Preparing to Replace Your Data Storage Array
If you’ve been using your data storage array for more than three years, you may be considering a replacement. Perhaps you just received a support renewal notification; or, if you’ve had this array for five years or more, the manufacturer may be announcing its End of Life (EOL). It could be you’ve simply outgrown the capacity and exceeded the performance of your current solution. Whatever the reason, there are some things you can do to prepare for your array search that will help you navigate the myriad of options and more quickly discern which solution is right for your environment.
In the next few blog posts, I will talk about some of the features/functions you’ll find in current data storage offerings and provide my take on how to determine their value for your environment. In this blog post, I want to describe some of the information you’ll need to understand about your current environment before you begin the search. Specifically, you should know what your usable capacity requirements are, the characteristics of the data on the storage system, and the workload profiles your array supports. If you understand these three things, it will be far easier for you to discern what features are important to you and which are not.
Usable Capacity – While understanding the number, size, and speed of the drives in the current array is useful information, it doesn’t tell the whole story. Most storage arrays utilize some level of RAID protection (RAID 5, RAID 1, RAID 6, etc.); and, consequently, you will need to understand how RAID is implemented on your array and what capacity is provided AFTER taking these RAID levels into consideration. Beyond just capacity, this information can help define the levels of service you currently support for your applications. With this information you’ll be better prepared to have discussions around how a particular solution can provide the same levels of service you currently offer while also meeting your capacity requirements.
Data Characteristics – It’s important to know about the data you store on the array. Does the array store Image files, User files (Word, Excel, PDF) Databases, a mix? Additionally, what are the attributes of those files? When were they created? How long has it been since these files were last accessed? Answers to these questions will help you more clearly define storage tiers and give you an indication of how well you could utilize array features like de-duplication, compression, and thin-provisioning. Classifying data in this way can also identify whether or not you have files that could be archived or deleted altogether. Capacity requirements can be significantly reduced in many cases if you discover that you are storing stale data that doesn’t need to be taking up space in your production environment.
Workload Profile – Measuring and documenting performance metrics such as IOPS, Throughput, and Latency, on the array as a whole and on each Storage Volume, will tell you the level of performance your array is being asked to support today. You should also understand the Read versus Write ratio and the Sequential versus Random operations. These metrics should be collected across different periods of a day and a week so that you can account for peaks when sizing a storage solution. It’s also important to associate performance to particular Applications. Understanding these metrics will help determine what storage tier Application volumes should be placed on and which could benefit from active-tiering versus those that should be pinned to a specific tier.