The classic work of English historical literature “The Decline and Fall of Roman Empire” is a book written by the English historian Edward Gibbon, which traces the trajectory of Western civilization from the height of the Roman Empire to the fall of Byzantium. I’m using this comparison to bring to light discussions that my team and I have conducted over the past few years on the topic of as- a-Service (IaaS, PaaS, SaaS, etc.) in the Public and Private Clouds.
Shifting your workloads to the cloud, whether public or private, looks attractive in a number of ways. You conceptually see the gains from quick and readily available infrastructure by just clicking a button or two, your service from a new virtual machine in the cloud appears ready as you need it. The initial gains materialize as on-demand capacity, high availability, and disaster tolerance to name a few. What about the costs of building all of this and has anyone ever seen a positive gain? Has anyone really seen a gain from IaaS alone?
Public and Private cloud services models are still maturing but the overall question that we are hearing, is it worth it? We’ve come across several articles that look at the features, and functions of as-a-Service offerings (to include PaaS, IaaS, SaaS, etc) along with theoretical return on investment (ROI) of each. What we have seen is the shift in focus that sole IaaS eventually plays into higher delivery models like BPaaS (Business Process as-a-Service), or SaaS (Software as-a-Service) etc. Of course the message is different between Enterprises and Service Providers where this could help focus more reliable revenue flows for Service Provider’s and a more deliberate approach for Enterprises.
In the months spent researching this, we never found a definitive paper or published research outside of system integrators or service providers that had actual projected financials for SP or Enterprise. Also, given the financial calculations were heavily weighted on the ROI models from specific vendor equipment vs any diversity in mixed infrastructure environments. In further calculation of the costs for IaaS, requirements from Service Providers or Enterprise do not involve simple scenarios where the predictable medium based Virtual Machine would suffice as a definitive control point for those calculations. We’ve seen the requirements need to align in the form of complex workloads such as database and transaction processing that require more robust, and more expensive IaaS-class VMs within diverse infrastructure, distributed about multiple tiers. Regardless of the requirements category, multiple small scale and diverse control projects are needed to gather precise cost, performance, and availability metrics to validate the real cost and ROI IaaS models. IaaS, for the most part, has to increase it’s service offerings to go further into areas like Virtualized Desktops (VDI), offer enhanced security for data, and potentially pay-per-use capacity on demand services just to name a few. At that point, IaaS is moved from it’s rudimentary form to more of a superset like PaaS, BPaaS, SaaS, etc. One thing to keep in mind, PaaS is very closely associated to the lower end services similar to IaaS where it’s monetization and revenue generation is almost identical.
In summary, we see the Cloud is here to stay but there is a decline in the need for just a simplistic offering of services beyond what is IaaS. The enhanced subset of services must move away from IaaS to be more like BPaaS, SaaS and other models to cost effective. Businesses, whether SP or Enterprise, are going to leverage those services in their market and effect significant changes in the way they operate. The budgets that once filled groups of individual business units (speaking in the context of the Enterprise) to accommodate for their own IT presence, are now consolidated to larger capital and revenue budgets for enhanced IT subscription services that go far beyond what used to be just cloud infrastructure.
Authors:
Marc Buraczynski, Solutions Architect, Cisco Systems, Inc, Cisco Advanced Services (Boston, MA)
Chris Shockey, Solutions Architect, Cisco Systems, Inc, Cisco Advanced Services. (Seattle, WA)
Nice write up Marc and Chris