What a fun week for a technology watcher/analyst/would-be geek like me! The IT world had gotten so boring over the past few years, that I have been happy to turn my attention to customer-led business innovation and customer experience (with IT enablement). This week’s news has taken me back to my heyday as primarily an IT analyst and pundit (as opposed to my current self-image as a customer-centric business/IT strategist).
First, Cisco threw down its gauntlet to IBM, Dell, HP, and Sun and
declared that it too is entering the market for blade servers to be
sold for IT data centers and virtualization. (Translation: there’s not
enough growth in telecommunications and Internet infrastructure, we’re
adding data centers and cloud computing to our remit).
Next, and in parallel, 1) Sun announces its eagerly anticipated “Sun Cloud.” This is Sun’s cloud computing offering, which will compete with Amazon’s cloud, Microsoft’s coming cloud, IBM’s Blue Cloud, and the existing Google cloud (which oddly, no one talks about)—and 2) the rumor leaks that IBM will acquire Sun for $7.5 to $8 billion in cash. Sun’s stock soars; IBM’s stock falls. And the pundits start punditing. What fun!
[UPDATE: First IBM Lowered the Price, then Sun Refused the Deal, but my take is that this will still happen.]
What made this little flurry of news and analysis more exciting was the process I have used to come up with “my” analysis. Over the last 12 hours, I used a combination of social media (twitter search filters to watch the pundits’ tweet their commentary and tweeples’ comments on their punditry) and outreach to my own community of technology architects: Patty’s Pioneers. As their comments rolled in from Hong Kong, Sydney, California, Boston and Chicago, it has been fun to debate and ruminate together (and the discussion continues!).
Steve Siu, the chief technologist at Orient Overseas Container Lines (OOCL) in Hong Kong, summed it up best: “This deal potentially marks a milestone of a new generation of computing economic model—a new chapter. Commoditize all computing activities, back to business model innovation to create value.”
Peter Horne from Blues Point Partners in Sydney, Australia, commented: “Sun is a hardware company with a software community (Java) attached. The hardware company has been dying for a while, and as a hardware company they have proven that it's hard for a hardware Co to understand and leverage a software community. IBM transformed themselves in to a services company and has probably made more money out of Java than Sun (Ditto Oracle). While a single brilliant Sun engineer created Java, they were pedestrian at managing it. I'm not worried about who owns Java now as it's owned by the community of Java developers and can't be directed by any single commercial interest anymore. Linux also becomes the final word in x86 OS alternatives as well. IBM is a Linux shop and I can't see x86 Solaris continuing. IBM also gets MySql which gets them some head space in open source databases. They also get Open Office which may make things a bit more interesting for MS given IBM’s incredible corporate sales ability. Summing up—inevitable outcome for Sun, great outcome for IBM.”
The IBM/Sun deal makes so much sense to me that my first reaction on seeing the news (on Twitter) was “of course.” I realize that Cisco or Oracle may compete with IBM, turning the sale into an auction. But frankly, I doubt it. I believe that Sun’s board and CEO, Jonathan Schwartz, have already shopped the company around. IBM is ready to make the best offer. Cash. Now. At a price that is not embarrassing. (2x Tuesday’s market valuation). I also realize that the deal will probably be slowed down by antitrust deliberations that are stimulated by HP, Cisco, Dell, Microsoft and Oracle—all of whom have the most to lose.
Most of the other pundits are busy comparing the two companies’ hardware, software and services’ portfolios. There’s lots of speculation about what the combined road map will look like.
Hardware: The two firms will need to rationalize their hardware offerings—Sun moved away from its own RISC chips to AMD some time ago. IBM remains a leader in computing chip technology, but derives less of its revenues and profits from hardware these days. Sun’s hardware engineers will probably find a happy home in IBM’s labs.
Software: Both firms have big investments in Java, Linux, and open source. They will need to rationalize their Unix/Linux offerings. Both use Linux; both have their own proprietary versions of Unix (AIX/Solaris). Both support the open source Eclipse development environment. Sun’s software engineers are already in bed with their IBM counterparts.
