Is Private Cloud the Next Frontier

Business Models

Looking at the state of Cloud Computing from 2006 to 2012, it is interesting to see how different companies modified their postures in the ecosystem.

Amazon has leveraged its data center operational know-how and turned it into a new line of business that is expected to generate north of $1.5 billion in revenue. This is one of the few instances where a big-bang approach for a brand new market that had worked.

Take Salesforce.com, on the other hand, it has focused on creating user stickiness by creating platforms where other vendors can participate in its ecosystem. Instead of jumping into an unrelated business, like Amazon, it has stayed closed with what it knows. And, I can only speculate that the fact that Oracle is breathing down its neck is helping them focus.

For those more "conventional" players, Rackspace had probably done the most in reinventing itself. It not only cobbled together a Cloud OS that is meant to offer Amazon like capabilities, gave it up as an open source project, then wanted to turn its business into managing other companies' data centers (running OpenStack). This is interesting in two respects. One is that it seems to acknowledge Amazon's dominance in public cloud. At the same time, by offering operational management services, Rackspace seems to be betting on its ability to offer differentiated solution that cannot be easily copied by Amazon (who by definition will focus on standardizing their offerings to maximize operational efficiency.)

As for the VCE crowd, with the large data center operators - Amazon, Google, Facebook, etc. - all opting to buy directly from the component makers given their massive scale, the combined VMWare, Cisco, and EMC offer is impressive. With a single stroke, it allows these companies to tap into each other's install base while minimize the complexity and overhead of creating customized solution for each deployment.

Private Cloud

Taking a step back, however. The industry dynamics also seem to be shifting. Amazon has effectively dominated the Public Cloud arena. Amazon will continue to improve its service and coverage along with the bevy of new companies that rely on its infrastructure, like DropBox. Similar to utility, it is unlikely that anyone can take on Amazon at this point.

At the same time, the implicit strategies for companies like Rackspace and VCE is to shift their focus to private cloud. Private Cloud has two advantage from those offered by Amazon. One is that it is (somewhat) customized to each deployment which increases its switching cost. The other is that private cloud users, by the fact that they wanted their own cloud and can pay for it, are less price sensitive and are willing to pay for value-added services.

Distruptions through Cloud

From $0 to $4 billion

It is the story that makes the Silicon Valley lore. Drew Houston not only founded the ubiquitous Dropbox that every other person is using, not only just got a B Round valuation of $4bn, not only turned down a cool $800 million buy-out offer two years prior, he turned down Steve Jobs.

Makes your knees feel just a tad weak, don't it!

And, for the record, DropBox could not have happened so rapidly without a solid Cloud backbone.

The Calculation

On the heel of my recent musing on the status of OpenStack, it is interesting to see how RackSpace CTO sees RackSpace's relationship to OpenStack as its unofficial guardian.

Apparently, the master plan to dominate the world involves widely deploying OpenStack whereby RackSpace can offer its consulting/management services. Seem reasonable enough, but it also seem to require quite a number of pieces to fall into the right place. More importantly, does this mean that RackSpace will take a more assertive role with OpenStack Foundation in the style of MySQL instead of Mozilla?

And I thought Cloud Computing is all about software

A final bit of noteworthy development. Given the relatively simple computational needs of Cloud Computing and its voracity for electric power, there is an arms race to create alternative chips and architecture away from the mainstream Intel x86 design. ARM chips are a favorite right now, but there are definitely chip shops working on other solutions to be followed by the infrastructure/support for the new chips and/or architecture.

Microsoft was disrupted by the internet as embodied by Google. Now, the other half of Wintel needs to watch out for the next disruption.

Related Entries

* Dropbox gets a $4bn valuation in B Round: https://plus.google.com/u/0/104462346399481715282/posts/P4bKc7cUX6y

* RackSpace's master plan for OpenStack: https://plus.google.com/u/0/104462346399481715282/posts/ZeRHQVheEQ8

* Simpler chips for Cloud computing: https://plus.google.com/u/0/104462346399481715282/posts/DFNx4V35xdm

OpenStack and Enterprise Applications

OpenStack

There are any number of Cloud related initiatives: Eucalyptus, Scalr, you name it. But, the "it" initiative these days is OpenStack.

Backed by RackSpace and a team from NASA, it is an open source effort that aims to create a complete Cloud Computing solution that is not encumbered by the VCE cartel (VMWare, Cisco, and EMC - in case you are wondering.)

With its hybrid commercial and tech origin, it also has a duel personality. On the technical side, there are groups of dedicated architects and developers driving the core products and related extensions. On the commercial side, a lot of companies, 100+ as of my last count, have signed up with OpenStack as corporate members.

It is worth watching who will drive the development of OpenStack as a project. On one hand, it could become Mozilla/Linux like with a clear technology focus. On the other, this could become mostly a standard setting body driven by corporate sponsors.

