My last post was one in a series that looks at Cloud, both in terms of technology evolution and also in terms of a business model that is highly disruptive. However I left questions about the road ahead unanswered. Such questions are almost always addressed in terms of technology adoption roadmaps. This clearly has great value, but risks missing the purpose of going Cloudy, i.e. to enable organizations to be more productive, efficient and agile; to be more competitive; to focus on their core differentiating value and/or executing on their mission. What we tend to overlook is a discussion around how to decide which IT services (and thus which business processes) to source where, when, and why. Defining such a strategic technology roadmap requires a clear business strategy that takes account of both the organization’s goals and the technology opportunities available. The emergence of Cloud has fundamentally broadened those opportunities.
So let’s rewind a bit…
The ongoing disaggregation of applications into services is equivalent to the disaggregation of business processes. This allows organizations to focus on the processes, and the IT services, that differentiate them and affords the opportunity for them to source non-differentiating IT services from somewhere else. That somewhere else is now the Cloud, and it is the Cloud because of a set of very compelling attributes from the perspective of the consumer of those Cloud services, whether they be Infrastructure (IaaS), Platform (PaaS) or Software as a service (SaaS). Very specifically Cloud services are ubiquitous and generally relatively cost efficient with a transparent, or pay per use, cost model, self-service and instantly provisioned, scaled up or scaled down. These attributes are enabled by economies of specialization. These economies are possible as the Cloud providers are able to massively automate the management of their services because they only manage a very tiny number of service patterns. This combined with the use of virtualization, and in particular server virtualization, allows those services, and also operating system instances, to be delivered very cost effectively (through co-tenancy) and to be very rapidly provisioned. Finally, the mere existence of the external, public Cloud providers has fundamentally changed the expectations of organizations, with respect to the delivery of IT services. They now know that these can be delivered rapidly; that they no longer need to wait nine months to get a new service into production. They know IT services can be delivered almost instantly, and they now expect that of their internally sourced IT services, even if their service level expectations are substantially different. So now the IT organization has to provide internal, or private Clouds, i.e. IT services that have the same attributes of transparent cost, speed of provisioning and self service, that public Clouds have demonstrated.
Let’s build on the assumption that larger organizations only want to invest and spend money in those areas of IT that enable them to differentiate themselves, to execute on their core mission, and/or to be more competitive. It follows that in general they will seek to acquire non-differentiating IT services as commodities.
Since infrastructure and packaged applications (HR, payroll, email etc.) do not generally differentiate an organization, unless providing these services is their core business, these IT services will tend to go to the Cloud first. This is reflected in the adoption of private/hybrid/public Infrastructure (IaaS) Clouds and of Software as-a-Service (SaaS) solutions, as a means of saving expense. These savings may then be redirected or reinvested elsewhere in the organization, if so desired. Of course short to medium term agility benefits will also accrue, especially for private IaaS clouds where services can be rapidly deployed and scaled, but eventually, as adoption becomes pervasive, the competitive advantages this bestows will diminish.
The true, long term advantages in Cloud eventually present themselves through Platform-as-a-Service (PaaS) implementations. Platforms allow the rapid development of new services, minimizing the time to delivery or time to profit. By their nature they free the developer from having to think about plumbing (building communication, persistence, scaling and availability mechanisms etc.) and allow them to focus on delivering value. Such platforms also provide an opportunity for organizations to expose their shared services that differentiate their organization, allowing internal and external partners to rapidly innovate and deliver new value.
The best way to understand this is to use a hypothetical example. Let’s say one of the world’s larger e commerce sites chooses to expose many or most of its APIs that it uses privately, as public APIs. This makes it a platform that other people can leverage. They can build new value, without having to worry about basic e commerce services like searching for goods, or making payments. After all, these capabilities were created by the first movers in this space, and differentiated them in the early marketplace. These same first movers now provide performance and scaling, at a cost that no entrant to the market could hope to match. It makes sense for them to expose these e-commerce services as a platform that others can leverage. This encourages the creation of an eco-system that allows the platform provider to benefit from the innovation of others, as well as themselves. And if their eco-system builds a better general-purpose e-commerce site that eclipses their own, they would still get a share of every transaction.
This is one example of creating and exposing shared services as a platform. If this is done, specifically as a PaaS, it allows an eco-system to be quickly created that enables more rapid innovation. This can be done within organizations in the form of private PaaS, or publicly via a public PaaS. This is the place where one could argue, significant investment will yield massive opportunities – for savings – by having shared services across multiple units within an organization, and for profit or other benefit, through reduced time to deliver new services, or value, on top of the shared services, exposed through a PaaS.
Next… Deciding How or Where To Source/Host Services