karatekid wrote on Feb 12
th, 2013 at 9:25am:
I gave up a while ago... it used to be thrudh and leslie doing that "it's all client-side shit.
I loved the 768 mbps DSL comment - he'd shit a brick if I piped in and said I'm getting the same rubberbanding, dc'ing, and chat channels dropping and am on 40 mbps fiber.
He also tends to say the amd bug in his sig's the problem, 'cept that only applies to Windows users.
I tend to believe glitches on random players is caused by something wrong on discrete servers within a given world's cluster (i.e. not all of Thelanis or khyber or what-not). If there's a bad node in the storage fabric or a bad database within the cluster, it WON'T necessarily affect all players - only a subset.
Last night, I had double my normal latency for about an hour, then it fixed itself - noone was using any network-connected devices at the time. This kind of behavior and the server crashes tells me there's something majorly wrong inside their datacenter (and/or their ISP).
There's nothing you can do to convince them there's a problem on Turbine's end.
Well as I pointed, from Turbine point of view most of the latency sources are client side.
Now I also pointed out that server side can just go locking up 11 persons because the 12th is DCing and the server is not receiving any acknowledge back from that DCing person.
There's more than net involved and I could go all into the latencies involved into database queries, and all the back end stuff at the server side. But what's the point it is going to be lost for 99,99999% of the people on the official forums ( and it will probably earn me an infraction and a ban for exposing guesses on the back end infrastructure )
At the same time what we call server is most probably a few cabinets filled with PC blades ( or similar ) and that can generate lots of fun stuff.
There's definitely, in some case ( typical examples are instance 2049 and 2050 ), a problem at Turbine side... but in 99,99999% of the cases the problem will be at ISP or client side.