The easy answer is that this all depends on whether OpenStack gains traction in deployment. And, preferably, large scale ones. The real question is how would RackSpace manage the process?

With the announcement of the OpenStack foundation with donated IP at its recent Boston semi-annaul meeting, the foundation's development would offer the first clue on which path OpenStack will go down.

What does Cloud mean if you are not Google, Amazon, or Facebook

On the question of traction, I cannot help but look at the enterprise side of things. Cloud Computing as a utility is likely to end up in the hand of two providers in the US (look up your economics textbook for why so few and think of the US anti-trust history for why more than one.)

On the other hand, there are plenty of enterprises who need Cloud Computing in a meaningful scale. The most interesting one that I have heard so far is how DreamWorks, the animation film producer, is using Cloud for 50% of its rendering work. Here is an example of mission critical work where Cloud makes a lot of sense for.

Similarly, it is gratifying to note that corporate IT chiefs are thinking about what strategic advantages Cloud could offer beyond cost reduction through consolidation. We are likely to see a new batch of Cloud ideas in the next two years.


Related entries

* What's in it for corporate memebers, modular Cloud through OpenStack: https://plus.google.com/u/0/104462346399481715282/posts/GXPX1m5KDHD

* OpenStack Foundation annoucned: https://plus.google.com/u/0/104462346399481715282/posts/Tfig9mTkr59

* DreamWorks offload 50% of its computer processing to Cloud in 2011: https://plus.google.com/u/0/104462346399481715282/posts/9SnKr2q5DLy

* Shifting corporate view on what Cloud is for: https://plus.google.com/u/0/104462346399481715282/posts/8CgTJNQ8zme

Looking for More Distractions?

Google Plus

I am fairly agnostic when it comes to social media. I have tried a good number of them and found most of them wanting in one way or another. Typically, small things that imposes a time tax which, after some good-intention'd efforts, my usage starts to drift.

Hoopla over its growth and its demise (how time flies in the social media scale), aside. I like Google Plus for its open environment instead of being a walled garden. I also like the Twitter like capabilities.

Prometheus Reconsidered on G+

Prometheus Reconsidered now has an open curated stream on G+. It is updated almost daily but features articles that are related to technology, innovation, and business.

Please join the fun: Prometheus Reconsidered on Google Plus.

Cloud Computing with Chinese Characteristics

Rapid enterprise cloud deployments

In a recent interview with GigaOm, Marten Mickos noted that there is a lot of installations in China. On the face of it, it does not seem particularly remarkable since China seems to have a veracious appetite for just about everything else.

The obvious benefit of public cloud

In a recent conversation with a VAR in China, I was told the challenges of deploying new IT systems. Money, everything else being equal, is rarely the problem. Buyers understand the importance of ERP, CRM, and other enterprise solutions as their growth accelerates both domestically and internally. However, over the deployment phase, the physical installation of things like fiber optic lines and configurations for routers, VM, etc. suck up time and attention of highly skilled engineers which are difficult to find and retain.

My question to the VAR was why not use a public Cloud architecture and solve all these problem with a single stroke? After all, this is exactly how SalesForce.com has become a multi-billion dollar company by solving exactly these type of problems.

The subtle problem of public cloud

It turns out that many business owners in China are reluctant to let any of these data go into any public network for fear of being sniff'ed. (Interestingly, this is not a uniquely Chinese concerns, given DT's attempt at fencing its Cloud operation from the prying eyes of the US.)

Looks like private cloud will proliferate in China for a while yet.

Links

* Eucalyptus refreshes IaaS platform: http://gigaom.com/cloud/eucalyptus-refreshes-iaas-platform-isnt-dead-yet/

Googorola and Smartphone Ecosystem 2.0

Google + Motorola Mobility

Okay. Unless you have been kipnaped and gone off the grid in the last week, you have heard about what likely will be considered the deal of the year, Google buying Motorola Mobility (MMI) for $12.5bn.

The most obvious reason is that Google is in serious need of strong patents and in great multitude in order to protect the Android ecosystem that it has built in the past few years after failing to secure the Nortel IP portfolio.

Fear, aspiration, and other reactions

Having the patents from the granddaddy of the mobile handset, Motorola, puts Google in a much better/equal footing with the likes of Apple and Microsoft/Nokia as far as the on-going proxy war in Smartphones are concerned. On one hand, this could potentially re-enforce the alligance of Android makers such as HTC and Samsung to the eco-ystem by minimizing uncertainties for the long term (the existing legal actions between Apple and these markers still need to be played out, but the outcome is likely to be more favorable to Android makers if Google gets involved in a multi-party licensing arrangement.)

