and year out for nearly 2 decades straight.
Perhaps it is not as you say. Consider instead that (just maybe) the products (and the company) are truly that good at meeting customer requirements. Maybe EMC marketing is merely adding the sparkle, not the shine, to the products.
Could it be?
EMC’s approach to marketing is much like the majority of New England companies - conservatively aggressive, never promoting what the company cannot deliver, speaking loudly about what it is delivering and where it intends to go. And EMC’s is not big-budget marketing, either - you don’t see any TV or even web-based advertising.
In fact, the FAST announcement consisted primarily of a pre-announcement in April, a press release, journalist pre-briefings and a handful of posts by EMC bloggers. What made it bigger than life was a) competition’s attempts to pre-empt the announcement, b) broad coverage across the entire storage blogosphere and c) competive attack responses (like yours) - all these brought a massive amount of customer attention to the announcement and made it larger than life.
And as Brendan Behan said: “There’s no such thing as bad publicity.”
Thanks!
Pete Gerr on 11 Dec 2009 at 9:34 am
So, just to close out my part of this thread, Barry, because I think we’re dancing around a rat hole here, you’ve made a claim that is most certainly false, and I’ve got to ask you to please share the source of data (I would hope it is 3rd party reputable data such as IDC, and not your own internal EMC data) you’re using to arrive at the conclusion that “Sym has maintained a #1 market share position for “nearly” the past 20 years” continuously. That claim is absurd.
I could be wrong, but I’d bet a coffee that IBM was actually the market share leader in the high-end, enterprise, “Tier 1″, the market that Sym, USP, and ESS, Shark or the DS8000 serve, whatever you want to call it, 20 years ago, in 1989. I think we all know anyone can make quantitative analysis support any conclusion you want to draw, but the one you’re making here is not a perceived dishonesty, it’s a real one.
Regards, and thanks again for this discussion.
Pete
the storage anarchist on 11 Dec 2009 at 10:26 am
I said “nearly 20 years”…and if you want me to be more precise, I’ll say that Symm has been #1 in the high-end since the mid-1990’s.
So indeed, “more than 15 years” would have been more accurate. My bad.
Today, using IDC’s Storage Tracker data, Symm remains #1, USP(et al) is #2, and IBM DS8K is #3 in the high end.
And EMC is #1 overall for external controller-based disk storage (Gartner) and #1 overall for Worldwide External Disk Storage Systems Factory Revenue (IDC), as announced in their respective Dec 4th press releases.
