interleaving

[alexatkin]

Good to know, I always wondered how these sorts of things worked.

Interesting that it takes a while for peering to fully restore itself once the problem is corrected, you would think it would happen quicker.

This does emphasise one point about the TBB pingtest though, its ONLY testing the latency of a single peering. So faults like this show major disruption, when in real-world use many of the sites I use won't even use LINX at all.

[jody-ASK4]

It only takes a while to get back to where it was due to network admins disabling their ports/withdrawing their advertisements until the peering exchange is stable again. The BGP conversations usually just take a few seconds.

[aceado]

i've asked to be back on interleaving with very minimal settings the hit for this is same speed i think a very little drop on my upload speed :/ but my ping has change from 10ms to 16ms to thinkbroadband

[alexatkin]

Personally I doubt it was fastpath causing your occasional stalls, but we shall see.

Its certainly interesting to see what different settings Thales are willing to try. I'm itching to get switched to fastpath now but there is no point when I will be switching to Max soon which will likely require switching back to fastpath again at that time as I doubt they will risk putting me on Max WITH fastpath right from the start.

That is one reason I really REALLY wanted to have the Huawei H612 on my line, because I can know for sure if things are stable by monitoring any errors. Though granted, my line is a LOT shorter.

[ThEGr33k]

Well, not sure what caused the slower speeds last night, the outage? Either way, tested this morning and here are the results; (After and before...)



Will add that although I got "only" 38.7Mb/s above on DL it did go to 39.5 for the majority of the DL test and dropped briefly at the very end to ^. I only ever got 39 tops before so in theory I have gained about 512k... Not bad really.