More flexible software licences needed for Utility Computing
Tap on. Tap off
The basic ideas behind utility computing, also known as computing on demand, are very simple, writes Tony Lock. At the core can be found the concept that today's rapidly changing market conditions place a requirement that every organisation's IT infrastructure needs to be flexible in order to satisfy the fluctuating demands for service. In other words, IT must be able to support its users whatever the workload whilst still maintaining tight control over the cost of service delivered.
One approach to supplying service in such potentially variable loads is to build an infrastructure scaled to meet peak demand. Such a solution may, however, entail a significant cost penalty.
The idea behind utility computing is to allow organisations to allocate sufficient resources to meet the service level requirements of users as the demand changes. When more users require use of the system, the infrastructure allocates more resources (scales up) and when demand lowers the system scales back down and releases any excess capacity.
Clearly any attempt to implement such systems demands that all of the solution components be capable of operating in such a flexible fashion. To this end, several of the leading vendors have produced a range of tools and products designed to operate in such on demand environments. Typically such offerings have included server engines, storage systems and sophisticated management tools to allocate and de-allocate resources in response to varying user demand.
At this moment in time a number of vendors, including IBM, Hewlett Packard, SUN, Dell and a number of others, can supply IT infrastructures capable of operating in such an on demand fashion. However, while these systems are very good at allocating and de-allocating physical resources as and when required and can manage the application software deployed on these systems, there is one area of cost associated with this type of operation that has received comparatively little attention, namely the licensing of the software applications and tools deployed.
Currently most of the software vendors operate on models whereby they 'sell' licenses for each product. The licenses concerned may function in a variety of fashions, but most often operate on a 'per server', 'per CPU' or 'per user' fashion. Some applications may function on a 'concurrent' basis whereby the license covers simultaneous use by a specified number of users rather than of individually identified users.
In order to function cost-effectively, utility or on demand computing requires that software licences be made available on the fly and that consumers pay only for the software while the licenses are in use rather than for periods when the infrastructure has been scaled down.
Independent software vendors need to address the issue of licensing by usage in the very near future if the development of utility computing is not to grind to a halt. © IT-Analysis.com
Sponsored: Hyper-scale data management