Myth #1: Storage is Getting Cheaper
While it is true that disk storage per Terabyte has been declining by 35% to 40% per year, the average amount of storage installed has increased by about 45% annually during the same period. In fact, storage is the fastest-growing capital cost within the data center. Research from Gartner’s Enterprise Storage Management service has found that storage is increasing, on average, about 45 percent per year. Based on this level of growth, storage capacities double every 18 months. For many datacenters the growth rate is in excess of 100% per year. Yet only 65 percent of enterprise storage is effectively utilized.

Myth #2: It’s Cheaper to Add Storage Rather than Manage It
The growth in midrange storage levels far exceeds that in mainframe environments, thanks in no small part to the appeal of this myth, yet the ability to manage and leverage storage is usually nonexistent. Ironically, a significant percentage of installed storage is unused or unavailable; RAID overhead can consume 20% to 50% of the installed capacity for event recovery and protection from device level failures. It’s also important to remember that the greater the storage volume, the greater the risk of data corruption, lost files, and backups and disaster recovery complication issues.

“The greater the storage volume, the greater the risk of data corruption, lost files, and backups and disaster recovery complication issues.”

But the real issue is uncontrolled expansion. Investments in storage management tools, automation and storage administrators have certainly not increased at the rate of storage capacities – as much as 45% per year in midrange environments. An ever-expanding base of storage is being managed by a smaller proportion of storage administrators. The results are reduced availability, security exposure and potential data corruption

Myth #3: Capacity On Demand Is A Great Idea
Capacity on Demand (COD) is a good idea for applications where there is extreme volatility in storage demand or where a sudden out-of-space condition would send a mission critical system crashing. COD gives the datacenter access to “just-in-time” storage already installed in the storage environment, waiting to be used. Emergency and peak storage are readily available.

COD is better for vendors, who are able to house storage devices at the client’s site instead of in their own warehouses. It’s likely that the storage is not the latest technology either, especially in multiyear contracts. So the vendor is essentially able to sell a datacenter older hardware without having to pass along the annual 35% to 40% decreases in disk pricing that Gartner sees. Even if the vendor offers a COD pricing discount of 10% to 15%, the datacenter that relies on COD will be paying a premium for that capability.

Myth #4: Backed-up Data is Always Recoverable
Many datacenters are diligent in establishing backup processes, rules and procedures. However, without regular recovery testing and audits, nobody knows if backups are really recoverable. Gartner estimates that 20% to 30% of current backed-up data is probably not restorable due to hardware and/or software failures and technology obsolescence. Snap copies and remote replication have been effective where deployed, but they carry management and planning overhead that must be considered, and for long-term data retention, they fall under the same constraints as traditional tape or optical storage.

Myth #5: SAN Storage Will Lower My Costs
SAN implementation can reduce costs for high availability, operations labor and facilities. Hot-recovery space is easier to set up and deploy than in direct-attached storage (DAS) environments. Operations labor can be reduced since storage and backups can be centrally managed. Smaller overall footprint reduces facilities costs although this is a minor portion of the storage TCO.

On the other hand, SAN is far more complex to configure, manage and maintain than DAS. Supporting this environment requires more highly skilled (and expensive) storage administration resources. SAN will drive higher costs for hardware, software, and connectivity, but once these investments are sunk, the SAN infrastructure is far easier to grow and utilize than DAS.

Myth vs. Reality
Each of the myths we’ve explored is influencing the direction of storage implementation in midsize businesses. The idea that it is cheaper to add storage than to manage it, for example , distracts IT managers from a deeper analysis of storage effectiveness that could suggest a solution based on reality, not dogma. In the same way, SAN and NAS, or FAS (Fabric Attached Networks) as they are collectively known, are the only storage technologies capable of effectively sustaining the massive storage growth we see today. But the truth is that while incremental costs are less, total FAS cost of ownership trends higher.

This Case Study covers the innovative use of block-based SAN virtualisation and replication technology – delivered as a service – to support the logical data migration to a new data centre. The logical data migration service had to integrate with the overall data centre migration which included the physical move of all servers and their associated applications. A key part of the solution was the use of “IO splitter” technology in the SAN to allow in-flight replication without any integrity risk or change to the existing production disk volumes. The solution incorporated a centralised application and configuration database, developed for the project, which managed the logical data migration workflow, and allowed integration with existing disk based asynchronous replication technology over 1,000km for Disaster Recovery. The migration workflow and architecture allowed for the re-synchronisation of data volumes in the new production data centre without the requirem ent for a re-synchronisation over the WAN. The solution provided automated processes to manage over 1,200 disk volumes across three locations; over 3,600 in total. Significant transformation elements were delivered as part of the project including new zone configurations, Fibre Channel port assignments, RAID set and LUN optimisation, VSAN separation, Inter-VSAN routing, N-Port Virtualisation, and FCIP.

Visit SNW Europe 2010 for more information.

Retention policies dictated by corporate and regulatory mandates exacerbate primary information growth as organizations are required to save data for longer periods of time. Additionally, companies are saving historical information in an effort to improve business intelligence processes as decision support functions benefit from having access to more data.

(Source: ESG REPORT – Why CIOs Should Look To Data Deduplication)

Modern tape technology like LTO has tremendous capacities and speed – if used in the right way. If you are able to stream data at high speeds and continuously then physical tape can be blisteringly fast – big databases are perfect for streaming to physical tape.

VTL comes into it’s own when there are lots of files to backup. Effective use of caching and de-duplication technology maximises the throughput and ensure VTL space is used in the most effective way.