Disk Products Manager
The current struggle to control storage growth consumes not only monetary capital but human capital as well. It is a challenge to stay ahead of growth requirements and integration of new hardware into the current infrastructure. Over time, a significant ongoing effort to ensure efficient and effective storage management is required, and it evolves constantly.
One phenomenon that has been ignored but contributes to a significant and growing percentage of these costs is the proliferation of data copies. Production data is copied for several reasons: local protection from data issues such as database corruption; remote copies for protection from disaster; and many copies used for testing, development, QA certification, BI/BA and so on. All these have contributed to an amazing statistic: last year the cost of data copies surpassed the cost of the original production data. Why is this surprising? Traditionally we think of data copies as requiring only the least expensive storage — tape, SATA drives, etc.
So how do we deal with this phenomenon? By repurposing some highly effective concepts that we have already seen succeed. One is server virtualization. Beginning with the IBM mainframe in the ’70s, processing images began to be separated from the hardware they were running on. The multiple virtual systems (MVS) operating system was the first viable OS to virtualize processing images. Others went before, but MVS prevailed. Logical partitioning (LPAR) images are still the de facto standard today. Modern systems such as VMWare owe their existence to MVS.
Storage virtualization followed next. Again starting with the mainframe, the connected storage layer was abstracted from the OS, which allowed non-IBM vendors to participate in the market for mainframe storage. It doesn’t matter whether you connect an IBM, HDS or EMC device to a mainframe, the OS sees just a virtual volume image. Technologies later appeared through software (FalconStor) or hardware (IBM San Volume Controller) to provide abstraction of the storage layer. So just as with the mainframe, Windows/Unix/Linux platforms can now attach to different storage models with one toolset and one management interface.
The next frontier is virtualization of specific challenges to gain the same benefits of efficiency and simplicity that other virtualization environments created. The data copy challenges outlined earlier can be tamed through copy virtualization. The leader in this space is Actifio.
Actifio’s value lies in the dramatic reduction in data capacity through virtualization (up to 90%), the speed of the data copy/recovery process (minutes, not hours or even days) and the simplicity of the SLA construction process.
To set up an online demonstration of Actifio’s capabilities, please contact me or your Mainline account executive.