| These NetEdges seem to have three different possible operating states: | completely working (which doesn't happen often enough); broken (often, right | out of the box); and kind of working (which happens all too often). That these things work at all under load is nothing short of miraculous. There are plenty of mixed-media-bridging devices that have been tried and which have failed miserably in the past, most notably: -- Magnum 100s (ethernet<->funny framing<->ethernet), which don't meet the standard IFG specification; jam too many back-to-back packets at them, and it will lose most of a burst of traffic. This hurt the old MAE-EAST. -- various ADSUs (FR<->ATM<->FR), which lack buffering and perform SAR too slowly, which delayed the PAC*Bell and Ameritech NAPs for months and some of which still exhibit flakiness under load -- a horrible idea (ethernet<->MFS ATM<->ethernet) which didn't work under load; perhaps a veteran user of these neato little things could explain the failure mode to anyone interested. I remember one service provider who had a national "10Mbps" ethernet backbone who had various horrible problems including the breakdown of the LIS (such that some routers couldn't talk to others), connections going simplex, frame loss and other wonderful things -- another horrible idea (FR<->MFS ATM<->FR) this was pretty neat; some of the problems above plus a brand new problem: the ADSU would strip the FR frame checksum, perform SAR, send the cells, and the ADSU on the opposite end would reassemble the frame and produce a correct FR checksum. All fine and dandy, unless cells arrived out of order, SAR was done wrong, or there was data corruption under load in the DSU or in the network. This interesting technology advanced the state of IS:IS in one vendor's software rather considerably -- mixed-media bridging (NetEdges, FDDI/Ethernet bridges) these break in all sorts of interesting ways. In particular, NetEdges have an annoying habit of confusing FDDI stations in particularly toxic ways, and some FDDI/Ethernet bridges resemble roach motels: frames check in but they don't check out. Essentially, most of these things worked mostly perfectly under low load, but when faced with the kind of traffic one sees at a busy exchange point, most bridging technology has failed in really awkward ways. My advice is that if you can avoid talking to something across a bridge at an exchange point, you should do so. The keep-it-simple principle is a bunch more expensive, but probably not as expensive as a very public failure. Finally, why is it that most vendors never test their products in a serious battlefield environment like an ISP of size medium to huge? These places tend to be excellent worst-case testing grounds. Sean.