• United States

FRoDSL and mean time to repair

May 27, 20042 mins

* More about why FRoDSL is a sound choice for accessing frame relay nets

We’re interrupting our discussion of frame relay over DSL for some hot-off-the-press results from a research study that Steve just completed at the Webtorials site. The Webtorials community was surveyed last month concerning their plans for and attitudes toward MPLS-based IP VPNs. The results from more than 200 networking professionals around the world confirmed a lot of our thinking and revealed a few surprising tidbits.

In last week’s discussion of frame relay over DSL, we pointed out that it provides a major price advantage over other methods of accessing frame relay networks because FRoDSL tends to have a less stringent guaranteed time before service is restored in the event of a failure.  This stems from the guarantees from the providers of the copper cable loops over which the service is provisioned. 

Remember though, that a longer MTTR (mean time to repair) is separate from MTBF (mean time between failure), which indicates the likelihood of an outage.  This is a measurement of how long it takes for the service to be restored in the event of a failure.

For most of the target applications, we don’t see a slightly longer MTTR as being a showstopper.  If the FRoDSL service is deployed for an application that’s truly not critical, then it’s not an issue.

But the more likely case is that your application is sufficiently business-critical that it needs backup whether it’s deployed over traditional access or FRoDSL.  So if you’re already backing up your traditional access, that same backup can be used for backing up FRoDSL access.  And you still save money as compared with traditional access.

The even better news is that excellent back-up options are available for the target speed range for FRoDSL.  Again, as we pointed out last week, the “sweet spot” for FRoDSL tends to be in the fraction T-1/E-1 speed range of around 512K bit/sec to 768K bit/sec.  This speed range is especially well-suited for sites that need a bit more than 56/64K bit/sec but still have a hard time justifying (or affording) full T-1/E-1 access.  And in this speed range, dial backup via either analog or ISDN service is quite appropriate.