Green storage sounds good, but for most of us, it's a goal we'll keep reaching and trying to reach again and again. Here are some practical tips for greening your storage to help you save storage costs.
Tiered storage
Spinning disks consume energy both directly and indirectly. They require energy to spin and generate energy in the form of heat, which then requires cooling systems that consume more energy. So the more spinning disks you have, the more energy you consume. But not all disks consume the same amount of energy. Fast, high-performance disks consume far more energy than their slower counterparts.
One way to manage storage costs is to tier your storage as much as possible. But this solution requires a lot of hard work. However, keep in mind that using lower performance disks (and the arrays that house them) can dramatically reduce the overall energy consumption of your storage infrastructure. Also in regards to infrastructure, you should see how much it costs you to maintain old arrays. How does that compare to the cost of purchasing and managing new ones? Eliminating your shop of old equipment and consolidating your storage for the purposes of energy efficiency should be at the top of every storage manager's to-do list.
Thin provisioning
It's no secret that storage is often overallocated. An application requires 1 GB, a database requires 2 GB, and the systems and database administrators often require a buffer. If you factor in growth, you can wind up allocating 10 GB -- a huge waste. As the storage manager, you have little control over that wasted space. But what if you had an intelligent array that could spoof the 10 GB and allocate only the 1 GB that the host truly needs with a promise to make that extra 9 GB available should there be a dire need? This is where thin provisioning comes in. Now instead of provisioning storage in a siloed manner, you can create "global" storage pools that all systems share or dip into as needed. Suddenly, the wasted 9 GB can be put back into the global pool and used to overprovision other systems that (similarly) require it.
What about tape?
Trying to save money on data backups is a topic unto itself. Backing up the same data over and over again wastes both tape and disk, and isn't wise. If you have a gut feeling that your backups may not be efficient, you should examine how much you spend on backing up data, and ensure that whatever you do spend is spent wisely.
If you send hundreds of tapes offsite each month, take a look at options for eliminating redundancies. Perhaps operating system images or application binaries are backed up more times than needed to restore the environment. Are log files and other types of disposable files being backed up? Be smarter when writing those "exclude" lists and policies. It's easy to implement and configure a backup solution, but it can be very difficult to configure it efficiently.
You should also consider investing in suitable data deduplication technology. Like thin provisioning, deduplication has the potential to cut down the amount of data that gets backed up. But beware the pitfalls of virtual tape libraries (VTLs), which are nothing but a bunch of disk drives at the back end. When purchased without any deduplication solution, VTLs can be more expensive consumers of energy than older tape libraries. Right sizing your infrastructure should therefore apply to your backup infrastructure in the same way it does to your online storage infrastructure.
Examine your business continuity/disaster recovery practices
Business continuity/disaster recovery (BC/DR) practices are a double-edged sword. Not having a BC/DR plan can be considered a big vulnerability in your company's IT strategy. On the other hand, its cost and upkeep can easily exceed its usefulness. But that shouldn't stop you from implementing a recovery strategy. You should always have a business continuity/disaster recovery plan, but don't be naive about it. Be creative, but rational. Ask yourself how you can reduce what goes into that infrastructure and find ways to put it to good use during its peacetime assignments so it's not just sitting there consuming energy. If you can move some operations to the remote/recovery site, then explore those options. If your recovery timelines are so relaxed that you don't need a hot recovery site, implement tape recovery. Even if a lot of people are implementing hot-standby recovery arrays, that doesn't mean you have to go along with the crowd. What's cool often comes with a heavy price.
Adopting green storage practices is all about being a good corporate citizen and trying to save storage operation costs in a rational way while making a positive impact on the environment. While the world talks about climate change, you can start examining the temperature of your storage operations. By the time solar or wind power fuels your data center, you'll need a new set of operating instructions.
This article originally appeared in Storage magazine.
About this author: Ashish Nadkarni is a principal consultant at GlassHouse Technologies.
