Cloud computing offers many benefits for the masses, but what does it have to offer a business, a research institution, or even an individual with specialized infrastructure needs?
The most important resource since the invention of the computer has been computer time. Cloud computing provides computational benefits in terms of sheer power and affordability that make sound economic sense.
As far back as the beginning of the public availability of computer resources, there have been business models based on time-sharing and batch processing schemes. Even remote connectivity to the mighty mainframe was possible, with IBM and Bell Labs leading the field in the early 1970s.
Time-sharing is the concept that computing is a utility no different from electricity or water. According to this view, computing power can be sold in slices or blocks of time. A consumer can purchase a contiguous block of time, usually billed by the hour. This model is good for planned computing needs, where the duration of specific tasks can accurately be estimated ahead of time and budgeted for.
A popular variation is to sell computing time on an ad hoc basis, usually measured in seconds. This model is especially suited for computing tasks that include data processing that emphasizes calculations on the data. For these tasks, the input data can cause the length of the computing time needed to complete the task to vary greatly. In such cases run time estimates of computing needs might be too inaccurate, and a free-form model is better suited.
Batch processing is the concept that a computational task can be broken down into subtasks. Every subtask may or may not be further broken down into more subtasks, but at some point a task can be resolved to specific procedures or steps. Such steps may be either standard or custom computer programs, or may be part of the basic services provided by the operating environment itself.
A task thus becomes a sequence of steps that can be prepared ahead of time, along with any required data, and then executed. The consumer can purchase computing time based on the number and type of steps in the task.
This enables the cloud computing provider to schedule the use of its resources ahead of time. In this way, the provider can maximize the overall utilization of its facilities, which also maximizes profits for the provider.
Today the new term “cloud computing” is used for old ideas that have never really lost their popularity, nor their practical necessity.
It is very expensive to attain the lofty level of raw computing power that the big names in cloud computing currently offer.
There are economic factors that deter most organizations from building their own powerful supercomputer systems. These include:
- The large initial capital outlay on computer hardware
- The cost of people with expertise in designing and building such a complex system
- The ongoing maintenance cost, which includes parts and labor
- The ongoing operational cost, which includes electricity consumption
There is a clear need for the new computing power utility companies: those that run the cloud.
Any organization with large but infrequent computing processing needs can rent computing power for a specific length of time, or for a specific task. The cost benefit compared to actually owning and operating specialized hardware – which would be underutilized for the most part – is obvious.
The business product that offers high-end computing resources for rent is known as “Infrastructure as a Service,” or IaaS. All of the resources of a computing platform, including storage, memory, networking and processing power, are offered as a unit to the consumer. The level and the duration of resource consumption determine the rental cost.
Data processing and computational power demands are very high already, and are steadily rising. The role of computing in business simulations, research simulations and data analysis is a crucial one. The more computing power available, the more detailed a simulation can be. More data can be analyzed in the same period if more processing power is applied.
It therefore makes sense to have a specialized source of computing power that can focus on providing the very best possible infrastructure and computing resources, just like it makes sense to have a specialized source of electricity. A specialized electricity provider can provide a much more consistent, safe and high quality resource than any other non-specialized organization could.
In the same way, the specialized IaaS provider can offer much more powerful, scalable and highly available computing resources. Few organizations can match this level of quality and affordability.
Cloud computing and IaaS will be the enabling factor for the next stage of civilization’s growth.
The impact of potential business strategies and other scenarios can now be easily simulated and analyzed by most businesses.
Medical and scientific research, which can only be effectively pursued using massive computational effort, is now available to organizations of even moderate financial means.
Using cloud computing as a computational resource makes sense, and can provide undeniable benefits for businesses and other organizations.