Storage Solutions for Enterprise: What, How and Why

by Sudhir on March 11, 2014

 

Sandeep Kulkarni - Sr. Director, IT BMC Software Pune
Mr. Sandeep Kulkarni
Sr. Director – IT
BMC Software, Pune

Cloud infrastructure — the buzzword today, stands on a three-legged stool. These three legs are compute, network fabric, and storage. Any short fall in design of any of these legs will collapse the infrastructure. While the requirements for right sizing compute and fabric are simple to define, the storage is generally neglected. Storage solutions can never be ‘one-size-fits-all.’

There are two main types of storage solutions available — block level storage and file level storage. Block level storage is flexible and versatile and will be familiar to anyone who has worked on storage area networks. Block level storage presents itself to servers using industry standard fiber channel or iSCSI type connectivity. In common terms, it can be thought of as a hard drive in a server except that it is not in the server, but is attached to the server through connectivity mechanisms. Raw volumes are created and the operating system on the servers connects to these volumes as if they are individual hard disks. This makes block level storage usable by almost any type of application, database, and file storage. Block level storage is used most commonly with database, exchange, VMware etc. File level storage is often used to share files with users. By creating block level based volumes and installing an operating system on it, one can share files on the system out. File level storage has very clear applications in infrastructure where sharing of the files is required.

Businesses are always looking at decreasing the storage cost to lighten the management burden on IT. Though there are many solutions available in the market that claim to solve all of the storage problems, increasing storage as data grows does not scale well as an enterprise solution. Hence, it is important to have complete visibility into data growth and requirements before deciding the storage strategy. In addition, it is absolutely essential to understand business data storage requirements before investing in a potential solution for storage.

To ensure the right investment is made, current storage environments need to be closely observed — what files are being created, at what frequency, what is the retention period, how frequently are those accessed, how much storage these consume and so on. Using this analytical data ensures that the right technology solution is chosen and will also help in zeroing on the right vendor who will help deploy the solution that is cost effective, scalable and yet meets all your business needs.

Some of the segregation of the data is done based on the following principals:

  • Importance of the data
  • Business value of the data
  • Constant change

Segregation based on the above three principals will help one understand the business value of the data. The next step will be to map the business value of this data to the right type of storage solution. For example, data can be segregated into Tier 1, Tier 2, and Tier 3. In Tier 1, data is business critical and is accessed very frequently. In Tier 2, the data is not so actively used and is not business critical and in Tier 3, the files are never or very rarely used.

While the market today offers a variety of the storage solutions right from high speed to low speed storage, the price is directly proportional to the speed of the disks. While having huge caches or fast SSD disks can be an excellent solution, the cost of such solutions supersedes the benefits they offer. Thus, having the right mix of cache and various speeds of disks plays important role in choosing the right level of solution.

Storage economics is another important aspect of the overall storage strategy. Most of the organizations divide the cost of storage with the capacity to arrive at price per GB. However it is not just the cost of storage, but the expense to run it that constitutes the price per GB. This should include the power required to keep the storage up, data center real estate rent, fire system for the DC, maintenance contract for the hardware, the fully burdened cost of FTE supporting the infrastructure, and so on.

For backup of the storage, many storage solutions implement replication technologies where the data is replicated not just across one storage location, but across multiple storages locations. Since replication technologies use only the delta of the changed data, this typically does not add large loads to the WAN links. First time replication is the only one that utilizes WAN heavily.

The market offers a slew of solutions for the above storage needs, right from NETAPPS DE duplication technology to virtual file servers. Today, solutions implement intelligent data storage by looking at business critical data and placing it automatically in the right storage based on the rules defined are also available. However, the key to selecting the right storage system, is to know your own environment, its expectations, and limitations. A checklist of a few important facets you should look for:

  • Capacity
  • Scalability
  • Cost
  • Performance
  • Reliability
  • Manageability

Related Posts:

Previous post:

Next post: