Types of Cloud Computing Including Consideration of the Private Cloud
Posted Apr 13 2011 12:00am
I have been commenting about the feasibility of cloud computing for more than three years . The technology has now reached a state of maturity such that CIOs, at least outside of healthcare, are seriously considering deploying this architecture. Healthcare computing tends to trail that in most other fields so most hospitals continue to run their own "machine rooms" with dedicated servers. A recent article provides some useful definitions of the various types of cloud computing (see: Here's Why Cloud Computing Is So Hot Right Now ). Below is an excerpt from it:
What is cloud computing?At its simplest, cloud computing means that users are connecting to applications that run on a set of shared or pooled servers, rather than running on a single dedicated server. This is a subtle but important change from the client-server computing that has dominated IT for the past 20 years, where each application was "assigned" to a particular piece of hardware in a data center. If that hardware was down, the user either had to connect to a backup (which had to be standing by and ready) or would suffer an interruption in service. In the earliest instances, these shared resources were physically located away from the company's premises, and users would connect to them over the public Internet....So the "cloud" in cloud computing was the Internet, which is often represented with a physical cloud on software design and networking charts. But in the last two or three years, vendors like Microsoft and IBM have been pushing the idea of "private clouds." These are sets of servers run by large companies or government agencies for the exclusive use of their employees, who usually connect to them over a private area networkTypes Of Cloud Computing
Software as a service (SaaS). These are end-user applications, like productivity or business management software, that run on a set of pooled hardware resources and are accessed over a network. Pioneers here include Salesforce.com's CRM and business management services and Google Apps, as well as startups like Zoho, but other companies have recently gotten into the act -- particularly Microsoft, with its Office 365 services. SaaS is usually contrasted with products like Microsoft Office, which run on a user's PC.
Platform as a service [PaaS]. Here, service providers offer a set of application "components" which developers can put together into applications....Examples here include Salesforce.com's Force.com platform and Microsoft's Windows Azure platform.
Infrastructure as a service [IaaS]. Sometimes called "utility computing," in this case a developer builds an application from scratch -- no building blocks are provided. But when it comes time to run the app, they run it in a virtual machine (software that imitates a dedicated server) rather than on a dedicated physical machine. That means that as more users need the application, the provider can quickly easily serve them by spinning up new virtual machines. Amazon Web Services, Rackspace, and Xerox (through Affiliated Computer Services) are among the players who offer this kind of cloud computing.
I continue to have serious reservations about the idea of a private cloud referred to above. To me, it does not seem to differ greatly from the client-server architecture that has been the norm in recent times. In an organization such as a health system, organizational power of an executive is a function of three things: the number of FTEs, the size of the budget, and space. The pursuit of cloud computing will result in a diminution of all three for a CIO. Hence, this is not a strategic direction favored by most of them unless they are under intense pressure to reduce costs.
However and as noted above, let assume that the local IBM rep comes to a CIO with the following proposition: "Let's call your existing set of servers in the machine room, after some tweaking by IBM, a private cloud. After we are done, no one will bother you any more with this nonsense." The essence of cloud computing is that economies of scale can be created by the transfer of computing functions to huge, external server farms that support a multitude of users. Also and as noted above, this model can be theoretically achieved by large multinational corporations and governments but certainly never by an individual health system.