The cloud-based hosted services market is an exciting and fast moving space that has received a lot of attention in recent years for having the potential to completely shake up the technology infrastructure of businesses. The space is evolving so rapidly that it’s hard to keep pace with what is what, with new acronyms and even business models seemingly popping up all the time. Several key trends have converged to really change the dynamics of the game. First, the processors and other HW components of servers have evolved, with multi-core processors and multiple processors/server the standard today, causing the computational capacity of individual servers to explode. Second, the emergence of virtualization software has made it possible to get even more out of servers by partitioning the compute capacity of servers into smaller virtual chunks – called virtual instances – which can be leveraged by different applications, or even different businesses. Third, the network bandwidth available to end-users of applications – a historical bottleneck in running applications from servers in the cloud - has skyrocketed in recent years, making it possible to run cloud-based applications from almost anywhere, even when using a wireless connection. Finally, and not unrelated to the above changes, larger players have rolled out cloud-based services (e.g. Amazon, Google, Microsoft) that offer global footprints - and a sense of stability and credibility - that smaller hosted providers historically fell short of. Taken together these trends will eventually transform the IT landscape as we know it today.
Three general classes of hosted services have emerged, ranging from leasing servers that are housed off-site - which is similar in many ways to the historical “managed” approach - to the provisioning of end-user applications that directly run off servers hosted in the “cloud”. Leasing the raw server infrastructure, or portions of servers, is generally classified as Infrastructure as a Service. This is more the domain of IT professionals looking to run servers as in the past, but from the cloud, with the need to configure the servers and specify technical workload demands and activity requirements. Amazon’s Web Services offering is a good example. The second class of services is called Platform as a Service, which is distinguished by a richer set of services for building applications in the cloud - typically targeted specifically at developers of applications. The PaaS offerings are intended to be more turn-key in nature, with a deeper stack of services and utilities, and often with ties to specific programming languages. Microsoft’s Azure offering and Google’s AppEngine may be the best examples today. Finally, and perhaps most well-known, is the class known as Software as a Service, which includes standard canned applications that historically have run on local computers, but in this case are run off servers located in the cloud. There are many applications available, with Saleforce.com’s CRM software and Google’s Gmail email offering two of the most famous examples. The research reported here actually cuts across all three arenas, including back-room technologies such as servers, business applications and file storage/backup, as well as end-user tools such as email, communications, productivity, and collaboration applications.
