In the past 12 months Hitachi Data Systems (HDS) has been doing some interesting work on storage lifecycle costs behind the scenes. The fruits include its ‘Storage Economics' strategy backed by a simple-to-use calculation tool so customers can see quickly where storage cost savings are likely to be made.
Was it serendipity that HDS produced such a tool and approach just as the credit crunch started to bite, or did someone see signs of the downturn early? Either way, it is exceptionally good timing.
In a presentation earlier this month Michael Väth, HDS's senior VP and general manager for EMEA, set the scene with a few pithy statements which few would dispute: "IT budgets will go down" but "data storage growth is still there" and "money for investments is scarce." So IT departments are being pressured ever harder to do more with less.
He added that consolidation was a major driver. Consolidation can help simplify management and reduce ongoing operational expenditure (OPEX). In turn this reveals one key to unlock real storage cost savings, namely that "OPEX needs to reduce [at least] in line with CAPEX [capital expenditure]." Väth said that companies should focus on return on assets (ROA) rather than ROI.
In its own survey, HDS found nearly two-thirds of IT managers needed help managing operational costs associated with their storage infrastructure; unsurprisingly, a similar percentage found it difficult to calculate these operational costs in the first place. Herein is a problem for many companies.
Ongoing running costs are invariably not carefully separated out. It was always thus but needs to change. For instance, power (electricity) usage may not be broken out across a building, let alone within an IT department or into separate segments of a data centre such as the data storage. There is actually a lot of waste, but this needs identifying if organisations are to wring out cost savings in the face of inexorably rising storage demands.
In past 12 months HDS has collated over 600 client engagement experiences from all over the world covering the past six years. It was from this that the HDS Storage Economics methodology emerged. It comprises a series of services and tools to help organisations better understand and assess their storage needs from an economic perspective.
One part of this is its desktop "Tiered Storage Economics Quick Estimator" tool which is populated with default values (variable according to country selected) taken from this research, and which drives a spreadsheet behind the scenes. It includes 33 different storage cost categories, including electricity, data centre floor space and backup server costs, as well as compliance risk and data retention issues. Straight hardware costs are in fact secondary.
In a one-to-one discussion with HDS with the tool displayed, an organisation will probably know some but not all of its data storage cost figures to apply. The defaults then fill the gaps so that an instant rough estimate of operational costs and potential savings is displayed. HDS believes it produces about 80% accuracy but, more importantly (including to HDS), it opens up a dialogue and acts as a stimulus for the client to investigate more deeply. Invariably the estimator reveals enormous potential savings.
Moreover, HDS thinks this should be a high-level discussion involving the CFO as well as the CIO or IT Director. It is something that can be done and acted upon quickly without the need for a review of all internal costs, which is itself a costly process.
HDS has also developed a chargeable one-day storage economics workshop and a 2–3 week course for those wanting to go deeper. It is an approach that every IT equipment vendor could try (although, for some, it might backfire as their customers might as a result go to another vendor to save money). Väth himself acknowledged this may mean selling less storage to customers as a result of their getting more value from their existing assets.
There is another reason why every storage user should be concerned to look at this approach. Bob Plumridge, HDS director technical marketing for EMEA, pointed out that demand for raw disk capacity was not slowing down but that the frequent migrations to higher capacity equipment were of no value to the business and would take longer through ever more data to move.
Higher disk capacities will come through new storage approaches (at an as yet unknown capital cost) but accessing them means migration. Also, disk spin speeds were not increasing, so transfer times for more data would increase.
The HDS approach going forward will be to eradicate major migrations through provision of, for instance, field-replaceable (protected) RAID, a virtualisation engine with thin provisioning, massively parallel switch architecture and intelligent virtual mega controller. The economics to users will be in incremental upgrades rather than any major rip and replace.
Ultimately, this highlights something I have been banging on about for a few years. The only ultimate solution to the data storage explosion is for an organisation to reach a point where the amount of data coming in ceases to exceed the amount of data removed from the system (either deleted or placed in off-line storage), so that on-line and near-line volumes cease to increase. It is only then that storage will become completely contained and manageable.
That does not detract from the HDS Storage Economics initiative. It deserves to be seen as a commonsense approach by user organisations.