Welcome!

Brad Vaughan

Subscribe to Brad Vaughan: eMailAlertsEmail Alerts
Get Brad Vaughan via: homepageHomepage mobileMobile rssRSS facebookFacebook twitterTwitter linkedinLinkedIn


Related Topics: Cloud Computing, SOA & WOA Magazine, Microservices Journal, Amazon Cloud Journal

Article

AWS and the Public Cloud Market

Amazon’s continuing role as a leading provider of off-premise infrastructure is almost guaranteed

The public cloud market is dominated by a 800 pound gorilla. We assume its 800 pounds and not 2000 pounds. It could even be 100 pounds or maybe not a gorilla, but an elephant of a paler color. There is no public information indicating its true size, revenue or performance. Let’s assume based on the marketing information it's 800 pounds and a gorilla.

The recent release of the dedicated instance feature within Amazon Web Services has shone a spotlight on a possible future of public cloud. A proof point that cloud and the traditional hosting model might exist on the same continuum. I think all along this was inevitable. To think that the Walmart of infrastructure service providers could possibly service the needs of all use cases was maybe a touch naive. As the Clouderati discussed on twitter, it might be technically possible to construct a commodity cloud service in such a way to support all application workloads, but the technical challenges are never the primary motivator for enterprise infrastructure decisions.

Amazon’s continuing role as a leading provider of off-premise infrastructure is almost guaranteed. It has almost single-handedly broken the back of the inertia that had dampened the hosting business. By leveraging the technologies of abstraction and applying a business model that defined elasticity, it has added a value proposition to hosting that has re-invigorated interest. In some ways it is disruptive and paradigm changing for IT. The impacts will be seen in hosting and on-premise infrastructure and development models.

For all its size, however, it will not be the monopolistic entity Google is for search or Microsoft for desktop. To achieve this, it must bridge the gap between consumer (individuals developers purchasing resources) and enterprise adoption. The reason for this is based on three (3) aspects of the adoption of cloud in enterprises;

  • Purchase decision - the way products are purchased by consumers and enterprises is very different, you cannot be all things to all people
  • Product complexity - if you scope the product as all inclusive (including services etc..) and not just components (ie. compute, storage etc..) then it is complex and not commodity
  • Business model – the design of the business model around economies of scale limits its ability to service certain market requirements

Purchase Decision
One of the largest descriptors of an “Enterprise” is in the purchasing decision. Consumers purchase as individuals and Enterprises decide based on the needs of a range of people. Bake-off, tenders, approved vendor lists, ROI, compliance are the recipe of a enterprise decision. Its impossible for a “low-touch self-service model" to address these needs in a comprehensive way.

Some people will throw iPhone/iPad or desktop computers on the table as a indicator that consumer technologies are infiltrating the enterprise. This is true, but does not apply to cloud. The devices that bridge the gap between consumer and enterprise computing are highly personalized and have use cases that span both work and personal use.

The other argument is that enterprise needs to change to be agile, and some of that will happen. Using the past as a predictor of the future, not all companies will make this transformation in the same way. This change will happen along a continuum, and along that range will be service providers to fill the needs. As much as open systems was a disruptor, client service was a disruptor and web services were a disruptor, we still have mainframes, forms-based client server apps and a  range of other technologies.

Product Complexity
Although compute or storage or network can be refined down to a building block that could be described as a commodity, none of these things is actually constitutes a service. The product that an enterprise purchases is a combination of these things, wrapped up with support, service commitments, compliance and a range of other features. These things are not a commodity and represent a complex product. When you find a complex product, then you will find a distribution of architectural (right) and service requirements. It’s difficult to know if Amazon represents the center of the distribution. Will it continue to dominate, or is another wave of adoption going to push it to the edges?

Business Model
The final aspect that is shaping the public cloud market for the future is the business model. The business model for Amazon is massive scale. The “dedicated instance” offering actually shows that Amazon might be backing away from that business imperative to attract different markets. It also might be a way to provide enterprises a quantitative way to measure the ROI of using a classic cloud model by putting a premium price on the dedicated instances. It certain shows recognition by Amazon that enterprise requirements might differ.

The business model of massive scale and the development of products designed for economies of scale has basis in a service provider's capability. If you have investment in a massive asset, you must keep it utilized to maintain profitability. It's very similar to the outsourcing services model for IBM or the big five consulting firms. Once you have the headcount (or servers), you must continue to feed the beast to stay on top. Economy of scale is a fickle partner. The true test will come when aggressive growth flattens and the flexibility of the asset lifecycle and the architecture's ability to service different business models will be tested.

In the end, there will be multiple public clouds and multiple types of providers for a range of customer needs. It is true, as Paul Miller of GigaOM writes, “ public cloud must surely be the most viable way forward”. For providers to note, as a cloud's scale grows, it ceases to be able to cater to the unique requirements of specific workloads. This will place even more importance on interoperability and portability.

The ability to service vertical industries, optimize infrastructure for workloads and adopt a range of business models still has its place in the public cloud. In the wake of Amazon, the Rackspaces, GoGrids, Savvis etc.. will find their place and their flavor of cloud.

More Stories By Brad Vaughan

Brad Vaughan is a twenty year veteran consultant working with companies around the globe to transform technology infrastructure to deliver enhanced business services.