We’ve been using this simple diagram as the basis for exploring a set of adjacent technologies in the world of high-tech information storage and retrieval. We’ve also discussed how disk arrays are large cabinets containing hundreds upon hundreds of disk drives that store this information. Within a disk array a variety of failures can occur without impacting the ability of the application to fully access the data. Disk arrays can take a lickin’ and keep on tickin’.
Corporate customers, however, are rightfully concerned with a different form of failure: a breakdown in the connection to a disk array. If an application has a single path between the server and the disk array, what happens when that path experiences a failure?
This situation is similar to Daffy Duck standing on the wrong side of the branch as he saws a limb off of a tree. It’s not going to be pretty. Information access will be lost.
For this reason customers often go to the extra expense of creating multiple hardware paths to a disk array. They will install redundant network cards in a server. They will deploy multiple network switches in between the server and the disk array. Each network switch will run multiple cables to two or more ports on the disk array. These types of configurations can withstand multiple failures, whether they occur at the port, switch, disk array, or within the cables themselves.
The hardware plumbing, however, is not enough. What’s needed is intelligent software which is path-aware and can react to interruptions in the flow of information. This software will often sit just below the application. When the application hands off a set of digital bits, multi-pathing software has full visibility into the health of the underlying infrastructure. It can navigate the different routes effectively. The diagram below shows a piece of software known as PowerPath.
There are many things to learn about this type of capability. First of all, technologies like PowerPath are often referred to as “device drivers”. Device drivers provide access to underlying hardware. Secondly, technologies like PowerPath can also load-balance read and write requests to achieve higher performance.
Given that device drivers touch every piece of digital information that is being passed down to a disk array, there is a chance to provide value-added services to data as it passes through. For example, a device driver can negotiate with a key server and acquire encryption keys which will protect the data from being intercepted as it passes down to the disk array.
I recommend learning more about pathing technology and device drivers in general. Knowing how they work and what they can do will be an advantage for you in an interview. For a more detailed description of PowerPath, you can start here.
Want to be found by top employers? Upload Your Resume
Join Gold to Unlock Company Reviews