Planning essential
IDC's Penn says businesses need to assume that their storage requirements are going to grow and plan accordingly. "For example, if you expect one terabyte of data a month you are probably underestimating," he advises.
"Most organisations find growth is far quicker than expected. Plan from the beginning: have bigger pipes into servers and allow for the fact that you may have to do backup and recovery every day and perhaps multiple times a day. This is one way to minimise costs--prevent mistakes before they happen."
Likewise, Brocade's Schultz says that any organisation which is finding that it's buying a lot of servers, and therefore lots of storage, will find that through storage consolidation there is a direct spin off and benefit on capital expenditure.
"Because when you start consolidating your storage--be it disk and/or tape--you have the scenario where you're not buying a tape drive for each server, you're not buying a disk for each server, and in some cases having to buy new servers because you've run out of capacity for your direct attached disk."
"The actual cost of buying individual direct-attached disk per server, when you add it all up, is infinitely more expensive in the long run than buying a large storage infrastructure and being able to just add disks to that infrastructure when you need it and not link it to any particular server."
Schultz says Brocade studies carried out in conjunction with KPMG on large organisations, revealed that direct return on investment for storage consolidation were about 296 per-cent with a payback period of four months.
In the same case studies, the back up and restore applications and infrastructure had an ROI of 123 percent and a payback period of nine months. The high availability environment based on the SAN infrastructure once it was deployed at an ROI of 525 percent and a payback period of two months.
Veritas' Elisha is another advocate of carefully assessing what your storage requirements are before the implementation is started. He says back up is often a last thought.
"People will go into a new application and [think] 'we're going to implement a new ecommerce application, it's going to have all these bells and whistles' and people will be very focused on the hardware that it needs, the performance level it has to reach, the new technologies that it's using--they will not think about protection of that data."
"It's most likely they'll think about it in the last few weeks before implementation, or when something goes wrong."
Elisha says at that stage it's very difficult to retrofit an effective solution. He says if storage is considered from the outset, then storage vendors could suggest possible solutions, such as completely backing up data every night, or by tweaking infrastructure in a certain way.
It all comes back to the design process. "Your system and your data is only as good as how available it is, so if, for example, you have a major system crash and you can't get your information back, your business has a very good chance of going out of business," Elisha warns. "And no amount of high tech will save you from that if you haven't planned it in in the first place."
Elisha says businesses often don't think about the availability of their system, but that it can be very easily planned in from the start and planned in to the overall strategy. "They know at the end of the day they're going to have an effective application that's backed up and recoverable and will work in the times that they expect it to work."
Hi,
Outside the operational and IO aspects, there are also costs associated with all these technologies. I thought readers may also want to check out an article which discusses Costs associated with SAN and DAS.
http://capitalhead.com/articles/san-vs-das-a-cost-analysis-of-storage-in-the-enterprise.aspx
- Mike