A goal of any security professional is to maintain constant access to data. In this video, you’ll learn about disk redundancy and using RAID to increase uptime and availability.
<< Previous Video: Multi-factor Authentication Next: Network Redundancy >>
In this next series of videos we’re going to talk about redundancy. This is how we duplicate parts of a system so that it’s always up and running, and available for the users. This is usually a reaction to some type of failure that’s occurred, and if something has failed, we still want to keep all of the systems up and available for use. The goal is for the organization to continue operating even though there’s been some type of failure with part of the systems.
This could be related to hardware, you might have a motherboard fail, or power supply go out, and you still need to make sure that everything is able to work. You also might have problems with the software. If there is a bug in the software that causes a service to crash or the application is not working as you would expect, there still needs to be some type of process in place to maintain the uptime and availability of those services. And of course, this might also mean that we’re building networks that are able to maintain and stay up and running even when part of the network may have failed.
Another way to maintain the uptime and availability is to create redundancy in a completely different geographic area. This is especially useful if the area you’re in suffers some type of major outage caused by hurricanes, tornadoes, flooding, or any other type of major disturbance. If you have some type of redundant location that is outside the scope of that hurricane or flood or natural disaster, like a separate data center, then you can keep uptime and availability with the services in that separate geographical location.
This might be part of a normal operation you have, there might be a data center on the east coast and a data center on the west coast, and if anything happens on either one of those coasts everybody can start to use the other data center to maintain the uptime. This might be part of the normal disaster recovery process. You might have a data center in Florida and one in Denver, and if there is a hurricane in Florida all of the processes can now be running from Denver.
When we’re connecting up network devices such as network drives, we often will configure multiple links in the network to provide redundancy if one part of the network was to fail. We call this Multipath I/O, or Multipath Input/Output. This allows us to create other routes on the network that we can use to work around any problems we might have. For example, if we have a Fiber Channel network we might have multiple Fiber Channel switches, that way if we lose one switch we can redirect the traffic through the other switch.
We can also create redundancy with the drives themselves, this would be RAID, or a Redundant Array of Independent Disks. Most of the redundancy in RAID comes from using multiple drives within a single array, where you can store some or even all of that data on that redundant drive. This way if you do lose one of those physical drives, you have separate pieces of that data stored on other multiple drives as part of that array.
When a drive fails in a RAID array the users usually don’t know that any problem has occurred. The RAID array continues to be up and running, and all of the data continues to be available. The system administrators will then, behind the scenes, replace the bad drive. That redundant information is then rewritten to the replacement drive, and now you’re ready for another failure should one occur.
If we review some of these RAID levels from our A plus studies, we can start with RAID 0. RAID 0 is no redundancy whatsoever, it’s usually called striping without parity. Where you have very good performance to be able to read and write to that array, but if you lose any drive in that RAID 0 array, you’ve also lost the data, and there’s no redundancy available. That’s why most places will start with a RAID 1 at a bare minimum. RAID 1 is what we call mirroring, where we can take one physical drive, and duplicate all of the data on that physical drive to a separate physical drive. It’s a mirror of the information. That way if we lose any one of those drives, all the information continues to be available and accessible on that separate drive.
Another redundancy type is RAID 5, where we have striping with parity where we’re putting pieces of information on separate physical drives, and then on a last physical drive we’re putting some parity information. If we lose any of the drives on that particular array, it will rebuild the data based on the parity information that’s put on that extra drive. And of course there are combinations of RAID that you can choose, RAID 0 + 1, or 1 + 0, RAID 5 + 1, and other combinations as well. By combining these RAID types together you can customize the redundancy for your purposes. And you’ll be prepared regardless of what physical drive might fail.