Steve Siu’s comments about the Linux/Unix variants: “In five years, the Unix server will become legacy; the same way the mainframe did, if we step back 5 years. In the future Linux / open source world, the application execution platform will further commoditize through virtualization. So, this potential deal will trigger corporate users to think about their choices. In the current market environment, all IT projects would require a more solid return. We will minimize special technologies to be employed. We did shut down our mainframe before Y2K, maybe we’ll do the same for our Unix server in 5 years.”
Services: IBM derives the majority of its revenues and profits from its professional services, SaaS, and outsourcing (including cloud computing) offerings. Sun has been terrible at monetizing software and services.
Cloud Computing Is Driving This Deal
In my humble opinion, this deal is all about grabbing mindshare and marketshare in cloud computing. IBM has its BlueCloud, which has been an “also ran” in mindshare and marketshare compared to Amazon’s Elastic Cloud (EC2). In fact, IBM announced its partnership with Amazon Services in February. Sun has just announced its SunCloud—which appears to offer some technical superiority to many of the current cloud computing offerings.
• The Sun Also Rises in the Blue Cloud. IBM and Sun will combine their data center virtualization and cloud computing efforts to offer virtualized data centers (with reduced electric power requirements) combined with cloud computing (offload your non-proprietary and peak load computing requirements to one or more clouds—ideally more than one, for redundancy).
• Google and Microsoft (not Amazon) are the players to beat in cloud computing. One of the things that IBM and Sun have in common is Microsoft as their historical nemesis, and Google as their current conundrum. Google’s free or close-to-free applications running in the cloud are the biggest threat to IBM’s software applications’ portfolio. If companies can run their businesses using Gmail and Google Office in a cloud that Google maintains “for free,” why would they pay IBM (or Microsoft) to provide these software services for a higher price? Privacy, security, reliability, and redundancy sound great, but when times are tough, “good enough” looks pretty good, especially to the fastest growing market segment: small to medium businesses.
• HP and Cisco are behind in Cloud Computing and at War. Allan Leinwand of GigaOm captured the essence of the current battle between HP and Cisco in a recent post, entitled: How HP Can Fight Cisco and Win:
“For the past decade, Cisco and HP have had a cold war alliance—Cisco claimed enterprise IT networking while HP focused on enterprise servers and compute. Both companies have been rewarded with exceptional market capitalizations. Now that Cisco has broken that alliance and declared war with their enterprise servers, HP must respond. I suspect bloodshed will soon ensue, so HP should move quickly—or risk losing the war before the first shot is fired.”
HP’s Blade Servers power many of the virtual machines used in virtualized data centers, such Tesco’s recent roll out of Citrix Xen virtualized servers as the first step in that retailer’s cost-saving infrastructure initiative. (First step: Virtualization; Second Step: Cloud Computing). In November, HP and NetSuite announced a partnership to create a small business SaaS/Cloud computing offering. Back in August, HP Labs announced a cloud computing testbed in partnership with Intel and Yahoo!.
Cisco's entry into the blade server/virtualization/cloud computing business makes sense as a growth strategy. That's why many pundits feel that Cisco may enter a bidding war for Sun.
Watch this space (Cisco/HP) for some counter-moves to the IBM/Sun deal.
Cloud Computing + SaaS + Mobile = the Future
There's one more important tectonic shift in play: the combination of mobile, embedded "edge" computing into every device. We are moving into a world in which all devices (and objects) are intelligent sensing devices which "phone home" to report their status and to request services. "Home" is now becoming a virtual cloud. This is a world that Sun has been designing for a long time. It's one that IBM understands really well, too. (IBM was among the first to promote "edge computing").
Mobile ubiquitous devices talking to a cloud is also the model used by today's telcos-all of whom are now offering computing clouds (as well as their communication and billing clouds). Many of these telcos have relied on Cisco gear to route their networks.
