While ensuring access to critical business data often drives disaster recovery (DR) planning, including network services in your DR plan can ensure business continuity. In this FAQ, Carrie Higbie, global director for data center solutions at the Siemon Company, discusses the role of network services in DR planning. Her answers are also available as a download below.
Table of contents:
How can network services be built into your DR plan?
How are network services affected in a disaster?
What is WAN optimization, and how does it improve DR processes?
Can network services improve communication during a disaster?
Are there any best practices for DR testing with regard to network services?
What other steps would you recommend taking in a disaster situation?
How can network services be built into your (DR) plan?
That's one thing that people often overlook. They think about switches and their local environment, but they don't think about what happens outside. They might have two carriers, but if both of those carriers are terrestrial, they might not have any network services at all. So depending on the disaster, they might need to look for an extraterrestrial or satellite type of connection for critical applications.
How are network services, such as DNS and DHCP, affected in a disaster?
Depending on how you're set up and what tier-level data center and type of redundancy you have, somebody might have to physically go in and change those pools of addresses. This may affect how you access your databases and some of your other applications. In some cases, depending on your level of redundancy, you might have a primary connection and then a secondary connection set up, or even a secondary profile set up so that people can get to those applications when there is a disaster.
What is wide area network (WAN) optimization, and how does it improve disaster recovery processes?
If you think of everything that might go over a WAN, when in a disaster and moving to a slower backup link, you want to make sure that only the applications that need to go over that link do, so that the ones you really need to get to become so slow that they're unusable.
So when you think about WAN optimization, you want to look at the bandwidth that's being used in the network, you want to look at the bandwidth that actually goes over those wide area links, and make sure that's optimized so you're using the smallest amount of the bandwidth possible. So when you do have spikes and surges moving from one connection to another with the bandwidth and change management, all of the connections do happen.
Can network services improve communication during a disaster?
They absolutely can, especially when you look at trying to get in touch with other employees to figure out where people are and if they can still do their job. Certainly when you pull in your disaster recovery, audit and security teams, if you've got some of those network services, especially the ones where you've got a good redundant path and people can communicate, that's an excellent way to get the word out to people that there has been a problem, things are down and that you're working on a resolution to keep your business sustainable.
Are there any best practices for DR testing with regard to network services?
You'd be surprised by how many people have a test plan and they've never actually dropped the server down and physically tested failover to make sure that the services still work. This really should be done quarterly at best and in the worst case, every six months. Every time you change a carrier, you should schedule sometime to plug your primary connections, make sure that everything fails over to secondary connections and it still works. That should really be part of the disaster recovery team's planning.
What other steps would you recommend taking in a disaster situation?
For instance, if you know a hurricane is coming, one of the best things to do is to go out and communicate to your employees to turn off all equipment that isn't necessary. Unplug it from the wall and make sure that it's not going to come up. You're typically going to be coming up on UPSes, generators and secondary circuits, and if you've got secondary DHCP and DNS pools, you may not have enough addresses for it to go out and find every telephone, desktop and laptop. So anything that's not critical to those specific applications should be turned off, things like hardening your data center, making sure that if you do have a data center with windows that they're covered up.
Other disasters like fires are obviously a little harder to plan for, but the biggest thing in any disaster is communication. People freak out when they don't know what's going on. So the more you can do to help them understand, by ensuring that you have a good help desk, a good command center where people have one number to call so that communication goes out to everybody, is going to be key.
Carrie Higbie has been involved in the computing and networking industries for over 28 years. She currently works with The Siemon Company as the global director for data center solutions and services in support of end-users and active electronics manufacturers.
