IBM Software Defined Storage to the Rescue

by Brenda J. Christie

IBM System Storage DS8800

An interesting article on mainframe storage in this month's edition of Enterprise System Media caught my attention.  The article, IBM Redefines Software Defined Storage, indirectly discusses the cost of storage in part due to its inefficient usage.  The product, IBM Spectrum, seeks to remedy this inefficiency and was announced by IBM in February 2015.  

On a much smaller scale, the problem would seem to be similar to a personal computer hard drive in need of a compression in order to squeeze out all the space left empty by deleted files and applications;  space which went unused because it wasn't large enough for a file or application, etc. This same scenario can play out with mainframe storage.  There are several basic differences, however.  

1. There is no ongoing expense related to a PC memory 
2. PC memory is comparatively inexpensive these days  
3. Mainframe storage can be quite expensive, especially given the explosion of storage needs to accommodate data captured from social media, emails, documents, images, i.e., Big Data.  The variety and size of system software used to manage mainframes today has also grown as has the use of third-party software installed on the mainframe.   Examples of this would include non-IBM Analytics software, Hardoop, BMC Subsystem Optimizer, and BMC Cost Analyzer.  

Given these realities, i.e, storage is expensive, programs use more memory, and in light of other costs, such as MLC which are revenue generators for IBM, it makes sense that IBM would provide its on solution to growing storage cost concerns.  It has done so with the announcement of Software Defined Storage.  Software Defined Storage, SDS, on one level, acts like an air traffic controller for storage management.  It purportedly, through the use of analytics,  will efficiently determine the type and size of data and where it can best be stored for retrieval as well as backup purposes.  IBM maintains that as much as 90% of storage cost can be reduced through optimum placement of data managed through SDS.

How much storage costs are reduced is dependent upon the choice of storage device as well as its configuration.  The possibility of reducing storage costs will no doubt be of interest to datacenter management as will the ability to quickly deploy enterprise storage for cloud.

Another interesting aspect to this product and how it is being licensed is that its model includes a pay-as-you-go feature in that it can be leased on a month-to-month basis or as long as it is needed.  

Summary

In summary, it appears that with the launch of the Software Defined Storage product, what IBM is is trying to do is twofold.  First, it is looking to capitalize on the inevitable demand for storage due largely in part to Big Data -- a growth which is accompanied by inevitable cost.  Secondly, in responding to that demand, it will also increase revenue from software licensing.  It may also, in the process, increase revenue from sale of its own storage.  With additional freed-up space, IBM may also license even more software.  Overall SDS and the possibilities it presents represents an opportunity to increase the Total Benefits of mainframe ownership.

Bye for now,

Brenda J. Christie