John Katsaros

Big Data Buzz

You want to stay one step ahead of the other guy. When someone asks you what's new, you can't say "The Cloud" anymore — that's the old new. The new new is Big Data. It's a perfect answer to the "what's new" question because most likely the person who asked you the "what's new" question doesn't want to show that he has no idea what you're talking about and he'll think you're a genius. — especially if you're coming from a world where data was mostly small. The 2011 prize for out buzz wording the other guy should go to EMC whose user conference was subtitled "Cloud meets Big Data" which made for interesting theme logo designs but not much in the way of content on this particular intersection. We work in an industry that loves its buzzwords so thank goodness we've got a new one. Jumping right onto the Big Data bandwagon we've assembled our initial list of venture backed startups addressing Big Data infrastructure. It's interesting that this category is already up to the $100 Million investment level with Cloudera topping the list with almost one third of the total. If you think we've missed your favorite Big Data startup send us a note: jkatsaros@irg-intl.com

 Our List of Interesting Big Data Infrastructure Startups

Company	Investment	Investors	Location
BackType	$1.32M	YCombinator, True Ventures, lowercase, Freestyle	San Francisco, CA
Cloudera	$36M	Greylock, Meritech Capital, Accel	Palo Alto, CA
DataStax	$2.7M	Lightspeed, Sequoia, Rackspace	Burlingame, CA
Datameer	$11.75M	Kleiner Perkins, Redpoint	San Mateo, CA
HStreaming			Chicago, IL
Karmasphere	$5M	Hummer, Winblad, USVP	Campbell, CA
Kitenga			Santa Clara, CA
MapR Technologies	$9M	Lightspeed, NEA	San Jose, CA
Pentaho	$32M	Benchmark	Orlando, FL
StackIQ	$3M	Antham, Avalon	La Jolla, CA
Total Investment	$99.45M+

Winter in New York

Last week in New York EMC had its analyst conference combined with its launch event (which they named "Record Breakers" for reasons that were more about showmanship than storage) introducing over forty new products. Although we won't spend time wondering why a motorcyclist named Bubba needed to jump over 40 EMC Storage Units the launch event does give us a chance to talk about EMC. But first there are a couple of "housekeeping" topics.

The Synthetic Mainframe Vendor

During the EMC Analyst Conference we had a session with Michael Capellas who joined VCE six months ago as Chairman and CEO — VCE is the Virtual Computing Environment company set up by Cisco, EMC and VMware. He's got a huge personality and the resume to back it — most recently he was Chairman and CEO of First Data. Before that he was head of Compaq and drove the merger with HP where he was president for awhile after which he moved on to rebuild MCI which ended up being acquired by Verizon. He's definitely what you would call an industry heavyweight. In May when it was announced that he would run the VCE coalition we pointed out in this newsletter that VCE would change to suit Capellas' goals and not the other way around and that he most likely would want to create something new and visionary in the computer business. With his credentials and capabilities it's unlikely that he took this job just to create some best practices building blocks. VCE has progressed rapidly since Capellas joined but they're still tight lipped about the long term plans. It's our guess that Capellas is building a modern mainframe company — you might want to call it a synthetic mainframe vendor since it sources key technology from other entities so that it can continuously take advantage of the best stuff that's available. Today the big three dominate the server market — HP, IBM and Dell. A year from now we may be talking about the big four.

Big Numbers

At its analyst conference, EMC's CEO Joe Tucci said that in 2010 EMC had shipped over 10 Petabytes of Flash (up from 3PB in 2009) with a 25% attach rate. According to Tucci, that's more Flash shipments than the rest of the storage vendors combined. 10PB is a lot and the growth rate is remarkable. Overall, IDC estimates that in 2010 EMC shipped about 2,506 PB which is about 33% more than 2009 so a measly 10PB may not be large by comparison, it's growing fast. Since Flash was still more expensive than spinning disks, going into 2010 it wasn't clear how fast Flash adoption would happen. Many of our pundit peers said that Flash would only take off once its price was less than spinning alternatives. That turned out to be wrong. Users didn't wait until Flash was less expensive than the alternative, it just had to be priced in the right neighborhood. As importantly, the software that allows users to optimize their use of Flash became "table stakes" during 2010 with most storage vendors making

Holy Cow

HP's announcement this week that it's acquiring ArcSight for $1.5B is a huge event in the infrastructure space for a number of reasons not the least of which is ArcSight's strategy. Five years ago there were 10 to 15 Security Event Management companies mostly funded by venture capital. Back then it was hard to tell the differences between these companies as they kept leapfrogging each other with new features and capabilities. Eventually ArcSight became the market leader when it re-engineered its user interface with a very graphic visualization of the network topology and the events being monitored. The company kept its leadership by continuing to enhance its value proposition by adding adjacent functionality like event correlation and compliance onto the ArcSight platform and by continuing with its direct sales model. For reasons which we may never understand, the competitors had trouble keeping up as ArcSight continued to gain market share. Cisco, IBM, Novell and EMC all made acquisitions in this space and in one way or another dropped the ball without gaining significant market share. But HP's acquisition of ArcSight is a big event for another reason - the consolidation of security. Last month it was Intel's acquisition of McAfee. This month it's HP and ArcSight. What's next? Who is next?

