Bladewatch.com

Chris was asking me about storage energy efficiency, one of the architects was saying how he was seeking to switch storage platforms to reduce power usage in the data center, as the new array they were looking at required more power.

“…We were thinking of switching from ******* to ******* having looked at the energy footprint of their array, we wondered what impact that might have in terms of SAN/NAS and wanted your opinion.”

Went the email (thanks Chris) who’s been watching the rain today in Canary Wharf.  Anyway he was asking me what I  thought.

Emm. Well before switching array shelves, storage platforms from EMC to HDS, from HDS to HP to NetAPP to Sun, Dell or IBM, can we look out our core?

• What are our current and projected utilizations and requirements for storage
• What applications/services are using the most storage
• What rules/procedures do we have for storage in the enterprise
• Do we have any best practice/rules for tiering storage where we offline data that’s not needed online?
• What architectural standards are in place for applications in relation to storage
• What de-duplication techniques, best practices or technologies do we use across the enterprise or application? Are we duplicating data, results, workloads and functions within related or different applications
• What combination of virtualization and storage processes/technologies are we using to our advantage? Are we creating separate C drives for all virtual machines, rather than one gold image and creating a change file per virtual machine?
• How are we addressing the real issue of data growth – the concept that the next EMC storage solution purchased is full before it’s delivered and connected.

The prime example I was using with a client yesterday during an interview (which is being published tomorrow) was:

The backups didn’t work, rather they worked, but the lead time to a restore felt like forever. This resulted in data being kept on the servers, which caused the backups to take longer to complete, which caused interference with some specific applications and at the same time caused the SAN storage solution to get full.

We can look at switching platforms/technologies at anytime, that’s the easy bit, we need to look how the business, the operational issues are acting as a barrier to success, above it was that the tape recall process was to slow, at another organization, their cross charging caused an issue. Teams were billed different amounts dependent on whether the storage was ‘Tier1′ (99.999%) available or ‘Tier2′ (99.9%) available, therefore to reduce individual application and business line costs, everyone went on Tier2 storage, “You sure it’s only Tier2?” IT would ask, “Oh yes, we’re sure”, suddenly two things happened, Tier2 got full within a few months, secondly a number of business lines turned around and said “That’s got to be Tier1 now, our application’s now Tier1, you can move us but we can’t have downtime”.

Address the non-technical, billing and operational issues and everything else will follow, in essence identify the issues and address them, not the resulting issues.

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.