To have experienced this October’s political activity in Washington is to know the meaning of the term “budget uncertainty.” Over a prolonged period, budget pressure appears to be pushing some agency IT acquisitions off course.
I’ve seen several recent requests for quotes and requests for information come through in which agencies ask vendors, generally systems integrators, for fairly traditional IT assemblages – the kind that are so often late and over budget. But here’s the twist: Agency specifiers, under pressure to adopt the cloud, intend to acquire the resulting systems as-a-service.
Acquiring traditional custom systems integration and overlaying it with an as-a-service buying model might be a way to avoid high upfront costs and control costs. But this as-a-service acquisition approach is based on a sort of IT fallacy: because cloud equals as-a-service, then as-a-service equals cloud. This is understandable, but wrong. Here’s why…
Whatever development and deployment costs a vendor incurs in building and maintaining the system or service, unique as they are to a given agency, will eventually find their way into the lifecycle cost. Final costs will be higher than those of true cloud implementations, even before the inevitable inflation from change orders and poorly-defined requirements.
Agency managers may imagine that by specifying development-as-a-service, they’ll get the cost benefits of cloud computing they’ve heard so much about. But, you get those only from implementations with the important attributes of cloud computing. Whether you are looking at infrastructure-, platform-, or software-as-a-service (IaaS, PaaS and SaaS), those attributes include these two features:
Particularly on the applications/SaaS side of cloud, custom is precisely what you can’t call “cloud” by definition. While you might acquire custom development or operations and maintenance “as-a-service,” with charges based on user headcount, that does not mean you are gaining the attributes of cloud computing. Under the SaaS model, agencies buy seat licenses for applications used by multiple tenants, with thousands and thousands of users sharing the application stack, and benefit from the continuous cycle of innovation and improvement that is driven by the scale of the shared application.
Broadly used application services – such as early SaaS products like Google Apps – have become tremendously more capable and powerful than when first launched, without additional fees for upgrades. SAP as-a-service, or similar enterprise applications, of course are uniquely customized or configured but large savings can still result from shared cloud management and maintenance of the platform (think PaaS). The per-seat license prices are driven lower, and more importantly, they are rendered predictable. The software environment itself continuously improves based on the specialized but shared expertise of users.
Federal agencies have waded into cloud computing for IaaS, SaaS and PaaS, but they can be forgiven for moving slowly. The vendor competition is fierce, the “everything is cloud” hype is confusing, and several big awards have been held up by protests. That will settle down. But agencies shouldn’t confuse the desire for “everything IT as a service” with true cloud computing. Both are good and maybe inevitable. But you leverage the economics of cloud computing differently.
Unisys is innovating an emerging model of cloud brokerage where the customer can choose from among cloud vendors, as well as their unique internal IT services. For example, we’ve developed an IT service management- (ITSM) as-a-service to give systems administrators single-pane-of-glass control across a hybrid infrastructure. It’s built out of best-in-class software components integrated and running in Microsoft’s Azure Government Cloud. This is essentially the same tool set we use to manage large commercial IT environments handling millions of trouble tickets as-a-service.
These examples show that by becoming trusted, value-added brokers and offerors of true cloud software, the traditional systems integrators can bring the benefits of cloud computing to the federal market. And to deliver that value you have to understand that everything “as-a-service” is not cloud computing. If we think differently and fight through the hype, we can do a better job with fewer resources.