Hardware - IBM mainframe On/Off Capacity on Demand

zAAP
Integrated Facilities for LINUX (IFLs)
IBMs On/Off Capacity on Demand Offering
Hipersockets

IBM Mainframe On/Off Capacity on Demand
Temporary scheduled or unplanned increases in capacity are a headache IT managers and planners have long struggled to deal with. The need to avoid expensive mainframe upgrades for what may be a passing peak conflicts with the urgent business need to keep up and running, meeting the demands that will fuel profit and growth. Failure to cope with these peaks can be costly, both in terms of current business but also in terms of potential future business.

IBM offer one solution that is well worth considering. It is On/Off Capacity on Demand and the best way to explain it is to reproduce the following quote from the IBM FAQ:

"customers can choose to turn on temporary processor resources to help meet e-business capacity growth, whether predicted or not. Some e-business may encounter predictable peak workload demands on a temporary basis. By choosing to purchase the On/Off Capacity on Demand right-to-use feature, customers have the ability to turn temporary processing resources to handle spikes in workload demand, and turn it off when normal workload levels resume."

And it really is as simple as that - you turn off and on processors as and when you need - you only pay for the days they are switched on.

On/Off Capacity on Demand gives you temporary access to extra processing units (general CPU, Integrated Facilities for Linux (IFLs), Integrated Coupling Facilities (IFCs), System Assist Processors (SAPs) and zSeries® Application Assist Processors (zAPPS). Of course if you feel or know that the increase in demand will be permanent then you can opt for a Permanent Capacity on Demand upgrade i.e. choose the extra processors you need to configure the capacity and power needed and have them available permanently.

From a zLinux On Demand perspective, it is a particularly attractive feature that IFLs are so scaleable. A number of industries (e.g. retail) know well that peaks and troughs, heavy reporting periods, and peak transaction loads have been dealt with well by simply adding IFLs to existing Linux Engines. See our section on IFLs and if you are not running zLinux give some serious thought to doing so, particularly the strategic DB2 Connect / zLinux option which is considered further.

Continue: Hipersockets >>>