Sun has been a major player in the telco market—both for network servers and for JVMs on mobile devices. As data and communications continue to converge, a Sun/IBM partnership may power many of the compute/communications clouds we rely upon.
The Sun and IBM Brands
From a customer experience/branding/customers' self-image standpoint, IBM and Sun customers have a lot in common. In fact, they're usually the same people. Gone are the days of the "IBM Data Center - CIO". Today the geeks (okay, the experienced geeks) rule. Smart technology architects have been stampeding to Linux, virtualization, and cloud computing for some time. These are Sun's most loyal customers and IBM's most advanced "lead users." These influential customers view themselves as "the smartest guys in the room," and they appreciate suppliers who make them feel smart and let them drive. IBM's open-hearted embrace and sponsorship of open source heralded the "new IBM" and brought its brand image much closer to that of Sun's.
Developer Community and Deployment Community Come Together in the Cloud
I was tickled to discover that the man who practically single-handedly built the Java developers' community—Sun's Lew Tucker—is now responsible for managing Sun Cloud—Sun's deployment platform. Lew designed and managed the original Java developers' online community which was largely responsible for the rapid uptake of Java. Then he was "promoted" to manage Sun's Internet strategy—a job he came to hate because it was too many meetings and too little engineering. So he left Sun to help Salesforce.com ensure that its SaaS cloud was robust and scalable. Then Lew returned to Sun to provision SunCloud. Lew Tucker's journey from developer community to deployment infrastructure actually mirrors a lot of what's currently going on among geeks. Developers are now deploying in the cloud and/or as SaaS. That takes a hybrid set of skills. On the one hand, you need to know less about operating systems and deployment environments. On the other hand, you need to know more about virtualization, security, and optimizing for loosely-coupled computing.
Bottom Line: Sun + IBM is a win for next gen cloud computing.
RSS 2.0 Feeds
Clarification and Citations
Clarification #1: By referencing "dumb terminal and mainframes" I was referring to the earliest organization of computer technology. That standard morphed into the client / server construct, which, in its turn, morphed into the Internet.
The Internet is now morphing again. I am not sure if there even IS a design, or if it is just happening and we are all like social scientists recording data and trying to understand what it is doing.
However, as pertains my argument about the return of “dumb terminals,” given the success of the MIT Media Lab, I doubt there will even be a "dumb" pair of socks by 2020.
Citation #1: The notion that the next morph of computer technology will follow an established path is not my idea. The thought that the dumb/mainframe that evolved to client/server that evolved to Internet will next morph into cloud computers interrelating with smart clients hosting rich Internet applications is not my original idea. I borrowed it. Page 4-8.
(Jeff Tapper, Michael Labriola and Matthew Boles with James Talbot; "Adobe FLEX 3 - Training from the Source"; Peachpit Press [Adobe Press]; Copyright 2008; pp 4-8.)
Citation #2: When Tim O'Reilly told that story about Thomas J. Watson's alleged 1943 comment (Time points between 5:17-6:17: http://www.youtube.com/watch?v=6PNuQHUiV3Q ), he was quoting Clay Shirky. (It must be in: http://www.youtube.com/watch?v=LabqeJEOQyI) So this is a citation of a citation.
Posted by: David Lance | March 22, 2009 at 03:01 PM
Dave DeB--
Participatory medicine using Cloud + SaaS + mobile is a great approach.. as long as we can be secure (which we SHOULD be able to be)..
Thx for the tip on Clay's new book. I'll see if I can Kindle it...
Posted by: Patty Seybold | March 19, 2009 at 06:45 PM
Dave deB--I'm honored to have You on my blog.. I think of Force.com as an e-cosystem.. It's a marketplace of Web Services and SaaS applets that is, to my eye, thriving.. It's a Cloud with lots of action...I think the only real difference between a generic Cloud like IBM's or Sun's and Salesforce's is that in Salesforce's case, it's really an environment that revolves around Salesforce as the central app.. and everything else is literally connected to it (I may be mistaken, I haven't looked at it closely.. but that's my perception).. This is somewhat akin to the Google cloud.. Google hosts it. Google's apps and Google's architecture are at the core. You can add your own apps and run them there, but they'll be google-architected, or at least interoperate with Google's architecture. Amazon's EC2 cloud has two pieces to it. On the one hand, you can just put your apps on a virtual machine on EC2 and run them there.. on the other hand, Amazon Services DOES offer services you can use, e.g. payment services.. A true cloud would let you host anything running in just about any architecture.. Amazon may be the one that does that the best right now..