On the other hand, by becoming the owner of a handset maker, Google has to maintain the tricky balancing act of ensuring that there is no hint of favorism with Motorola. For now, Android handset makers have expressed support for the acquisition, although, as the saying goes, the proof is in the pudding - we shall see what happens in the next 6-12 months. There has also been alternative scenario proposed suggesting that given the proliferation of Android handsets (300+ as of latest count), Google could be in a position to use Motorola to create reference designs for the whole eco-system so that the end user can get the maximum benefit of Android phone through streamlined design/production. Honestly though, it is hard to see how Google can do this with Motorola in its present form unless it plans to trim MMI down to an expensive hardware lab.

An area that is worth thinking about is how do the telco think about this. While iPhone has been a boom for the likes of AT&T, the consensus is that Apple is capturing most of the profit and accolade while the operators are stuck with footing the bills for massive capital expenditure and handling user complaints. So, what does the creation of another Apple-like 800 pound gorilla in GOOG/MMI mean? So far, the reaction has been very muted.

How about consumers? There was a fun naming competition at WSJ for the rest of us where the readers get to vote for a name for the transaction. The winner is Googorola - the people have spoken. Beyond the thrill of instant veiwer particilation, there was a suggestion that Google will offer free smartphone as an extension of its multi-side platform business model. Since, unlike software which is virtually free once written, hardware comes with a BOM (bill of materials) for each unit, not sure if that would ever happen with the current ecosystem. On the other hand, if you extend the scenario to the model of fab-less IC design houses, it may be a good time to think about who will be, or indeed think about starting, the next TSMC for handsets.

Did you know - a few loose ends

Finally, on the patent front, given Apple's aggressive legal actions, it is interesting to note that Palm (HP, for now) is the only company that Apple has not gone after. Both Apple and Microsoft are siting on a big pile of cash right now. (I am just saying.)

On the smartphone technology front, it is probably worth noting that Mozilla is starting a project to build a smartphone browser that aims to do to smartphone ecosystem what its FireFox (internet browser) has done to the PC ecosystem by rendering Microsoft's operating system playing second fiddle as users migrate to Cloud Computing. Given the success of Mozilla and Firefox, this is a highly credible project. However, with the on-going battle between iOS and Android instead of a dominant platform like Microsoft Windows, this vision may take a while.

* Google buying MMI: http://googleblog.blogspot.com/2011/08/supercharging-android-google-to-acquire.html
* Android Partner quotes: http://www.google.com/press/motorola/quotes/
* Google taking charge of Android Ecosystems: http://www.lightreading.com/blog.asp?blog_sectionid=217&doc_id=211132&f_src=lrdailynewsletter
* Telco and other poentail losers of the GOOG/MMI deal: http://www.fiercewireless.com/story/who-are-losers-googles-125b-acquisition-motorola/2011-08-16?utm_medium=nl&utm_source=internal
* France Telecom "Like" GOOG/MMI: http://allthingsd.com/20110815/u-s-carriers-silent-on-motoroogle-but-france-telecom-gives-it-a-thumbs-up/?mod=mailchimp
* Verizon hoping for peace on the patent war front: http://www.fiercewireless.com/story/verizon-hopes-google-motorola-deal-will-cool-patent-wars/2011-08-17?utm_medium=nl&utm_source=internal
* Does this mean I can get a smartphone for free from Google now? http://www.xconomy.com/san-francisco/2011/08/17/are-free-motorola-phones-in-our-future-a-vc%E2%80%99s-take-on-why-google-is-buying-motorola/
* Readers have spoken, Googorola it is: http://allthingsd.com/20110817/googorola-triumphs-in-snarky-nickname-poll-over-12-5b-bid/?mod=mailchimp
* Apple may fear Palm's (HP) patents the most: http://allthingsd.com/20110819/could-hp-turn-a-profit-on-palms-patents/?mod=mailchimp
* Mozilla/Firefox's Plan B: http://www.technologyreview.com/web/38308/?nlid=nldly&nld=2011-08-15

Smartphones and Billion Dollar Auction

Rockstar's $4.5bn

Well, it is now official. Nortel's 6,000 patent portfolio has been sold to Rockstar, a bidding consortium consisting of Apple, Microsoft, Research in Motion, Sony, Ericsson, and EMC for $4.5bn. To put this in context, the starting stalking horse bid was $900mn from Google.

It is said that the winner companies were bidding mostly for defensive purpose against Google and its Android eco-system. The most obvious targets, beyond Google, would be Samsung and HTC. Words on the street is that there is regulatory concerns on the new IP owners could inhibit the competitive nature of smartphone developments.

The saga continues.

Alcatel-Lucent's $20+bn

With Nortel's auction done, inquiring minds cannot help but wonder what comparable could be drawn in the IP world. An investment banking analyst did some number crunching and suggested that Alctel-Lucent, who owns the IP portfolio of the old (AT&T) Bell Lab, could fetch more than $20 billion alone in a similar auction.

