Disaster planning for the wide area
 
|
|||
 | |||
|
Sign up to receive this and other networking newsletters in your inbox.
Disaster recovery plans for the wide area are often difficult to justify - an additional monthly cost for a service interruption that may or may not occur. Furthermore, it's often tough to determine how encompassing a disaster recovery plan must be. Will the plan cover only major data centers and headquarters or will all sites and users be addressed? The scope of the plan is dependent on the nature of your business and the impact network downtime has on its ability to conduct its operations.
The two most common "disasters" an enterprise prepares for are local-loop failure - cutting off access to your service provider - and major data center failure - cutting off remote locations' access to resources. This issue will discuss the typical methods for addressing the local loop, while the next issue will deal with addressing reroutes across the network.
To address loop failures, your typical recovery plan involves a dial backup solution, whereby the router re-establishes connectivity via an analog or ISDN connection. There are also those organizations that opt to maintain two local loops on separate local access providers and facilities.
For the dial backup solution, you will most likely choose one (or more) of three ways to re-establish the link. The first would be dialing back in to the same provider port your dedicated loop maintained. By doing so, you would establish the original port connection, addressing and PVCs. Following in the same light, you might choose to dial back up into your provider's frame relay network - this time into a frame relay port on another, previously determined switch. This plan addresses not only the scenario when the local loop has been lost, but also when your provider's switch is having difficulties. Also with this plan, a provider's downed switch will not impact your ability to re-establish the connection.
Finally, you could choose to completely dial around the local loop and network, and establish a connection directly with another locations' router. For example, the remote site dialing completely around the network into headquarters' router.
Your provider will typically support one or more dial backup solutions, for which it will charge a usage component and/or flat monthly recurring fee. It is also important to enlist the assistance of your equipment vendor. Many vendors support detection in areas such as loss of signal, prompting automatic dial backup solutions to be envoked.
RELATED LINKS
AT&T offers facts about frame fiasco
Company tries to decipher their frame relay outage. Network World, 4/27/98.
Why AT&T's disaster recovery service failed
Services based on alternate paths through a knocked-out network prove useless during a nationwide outage. Network World, 4/27/98.
Were you affected by AT&T's outage? Talk about it in our online forum.
Cable & Wireless goes for ISDN redundancy
A look at its frame backup plans. Network World, 7/21/97.
Disaster recovery white paper from Deloitte & Touche
Disaster Recovery Planning Project Plan Outline
From the University of Toronto.
Sign up for the NWFusion Focus: Frame Relay newsletter
Additional NWFusion Focus: Frame Relay back issues
