In-memory computing can boost both CPU-bound and IO-bound jobs

In enterprise computing, usually there are CPU-bound and IO-bound jobs. In CPU-bound jobs, CPU utilization is high while IO tasks are minimal. In IO-bound jobs, CPU utilization is low because the jobs are mostly waiting for IO operations to start and finish. As we know, disk-based IO operations are several magnitudes slower than CPU-memory operations.


Good in-memory computing technology directly impacts performance of CPU-bound processes. You may then ask this question: how does better in-memory technology can have any impact on IO-bound jobs since CPU is already under-utilized and it is the IO that is the bottleneck? 

Look at this way you will find the answer: advances in data storage technology. Storage technology dictates how fast IO operations can be performed between CPU and where data is stored. The gradual adoption of Solid Sate Flash Memory as data storage mechanism by enterprises is boosting IO operations dramatically. IO-bound jobs are less IO-bound now. Some IO-bound jobs may have even shifted to CPU-bound.

Recognizing bottlenecks in your enterprise computing is crucial now since you are dealing with big data. The point is that good in-memory computing technology and good storage technology all play important roles to improve your enterprise productivity.