NZBMatrix down again, 'fiber cut' (i.e., no fiber diversity)

For those who 'live' on NZBMatrix (okay, live is a bit harsh), but I search for stuff or update maybe 10+ times on any given day; and I've brought to their attention several times that after reviewing the facility they are hosted at, that it has a LOT to be desired, and folks getting connection errors and the like (both on SSL and non) is due to the lack, within the hosting facility, of any diverse path (both logical and physical) between the entrance-way and the servers.

It also appears that Bell Canada hasn't provided the facility with diverse routes between the local Central Office and the facility either, as they are down today and reporting that they've had a cable cut.

Bad, really bad. I can't believe that anyone would first, design such a thing, or second, would host their servers at such a facility. Again, I've said as much to the folks running that operation several times.

The hosting facility is somewhere in Mississauga, Ontario, a place I've been to many times (Nortel HQ there, when Nortel still 'existed').

In no way should a single fiber cut bring down the facility. If it was a backhoe, they're looking at several hours to get a splicer in there to repair, if they're lucky.

Maybe, just maybe, they'll look into moving now, and get rid of all these minor and major outages. Having designed and built several major hosting plants for clients as diverse as Qwest (where my servers are, in South Seattle), MCI (all over the northeast, including the basement of the WTC/rip), Sprint (including their main fiber facility in downtown Kansas City), AT&T (all over the place), etc., and many others too numerous to mention.

Sad.