Work’s external network connection is rubbish. Even when it works properly. Which it isn’t.
Since Thursday last week, every afternoon our connectivity gets shaky until well into the evening. Emails get delayed, web connections time out and fail etc. It is very annoying. People at HQ have suggested that ‘rogue traffic’ on the network may be responsible. This is a reasonable suggestion and one that I had already investigated – last year the Blaster worm caused a ping flood which effectively knocked our network offline. So I have spent all day reading tcpdump logs to analyse all the traffic. And found nothing interesting. Nothing like last year’s ping flood traffic, anyway.
Some of the errors visible in the packet logs suggest that either the router or the actual leased line itself is faulty, which sounds like a tedious thing to have to troubleshoot, so we’ll probably have Connectivity Issues for a while.
The only moment of amusement in this whole mess is that when a droid at HQ opened the fault log, she used the term ‘killerstream’ to refer to our connection, which was her version of what I dictated to her over the phone. The correct term is, of course, kilostream. 🙂
UPDATED Thursday 5 August 19:00: BT replaced the network ‘black boxes’ in our comms room and at the local exchange and all seems to work fine now. Given that I’ve reported these problems before, I surprised this isn’t something that has been investigated previously. Any, we have bandwidth back (such that it is).