Disaster Recovery

Know the requirements for point-in-time disaster recovery solutions

If you're confused about the types of point-in-time disaster recovery solutions available for you to suggest to your clients, check out this article, which covers the offerings and the factors you must evaluate to help you choose the right one.

One of the most common questions my clients ask is, “Can you please explain the difference between the available disaster recovery tools?” The pool goes beyond the easily defined tape backup and high availability tools, which pretty much do what their names indicate. After all, there are numerous flavors of disaster recovery solutions, and they operate in different ways.

Disaster recovery solutions can be divided into two main categories: point-in-time solutions and real-time solutions. While point-in-time solutions mirror or “snap copy” your data at regular intervals, real-time solutions constantly transmit changes to remote data systems. Both of these solutions also break down into various subcategories.

Let's look at point-in-time solutions. These allow you to define a set of files, directories, and/or entire volumes that will be copied en masse to a remote data system on a regular schedule. Here are a few flavors:
  • Hardware-based mirroring solutions give you the capability to copy volumes that are defined on a storage area network (SAN) and other storage systems on the same storage platform. In many cases, these solutions can define changes between the two data sets but only transmit those changed blocks as a large push.
  • Hybrid solutions allow you to snap off what are commonly referred to as business continuance volumes (BCVs). These volumes are composed of the files that have changed between the two storage systems, either at the file level or at the block level. They prepare a separate data volume that's made up of either the files and blocks themselves or “pointer” data that tells the remote storage system what to change. The BCV is then transmitted on a scheduled basis over to the remote storage system, where the changes are applied. These solutions are hybrid because host-based software is generally required to track file-level changes.
  • Software-based solutions track changes from the volume level all the way to the byte level, and then ship those changes on a scheduled basis to a remote host. These solutions give you more flexibility, but there isn't a universal solution, which means you have to find a separate version or vendor for each OS you need to protect. While many engineers find this to be a boon to their disaster recovery plans, smaller shops may not wish to deal with multiple entities to accomplish a complete disaster recovery solution.

Which solution will work best for you? The answer depends on a few factors. If you’re considering a hardware-based solution, make sure your hardware supports the type of solution you want to use. Most high-end systems support all the options, but less-expensive systems may only support certain solutions. You must also have identical storage hardware in both locations. To date, the hardware-based solutions work only when operating between systems on the same storage platform.

If you’re considering a software-based solution, find out what kind of effect the software will have on your hosts. Generally, this is not an issue. However, on older servers, these solutions use a great deal of CPU overhead, which can cause problems. Also, to stick with one vendor, you must find one that can support all the OSs you have to protect.