The Azure Appliance

One of the most interesting announcements of the year was Microsoft's Azure Appliance — a container of 1,000 servers, delivered to your data center, running an up-to-date private version of the Azure Cloud platform. Last year we saw the introductions of Cisco's UCS, the Cisco/VMware/EMC VCE consortium, V-Block system blueprints and Acadia JV, all intended to help their customers achieve utility computing more quickly. Well if that's the goal, isn't something like the Azure Appliance an even better answer? Of course you have to have some respect for the importance of the Windows software and business ecologies, and you have to think Azure is pretty interesting on its own, but with those caveats isn't the appliance version interesting! The initial package (presumably similar to the container level packaging which Microsoft uses to build out Azure and their other large properties) is inappropriate for most but not all (huge chunk of platform, specific cooling and power requirements) but is a natural first step. With the issues of non-Microsoft multi-tenancy and new network topologies understood, a next step of creating smaller packages seems pretty straightforward (Microsoft has already said they are moving to a smaller granularity for their own data centers). What Microsoft clearly understands (or at least Bob Muglia clearly understands) is the difference between having a customer build their own utility solution out of the pieces (Windows Server, Hyper-V, System Center and Virtual Machine Manager, SQL Server, Exchange, SharePoint,...) and giving them an up-to-date and evolving platform is pretty dramatic. UCS and VBlock customers are still dependent on the product development and release cycle of VMware, EMC, Cisco and others, and still fully responsible for the pain and anguish of integrating new versions.

New kid on the bloc or the birth of a new computer company?

Last month, Cisco and EMC appointed Michael Capellas to lead the VCE coalition and named him CEO of ACADIA the Cisco-EMC joint venture with investment from Intel and VMware where he reports to Joe Tucci, John Chambers and the ACADIA board (consisting of two EMC execs and two Cisco execs plus Capellas). It's hard to imagine anyone able to have that many top level bosses but Capellas certainly isn't your average CEO. Capellas was Chairman and CEO of Compaq in 1998 leading up to its acquisition by HP where he spent time after the merger as president. After that he ran MCI during a difficult rebuilding phase. You might wonder why Capellas would take on an assignment like ACADIA - we've been told since it was formed in November 2009 that ACADIA was set up to sell Vblocks - preconfigured versions of Cisco's UCS server, EMC storage and VMware software. But maybe there's going to be more to this story than that - what might interest someone like Capellas is the launching pad to build something bigger - more than an entity selling preconfigured packages - more like a new computer company. ACADIA could be the perfect platform for doing this — with a birthright consisting of networking, server, storage, software and with Capellas as CEO can it grow quickly enough to compete with the likes of computer giants IBM and HP?

Seething

In last week's press release announcement of Diane Greene's sudden departure you could almost feel Joe Tucci's tension jumping out of the page. Tucci, who has multiple roles as the chairman of VMware's board as well as chairman, president and CEO of EMC, has had to sit back and watch EMC's share price drop -- being pulled down by VMware -- since Microsoft launched Hyper-V. That's because EMC owns such a big piece of VMware which, in turn, accounts for a significant amount of EMC's current valuation. Until VMware's IPO last year, EMC's stock price was fairly typical for a large high tech company. But for the past six months EMC's stock price has been as volatile as they come. You can't blame him for wanting to get things under control.

End Game?

The Information Leak Prevention infrastructure sector makes you scratch your head and wonder how this whole thing plays out in the long run. Last year we wrote a series of reports on ILP/DLP technology and basically concluded that, while these products do interesting things, customers wanted more complete offerings that delivered important business value and that this market needed more time before it would ignite. Last week's acquisition of Tablus by EMC's RSA division seems to underscore this and may even point out that ILP technology is better viewed as extending the functionality of larger information management subsystems. Websense is taking this approach by integrating its recently acquired Port Authority product within a broader framework. The value of the EMC/Tablus transaction wasn't announced, which more often than not is a disappointing signal. Couple that with the ensuing discussions pointing out that Vontu's 2007 revenue number may be closer to $25M than to $50M has us all wondering about the short term health of this sector. With about 20 plus startups in this sector and more than $500M in venture investments, we're all expecting an exciting business to develop. As we've done before in other sectors, we'll be keeping score as newcomers enter into our DLP/ILP Hall of Fame.

 

DLP/ILP Hall of Fame

Date	Acquiree	Acquirer	Amount	Estimated Sales Multiple
10/06	Onigma Ltd	McAfee	$20	?
12/06	Port Authority	Websense	$90	20x+??
8/07	EMC/RSA	Tablus	???	???
Total			$110+M

Thanksgiving at the Melvins

We caught up with Network Intelligence's EVP of Marketing and Business Development Jim Melvin shortly after EMC acquired the company at the end of last month.  Jim was instrumental in assisting Network Intelligence in turning the corner from a second tier SIM player into a contender which eventually caught EMC's interest as EMC expanded its enterprise security portfolio.  Jim was too polite to comment on the fate of the remaining independent SIM companies (it's our belief that with Cisco hitting on all cylinders with its MARS product -- as of August Cisco brought on 2200 new MARS customers -- it's going to be difficult for the independent startups still left in the space) but there still may be a few large systems companies that view SIM as an interesting form of product line extension.  Jim's view of the acquisition is "right time, right company" -- EMC came along as Network Intelligence was reaping the benefits of its new strategy.  If you're counting, the latest scorecard for SIM acquisitions is below.  Jim's brother Jeff is the VP sales at Network Intelligence so it's Thanksgiving dinner at their house this year!