Essential guide to business continuity and disaster recovery plans
A comprehensive collection of articles, videos and more, hand-picked by our editors
Data centers and wide area networks may get all the attention, but network administrators should also have a local area network disaster recovery plan in place for disasters both large and small.
The term disaster recovery conjures up images of catastrophic destruction, but in reality, a single leaky pipe or power surge can cause a small disaster for an enterprise LAN. Without a network disaster recovery plan in place for the LAN, the loss of even a single network switch could quickly turn into a major, time-consuming outage for the organization.
Basics of LAN disaster recovery planning
Network administrators should start their local area network disaster recovery planning with a complete inventory of what is actually on the network and where it is located. This inventory should include switches, WLAN controllers and access points, network appliances and any other device attached to the network.
There is a wide range of commercial and open source software applications capable of scanning a LAN and identifying and classifying each connected device. Administrators should use this scan as a baseline and define the physical location of these devices. For example, it‘s important to know what equipment is in a flooded basement.
An accurate inventory will help the administrator determine which spare parts, such as switch line cards or power supplies, to keep on hand for both hardware problems as well as smaller, more localized events, such as lightning strikes or burst pipes over networking racks. If budget allows, an administrator may consider having a couple of spare, production-ready switches on a shelf to replace any critical devices that fail.
Armed with the network inventory, network administrators should next capture the configuration data for every device. Whether this is the configuration for QoS and VLAN settings for Ethernet switches, or the configuration settings for the entire wireless network, having the latest configurations and profiles for the local networking gear in a secure alternate location is essential to a rapid recovery in the event of a disaster. An administrator should log any configuration changes in the disaster recovery plan. Network change and configuration management (NCCM) tools can automate this process, but administrators must back up the data in the NCCM system and make sure that they can access the information in the event of a disaster.
Collaborate on your local area network disaster recovery plan
To support the organization’s disaster recovery efforts, network administrators can work with the server, desktop and data center support teams to determine the relative criticality of each inventoried item on the network. In the event of a large-scale disaster, knowing which aspects of the network demand first priority will define the speed of recovery time and ultimately set the order in which network devices are brought back online.
The enterprise must consider the effort and costs needed to completely cover every aspect of the network and weigh that against the criticality of that network component. If, for example, an enterprise has deployed a wireless network only for guest access, it won't be worth ensuring redundancy in the wireless LAN.
Minimize the single points of failure in a network disaster recovery plan
A network administrator must ensure that the enterprise LAN is as resilient as possible during a disaster, with the ability to handle the loss of any single network component with minimal impact.
The typical enterprise LAN has a number of single points of failure that administrators should identify and remedy if possible. A aggregation switch that pulls together all edge and wiring closet network switches or a wireless LAN controller installed without a redundant backup device can lead to a significant outage.
As resources allow, administrators should introduce redundant devices at each critical single point of failure they identify. There is little that can change the impact of a total loss of a facility, but redundant network links, power supplies or secondary wireless controllers could minimize the impact of a smaller scale event.
Incorporate simplicity into a local area network disaster recovery plan
An effective network disaster recovery plan sets in motion the steps needed to restore network access to at least a minimal working state. The person who restores the network might not be a network administrator. Events might prevent the admin from having access to the network. Other technical staff, such as storage and system administrators, should be prepared to execute on a network disaster recovery plan. Ensure that staff from other parts of the organization can execute on the recovery plan through documentation and testing. During many disaster recovery tests, an enterprise may flag key personnel as “unavailable” for the test, in order to ensure that the required knowledge of these administrators is contained within the recovery plan.
The role that the network team plays in a disaster recovery event will vary wildly among organizations. Based on the planning and recovery in place, the local network administrators could find themselves at a third-party hot site location assisting the recovery of the data center, or at a temporary office space for displaced workers.
Virtualization brings even more options for local area network disaster recovery
Server virtualization has changed a lot of the rules for disaster recovery. Beyond improving the recovery time objective (RTO) for the server administrator, virtualization can also improve how the network administrator handles recovery of network appliances. Many hardware appliances, such as load balancers and application delivery controllers, now have virtual machine equivalents. Although many of these virtual machines cannot match the performance of equivalent hardware appliances, they could serve as temporary replacements in a disaster recovery situation. Running these virtual appliances at the recovery site would lower costs by eliminating the need for matching hardware at the remote site.