Saturday, May 26, 2012

Evil Assets, Expensive Clouds and Open Source

One of the primary focuses on Cloud Computing is the cost benefit, a focus driven primarily by those who adopt only the virtualization aspect of cloud.  However in reality cloud can be very expensive.  Sometimes being more expensive as a cost is still beneficial to becoming an asset, but sometimes expensive is just simply expensive.

The real economic value of Cloud Computing is in its delivery model; the ability for companies to consume the resources they need while leveraging the assets of another.  Assets are evil. Assets cost capital, a rare resource in companies which they protect and work to minimize the distribution of each year.  Assets also require maintenance, good old tender love and care.  CIO's today complain about the large percentage of their budget that just goes to the care and feeding of their IT systems.  Just like a pet, a server also requires its equivalents of food, air, a litter box, and play time.

Pretending these two issues didn't exist there is a darker, more hideous side to assets; one that makes them orders of magnitude more evil than we mortals truly understood years ago.  Assets create gravity.  First once there is an asset it tends to attract other assets like the formation of a planet.  Although the weakest force, gravity works over the longest distance pulling together other assets until soon something near the size of a real planet begins to emerge in a data center.  It becomes impossible to separate the assets and the collection of assets takes on a life of its own.  Soon companies are paying the legacy tax, often upwards of $1MM, just to make even the smallest change because layers and layers and layers of assets have to be fumbled through and moved out of the way over the course of any implementation.

And of course with gravity comes friction, that negative force which slows momentum eventually bringing everything to a stop.  Why have companies become so bad at innovation over the past 25yrs?  I argue because the weight of the things they own and thus are compelled to take advantage of keep them from doing what they should, most often leveraging things they don't own.  Remember that one of the key financial indicators for a company is Return on Assets, but only if you own assets.

Public cloud, even private cloud when done right in an off-premise model, is asset free.  However internal private clouds, where the Fortune 500 is clearly focused, are asset intensive.  Organizations see the assets as free because the initial set are already on the books.  CIO's are often blinded to the eventual reality of the albatross after a generation or two of changing out hardware.  Of equal concern is the cost.  Software vendors are known collectively for one thing: getting their money.  The greatest concern I have for private clouds is the cost basis on which they are being built.  First any proprietary elements one should assume will increase in cost as they increase in value to the company.  The larger the implementation, the greater the bill.  This is in contrast to existing models however once a company builds their own private cloud they will be extremely reluctant to make any changes to the core and the vendors know this well.  Second the cost basis often consumes the benefits of scale making costs scale linearly when they should in fact drop as more nodes are moved into production.  And third since there are no real cloud solutions in the market, the best being nothing more than hyped virtualization, what cost elements are not represented but will be required to execute native cloud applications?

There is no better application of open source than in cloud computing in all of its forms.  Open Source solutions deliver:

  • code level scrutiny required to meet stringent security requirements
  • a common base from which everyone benefits
  • innovation by encouraging contributions and participation
  • customization as required by a company
  • the ability to remand issues rather than waiting for a vendor

Add to the above the low/no cost and the availability of support for all the most popular tools today and the value proposition is compelling.  In addition the vast majority of innovations in technology over the past decade have been created in or been ported to open source.

Although futures tend to be uncertain there is one prediction that appears sound.  Assets will continue to be evil, on-premise private clouds today misrepresent their true costs looking for one of those suckers born every minute, and the lowest risk, lowest cost route to success is paved by Open Source technologies.  CIO's need solutions enabling them to leverage the assets of others while maintaining all the control they need.

Saturday, May 12, 2012

The Coming Fortune 500 Ready Cloud

Cloud computing represents a seismic shift in the delivery of technology services.  By combining internal and external resources a cloud maximizes flexibility and efficiency while minimizing the quantity of assets under management.  As a result IT budgets gain room for business critical innovation initiatives and IT gains the agility required to address rapidly shifting markets.  When properly implemented cloud represents a global collaboration platform upon which to develop new services, enter new markets and transform how a company does business.  It's no surprise it's on the agenda of every CEO, CFO and CIO in the Fortune 500.
Today the power of cloud computing is trapped between promise and responsibility.  Executives used to worry about security but today they take it personally.  Sarbanes-Oxley taught Fortune 500 boardrooms about the extent Congress was wiling to go on personal corporate responsibility.  As a result CxO's scrutinize security to a degree previously unseen; a reasonable response in light of growing threats and vectors of attack.  Adoption of cloud is hindered by one simple fact: the best solutions today require giving up control to a 3rd party.  What is needed is a solution that gives the Fortune 500 assurance that they have control over the who, what, where, when, and why of their applications and only outsource the how.
There is a fast growing trend across IT that 10yrs ago would have been heresy to repeat.  CIO's are looking to exit the data center operations and ownership domain.  Taking a page right out of Nicholas Carr's paper they realize the value of IT is at the application and data layers, not the underlying sausage factory.  I expect in 2013 and 2014 we'll see a downward trend in hardware and software purchases at corporations, in particular servers, storage, and operating systems.  Purchases by providers should be lower as additional contraction is effected through virtualization so overall volumes should decrease. Not the future that the big tech vendors want, but one they won't escape from either.  For this to happen Fortune 500 CIO's need to build synergy between their infrastructure and application development teams.
Today private clouds tend to be nothing more than large scale virtualization focused entirely on delivering a stable, low cost unit called a "server" to their customer, the rest of IT.  These customers then treat that "server" as they always have: an application silo.  Very few companies are building native cloud applications, sometimes for lack of talent but most often for lack of a cloud platform.  To build a cloud application requires a purpose built Service Oriented Architecture (SOA).  Most SOA implementations predate cloud and thus lack the refinements necessary such as the elimination of traditional languages like J2EE and.NET and adoption of their light weight replacements in the forms of Ruby on Rails, PHP, Python, Scala and others.  Other changes include the adoption of HTML5 and CSS3, adopting REST to replace SOAP, coding for security and reliability, leveraging resources as services, and of course exposure through prototyping and education. 
At some point soon these two major issues are going to be addressed.  A provider will bring to the market a solution that enables large corporations to outsource real payloads without having to worry extend their trust to a 3rd party.  And a technology company or cloud provider will deliver a compelling Enterprise Private PaaS solution compelling enough to get CIO's to adopt native cloud application development as the next step in the value chain above building virtualized infrastructure. When solutions to both problems are in the market ready and available for use I believe we will have reached the tipping point at which the Fortune 500 will begin to eagerly adopt cloud. Of course it will start with a trickle, but soon an all out flood will ensue as the "space race" of Big Data and Mobility will be on.
What cannot continue to happen is getting blood from the stone. Virtualization, in the form of the vast majority of solutions available today from IBM, VMWare, Citrix, and others, will only get you so far.  The cost savings are tangible but they are also limited.  To unlock the value of cloud one has to shift focus to revenue generation and that means applications.  Applications can be built in a cloud but not in a vacuum which today is the only option for most enterprise developers.  Fortune 500 companies need cloud to compete, to innovate, and ultimately to survive.