Posted by: Patty Seybold | March 19, 2009 at 06:41 PM
Dave,
I don't agree that Cloud Computing is "dumb" computing.. meaning dumb terminals... Rich client-side applications are one of the things that makes SaaS really work! Same thing goes with all the applets on our cellphones and blackberries. You don't have to wait for the latency off hitting the cloud and coming back for most of the things you need to do.. Of course, sometimes you ARE submitting a request and needing a response but the request is "skinny".. I am a big FLASH/AIR fan.. It has made my Web experience MUCH better and most of my clients are way ahead of the curve in using rich clients .. take a look at the National Semiconductor site, for example.... All that rich interactivity and user toolkits resides in the browser and most of it can be used offline--e.g. configuring your own LED or power supply...
Posted by: Patty Seybold | March 19, 2009 at 06:33 PM
btw, re "Cloud + SaaS + Mobile = the Future":
There's a huge initiative underway in healthcare for "connected health," and another push to open source solutions, with an intensely rebellious disruptive streak. (I just learned last night that Christensen just published Innovator's Prescription, his healthcare book. Buying it today.)
Meanwhile numerous sharp observers on the e-patient blog note that for the lower socioeconomic tiers, mobile is the platform for HC - health 2.0 apps are great but useless for people who don't have broadband at home. Pew data released last year talked about it.
So now I'm wondering whether cloud + SaaS + mobile is the future of participatory medicine.
Posted by: Dave deBronkart | March 19, 2009 at 05:39 PM
Great to be reading your prognostications again after all these years, Patty. I saw your tweet yesterday looking for a fresh angle - you sure haven't lost the knack for being prolific! Quite a post for one-day turnaround.
What do you make of Salesforce.com's Force.com platform? Is it an order of magnitude lower than also-ran, or is it just a horse of a different color?
Posted by: Dave deBronkart | March 19, 2009 at 05:32 PM
Hi Patty, great to see you back in tech/geek mode.
Adobe's rather intense initiative to infuse the Flash/Flex/AIR model into the technical bloodstream builds upon the idea that "Rich Internet Applications" will increasingly reside independantly on client machines, and will share data with "the cloud" only when the host client has Web connectivity.
This is all bBrowser independant - of course. This Web interactivity is built right into the application. It is a hybrid solution. Increasingly powerful client machines hosting local applications that routinely share data over the Web.
This approach models the client / server paradigm that naturally evolved over the LAN in the last several decades. Network connectivity to ever-more-powerful, smart, independent client PCs, enabling the sharing access to widely distributed data.
If I hear him right, Tim O'Reilly envisions "cloud computing" as the return (on the world stage now) to a single "mainframe" paradigm, with alot of dumb terminals (in a million different forms and functions) interconnected with it.
Speaking of IBM, O'Reilly explains that Thomas J. Watson's alleged 1943 "I think there is a world market for maybe five computers," was off by four.
If O'Reilly is right, and "cloud computing" eliminates the local, client app with one big central computer repository of data and applications as service, doesn't that implode the software industry?
And speaking selfishly, what exactly will I do with all of those terrabytes of space that will soon be available on my new flash hard drives? There are just so many digital photos I can manage effectively...
Could "computing in the cloud" really be the new model? What about the continuing, exponentially increasing power of the PC?
How do increasingly powerful "client" machines, each capable of generating and storing exponentially greater quantities of raw data, fit within a model of "one central computer" feeding data and computing services to an unlimited number of dumb terminals?
Posted by: David Lance | March 19, 2009 at 05:28 PM