What you will learn from this tip: You will become more familiar with what storage automation is, and what it is not, and what it really means to you. Also, you will learn why you should, or should not implement a storage automation product.
There has been has a lot of buzz around the term "storage automation" lately. It sounds nice, but what does it really mean, and do you really need to spend the money to make it a reality for your organization?
Depending on who you talk to, the term storage automation can mean a number of things:
- Automated provisioning
- Capacity on demand
- Host transparent data movement
- Hot spot elimination
- Transparent data migration
- ILM
- Utility storage
Storage automation is the ability to encapsulate time consuming or repetitive tasks into a best practices policy that can be initiated on command or triggered via an event with little or no human intervention; you can think of it as being like a script.
Automation is a good thing for routine tasks that take a lot of time. Automating routine tasks can decrease administrative overhead and save money by allowing fewer people to do more, and can eliminate the human error factor. Storage automation products are available from software and hardware vendors as software modules or hardware microcode that can:
- Provision storage to hosts based on your policies for switch zoning, LUN masking and performance criteria.
- Move application data from low-performance to high-performance disks (or visa versa) within an array based on application I/O metrics or statistics.
- Migrate data between different classes of storage based on data age or frequency of access.
- Automatically expand LUN sizes to hosts based on policies for how much capacity is left.
- And a host of other cool capabilities that are outside the scope of this tip.
Storage automation is not a panacea for IT administrators, so they can just forget about the day-to-day administration of storage resources to focus on more important stuff.
But let's face it: Stuff happens. Even though routine tasks can be automated, administrators still need to be on their toes and monitor the environment to make sure everything is humming along properly. An automation process may fail, or hardware may fail in the middle of a data move, which may cause data corruption. Automating routine provisioning is a great place to start implementation, since failures during that process present much less risk to critical data.
Do you really need storage automation? That depends on your environment. If your shop is small, and your employees are familiar with API scripting, buying off-the-shelf automation tools may not be required, since you can roll your own. If your shop is large, and your staff is continually fire fighting or dealing with outages, implementing vendor tools to automate most of the day-to-day tasks will allow your staff to be more productive.
In either case, make sure that the product you put in place includes enterprise-class monitoring and reporting capabilities, so your administrative staff will have the proper controls and problem notifications should something go wrong.
Do you know…
A guide to cost management?
How to build the ultimate storage team?
About the author: Christopher Poelker is the co-author of SAN for Dummies.
