But now I can happily report those problems are history. My recent tests on T-FRAP version 2.61 at the MCI Developers Lab found it running clean under load. And when I upgraded to version 3.0 beta, I was amazed by its new features. Version 3.0 not only raises the bar for frame relay-aware CSU/DSUs, it also includes capabilities--including simplified installation, automatic discovery of remote peers and per-PVC (Permanent Virtual Circuit) diagnostics--that you won't find in other frame relay management probes.

Indeed, as we noted in our previous review, the original T-FRAP exhibited some unique capabilities as well. In addition to its standard CSU/DSU functions, the T-FRAP can place frame relay management traffic on any active DLCI (Data Link Connection Identifier)--a frame relay PVC (Permanent Virtual Circuit) end-point address--including those with user traffic. This spares you from having to buy a separate PVC for management data, or shell out cash for Ethernet interfaces on CSU/ DSUs. But if you're a service provider, you may want to dedicate private PVCs to management traffic--and the T-FRAP can do this, too. These capabilities shrink the T-FRAP's competition to a handful of players.

Getting to the Source Through a technique known as LMI (Local Management Interface) Sourcing, the T-FRAP helps you quickly isolate problems when your remote equipment (a router or FRAD, for example) dies or the carrier's frame relay service goes south. If your equipment fails, LMI Sourcing to the frame relay switch lets you get to the CSU/DSU and troubleshoot over the WAN. With it, you'll quickly see that your equipment--not the carrier--is at fault, and the carrier will not issue an erroneous trouble ticket for your problem. Conversely, if the WAN goes down, the T-FRAP can hold the router interface up, so you can access the CSU/DSU and determine whether the frame relay service or the T1 access circuit is causing the problem.

Without frame relay LMI Sourcing in the CSU/DSU, all connectivity problems look the same--your network management system can't tell you if you have a bad router or CSU/DSU, local loop errors or carrier WAN failure. The LMI Sourcing technique works by embedding some frame relay protocol functions directly into the T-FRAP. In normal operation, your equipment sends regular "status enquiry" messages to the carrier's frame relay switch. In turn, the switch responds with "status" messages containing the status of active PVCs, among other things. Equipment on both sides rely on these messages to determine that everything is copacetic.

If you're relying on your carrier to issue trouble tickets when your routers die, LMI Sourcing will load the burden of watching remote equipment on your shoulders (no problem if you've built your own monitoring system). Second, if the WAN is down, you can only access your local router anyway, though it's helpful to use the CSU/DSU to troubleshoot. These techniques aren't a magic bullet, but they can help when used properly.

I tested LMI Sourcing by disabling a remote router but not its companion T-FRAP. As expected, traffic ceased to pass and alarms appeared in Envisage, Sync's Windows-based SNMP management console, which is available as an add-on product. I accessed the CSU/DSU through Envisage and was able to quickly pinpoint the real problem. In a second test, I disconnected my local access line and used telnet through the local router to the CSU/DSU to diagnose the source of the problem.