Thin Provisioning, Deduplication, and Automated Tiering are some of the proven technologies that help optimize cost and simplify day-to-day management of storage systems. These technologies, observes Forrester Research in its Market Overview: Midrange Storage, Q3 2011, have been in the market for some time now. But they have not yet seen massive adoption.
a) Thin provisioning Alastair Cairns, Network Manager at St Mary’s Redcliffe & Temple School, was initially nervous about thin provisioning. Though the FAS 2020 SAN from NetApp and VMware came equipped with this technology, Cairns switched this feature on much later, as he was yet to understand its implications.
He realized the importance of thin provisioning, when he set out to reconfigure the storage in VMware. This required him to move a number of his servers from one Logical Unit Number (LUN) to another. In doing so, he found it difficult to move a 30GB server drive, which was not thin provisioned.
b) Storage tiering: Tiering, a key function, was a manual and a tedious procedure for a long time, points out Forrester. Auto-tiering is now gathering momentum with a slew of launches from industry heavyweights like DataCore and Dell.
Dell’s Fluid data architecture, which allows for intelligent data management, is a key feature in its multiprotocol virtualized storage solution, Dell Compellent. “It allows data to be moved automatically from tier to tier depending on its usage. Accordingly, mission critical applications move to the most expensive discs while the infrequently used and archived data gravitate to the lower costs discs. This frees up space in the higher tier for more data,” says Antonio Gallardo, Storage Business manager, Dell EMEA.
CMA, a leading provider of intraday CDS data and OTC market solutions, employed Dell’s Fluid Data technology along with VMWare virtualization. The auto tiering feature, which was initially called data progression, allowed 60% of their data to sit on the fastest, most accessible storage, while virtual application servers sit in the middle and all historical data sit on the lowest tier. “This simplified my storage management, ensured high performance and precluded the need to purchase additional equipment,” says Ryan Sclanders, Infrastructure Manager, CMA Vision.
c) Deduplication: Deduplication retains the original copy, eliminates redundant data and frees up capacity. Jeremy Wallis, Systems Engineering Director at NetApp UK, remarked that this feature helped most of his customers save 80-90% of capacity in a virtualized environment. St Mary’s Redcliffe & Temple School, for instance, gets a great deal of file similarity across its drives. A teacher creates a worksheet for a student, which leads to 30 odd copies. Another faculty member replicates this. “As a result, we get about 50% deduplication across the board. With NetApp we save approximately 50% of storage space,” adds Cairns.
As previously noted, based on six-year forecast period from 2009, Sinclair anticipated a capital savings on infrastructure asset purchase of 144,000£. “Our ROI is ahead of forecast. We are now in 2011, three years down the line, and have already saved 60,000£ [$95000]. This includes the cost of purchasing the SAN outright,” added Watts.
This year, the company also brought forward a major upgrade to their ERP systems. Due to go live in November 2011, this upgrade required 18 additional servers for hosting and testing. This project was not on the drawing board when the planning was put in place in 2009.
Had the company been on their older machines (and not used virtualization), they would have incurred a cost of 115,000£ [$184,000] on physical hardware and operating system software. But the virtual servers in the SAN enabled Watts to create the entire ERP server infrastructure at no cost. “With the 115,000£ from this project, by the end of the year we will have saved a staggering 175,000£ [$280,000] of capital solely by adopting a combination of EMC SAN strategy.” Watts now anticipates a total CapEX saving of 250,000£ [$400,000] by end of 2014.