Contrasting with ALU's market cap these days at less than $20bn. I wonder if there may be an activist investor or two who are making plans as we speak.

At $15 per Smartphone

Valuing IP is a tricky business. And, the Nortel's auction, in no small part, is impacted by the recent explosive growth of smartphone eco-systems. To give it a more concrete context, MSFT is rumored to be asking for $15 licensing fee per phone that Samsung ships that uses Android. And, at 19mn phones for the last three months and continuing growth in the foreseeable future, this is real money that even MSFT cannot ignore.


* Nortel's $4.5bn: http://www.bloomberg.com/news/2011-07-01/nortel-sells-patent-portfolio-for-4-5-billion-to-group.html
* Bell Lab's $20bn: http://www.lightreading.com/document.asp?doc_id=209655&f_src=lightreading_gnews
* $15 per for MSFT: http://www.reuters.com/article/2011/07/06/us-samsung-microsoft-idUSTRE7651DB20110706

Three examples of open source projects

Firefox

Firefox came out of Netscape and is currently under the Mozilla Foundation. As an open source project it is most notable in that it competes directly and successfully against software heavyweights such as Microsoft (Internet Explorer), Apple (Safari), and Google (Chrome).

On the licensing front, it employs a tri-license strategy under the MPL (Mozilla Public License), GPL (GNU General Public License), and LGPL (GNU Lesser General Public License). This allows different projects to be combined and gives users the option to pick a particular license. Its revenue is derived from fees paid by Google and other search engines to be included in Firefox.

In a conversation with Asa Dotzler of Firefox, adoption started to take off when Microsoft stopped investing in Internet Explorer. With fear of hackers, both personal and corporate users began to massively adopt Firefox.

MySQL

MySQL is one of the poster child of commercial open source which was sold to SUN Microsystems for US$1 billion in 2008. Many have argued that Oracle's SUN acuqisition in 2009 for US$7.4 billion was really about buying its biggest database competition, MySQL.

On the licensing front, it employs a dual-licensing strategy. Its open source version is available under GPL. It also offers a more traditional proprietary licensing option for users who want to incorporate MySQL into their own system or use additional libraries developed by MySQL AB.

I invited (former) MySQL CEO, Marten Mickos to give a Forum talk at Xerox PARC in 2010. It is a talk worth hearing. In separate conversations, he indicated that one of the major drivers for MySQL adoption was the fact that it has a specific niche that nobody was filling at the time, namely providing database service for web servers.

Defensive vs offensive open source

Traditionally, open source has been a way for corporate giants to create potential insurgencies in their competitor's backyard. For example, IBM and Oracle have been active supporter of Linux operating systems as a way to counteract Microsoft's dominance in the PC OS market.

Google seems to have taken a new approach with Android as an open source project. Instead of being a defensive strategy like IBM and Oracle, Android is used as an additional avenue to engage users, i.e. beyond desktop, and foster new ecosystems, i.e. working with consumer device makers. By becoming the best selling operating systems for mobile devices according to Gartner, Google has demonstrated how to use open source as an offensive strategy to penetrate new spaces outside of a company's core business.

Open source

Open source in software

A quick definition so we are all on the same page. Generally "open source" refers to making software source code available to the public under relaxed copyright restriction, aka copyleft. It also requires any change/modification be made available under similar conditions. In practice, selection of one or a set of compatible licensing models, e.g. BSD license vs GPL (GNU General Public License) vs LGPL (GNU Lesser General Public License), is an important consideration.

The obvious attraction of open source is the ability to tap into a much wider spectrum of talents that may be too expensive or impractical to assemble into a development team. As a collaborative endeavor, it also allows for divergent perspectives, i.e. developer vs users and personal vs corporate, to push the product forward instead of relying on the same set of product managers in a company. With multiple constituents involved from the development cycle, open source can make the marketing costs lower and go-to-market faster. Finally, many have argued that open source can result in higher quality products.

Open source as a commercial endeavor

To succeed as commercial operation, open source companies have adopted different business models. An early model is to offer value-added services such as technical support around the source code. Several companies have done well by having a multi-licensing strategy where the public can access the open source version while enterprise users often opt for the proprietary version with feature extensions. While others would offer the latest version with a fee but becomes open source when a new version becomes available.

One common myth about open source is that it does not require dedicated engineering resources. Nothing is further from the truth. In order to start an open source project, there has to be enough ground work done to ensure that people take it seriously and can use it productively. As the project ecosystem grows, development and management resources are important to keep the work on even keel.

On the other hand, open source approach brings two distinctive advantages in the development process. One is that it allows dedicated people to be involved in a way that is not possible otherwise, e.g. language localization. The other is that it allows for the "survival of the fittest" in features where anyone can propose and add a new feature but only the truly useful ones survive the new release.