Tech Careers: Be Persistent

There are an incredible variety of software technologies and capabilities that can reside inside of a disk array. In the coming weeks I will write several articles that explain some of the more important features. Understanding these features, and how they are being used by customers, can give you an edge during the interview process.

For an introduction to disk array technology, see my last post. In it, I also encouraged high-tech job seekers to spend more time improving their understanding of the importance and impact of the disk array.)

The most important feature of a disk array, by far, is its ability to persistently store and retrieve digital bits of information. We’ve been discussing a simple diagram which illustrates the flow of information from an application down to a persistent storage device. If the persistent storage device is a disk array, the information is being sent to a cabinet which is actually a collection of disk drives. A simple picture of this scenario is shown below.

 

A disk array contains intelligent software that makes one of two possible choices when storing incoming information:

• It places the information on one disk, or
• It distributes the information across multiple disks.

In addition to these two choices, the disk array software usually has the ability to further protect the information by storing additional copies on one or more disks (mirroring), or by mathematically generating and storing additional information on one or more disks (parity).

In summary, a disk array has a wide degree of flexibility in its ability to persistently store information. These choices affect the performance of the disk array (how fast it stores and retrieves information), as well as the financial cost to store the information. The possible choices include:

• How many disks to use when storing the information (more disks can result in increased performance).
• How many extra copies to make when storing the information (more copies results in higher costs to store the same piece of information).
• Whether or not to generate mathematical parity (this option can have performance and cost implications).

For more reading on the cost versus performance tradeoffs of disk array technologies, I refer you to the landmark RAID paper (Redundant Array of Inexpensive Disks) that was influential in the growth of the disk array industry.

RAID is a technology that you should familiarize yourself with, and it will be the subject of my next post.

Steve
http://stevetodd.typepad.com
Twitter: @SteveTodd
EMC Intrapreneur