This is a basic guide, meant for the jack-of-all-trades, not the storage professional. There is a dizzying array of storage appliances on the market. Appliances with a list of features that would make a Swiss-Army knife blush. We could argue the pros and cons of specific features all day, but this article is more focused on the IT generalist. More to the thinking, “Will this storage appliance help my business?” As such, we will discuss three initial questions to answer before starting the process of determining the best storage appliance for your environment.
These three questions are 1) How fast and how much storage do I need, 2) For what purpose do I need this storage, and 3) What technology will I use to attach this storage to my environment?
How much, how fast?
So – How much Storage Appliance do you need, and how fast does it need to be? In simple terms, it is time to do your homework. You can collect data on your network using a multitude of tools. The common term for this work is a storage study or survey. In this study, you or an IT engineer will usually activate utilities that are already on your various systems. These utilities collect information about how you use your storage on your various boxes over a period of time. As far as the sampling period goes, the longer the better with about 7 days being a happy medium.
Depending on how many systems you have at your location, the collation of this data can be a bit of a beast. There are several vendors that have tools to help collate this data, and they are both free to download and worth the effort. Here is a sample report of a personal favorite that supports individual systems as well as VMWare. Click Here to download a copy to run in your environment.
This collection of data will give you insight into how much data storage your environment uses, how often that data changes, and how fast your systems are trying to write data to that storage. This is fairly equivalent to horsepower and gas mileage in a car. There are several subtle nuances that will become important later in the decision process, but this collection will get you the information to start. It is possible to put together a good picture of your environment, once you get this information. Data show where you are and allows some assumptions as to how you may use data in the future. You will run into terms like Input/Output Operations a second (IOP/s), latency, and queued operations. Save the questions for later, we have more work to do.
How will I use my storage appliance?
How you intend to use this storage is an important factor in what storage is best for you. Is this storage for a single application like a database, or do you plan to share this storage across several servers and allow them to each use a portion of it? This may determine the sort of connectivity that you need. In addition to the number of applications that may be using this storage, what are those application doing? Databases tend to need more structured access to data and the faster the storage is, generally the better the database will perform. If you have several people on the database at the same time, this may become a factor. Also, for general storage, speed isn’t always the issue as much as perhaps usable space.
Is virtualization a factor in your storage decision? To achieve a truly virtualized environment, there must be shared storage between physically separate hardware servers. If you are virtualized, or plan to, this can be a factor in the type of storage that you choose. If you have considered VDI (virtual desktop infrastructure – or the “virtualization” of the desktops in your environment), then that will most definitely be a consideration in your decision.
How do I attach?
How do you plan to connect your storage to the hungry applications that need it? Unlike traditional direct attached storage, a SAN or NAS doesn’t just clip into your server using existing slots in the chassis. There is a bit of planning involved.
The most common ways to connect NAS and SAN devices is using your existing network. No problem if your network is serving your needs well. If not – then there is some planning needed into how to best get your applications connected to the data that they crave. Fortunately, the numbers that we collected in the storage study will give you some insight into this. There are alternate ways to connect the storage, if your network isn’t cutting the mustard. Some are best for databases, some are designed to allow low-cost access and growth, and some are hugely expensive but fast as blazes – definitely used for specialized applications. Just collect your thoughts on this and we can use it in a further discussion.
Once you have this information collected, you are well on your way to the having the information that you need. Having done a storage study, determining how you intend to use your SAN, and determining how you will connect servers to the storage will start you in your journey to find the correct the SAN or NAS for you.
We will address each of these questions in more depth in future articles, so stay tuned.