Comments on: Windows Vista and terminal service http://webninja.org/archives/38 The digital biography of Justin King Sat, 04 Sep 2010 01:03:21 +0000 hourly 1 http://wordpress.org/?v=3.2.1 By: Brandon http://webninja.org/archives/38/comment-page-1#comment-3389 Brandon Fri, 13 Jul 2007 08:29:05 +0000 http://webninja.org/archives/38#comment-3389 Hm, okay, so I read a bit about it. It seems like it could be an actual software problem: Windows or its RDP client not receiving the larger packets fast enough, or it could be any router's problem: implementing QoS because of the large bursts of data. I didn't see anything about routers stripping ACKs, but I didn't read too much. Is it just a certain brand? But if it's the software problem, they'll probably fix it in an upcoming patch, and if that's the case, I'd suggest re-enabling the autotuning. If it works right, it supposed to be much much better. Hm, okay, so I read a bit about it. It seems like it could be an actual software problem: Windows or its RDP client not receiving the larger packets fast enough, or it could be any router’s problem: implementing QoS because of the large bursts of data.

I didn’t see anything about routers stripping ACKs, but I didn’t read too much. Is it just a certain brand?

But if it’s the software problem, they’ll probably fix it in an upcoming patch, and if that’s the case, I’d suggest re-enabling the autotuning. If it works right, it supposed to be much much better.

]]> By: Brandon http://webninja.org/archives/38/comment-page-1#comment-3388 Brandon Fri, 13 Jul 2007 08:03:10 +0000 http://webninja.org/archives/38#comment-3388 Why in the whole wide world of muffins would the auto-tuning in Vista start using larger packets without an ACK. Don't get me wrong, optimizing TCP throughput is an awesome idea, and, in Microsoft's defense (whoa, that sounds weird), there *is* an established RFC for it, but you have to have the error checking around it. I mean, it's software. It doesn't take every manufacturer having to change a standard. Unless, I guess, a false ACK comes back from the unsupported router. I guess I'll have to read more about it. Damn you. But, hell, odds are even then the false ACK could be predictably detected. ehrm, yeah. [/nerd] ..damn your technie corner too. Why in the whole wide world of muffins would the auto-tuning in Vista start using larger packets without an ACK. Don’t get me wrong, optimizing TCP throughput is an awesome idea, and, in Microsoft’s defense (whoa, that sounds weird), there *is* an established RFC for it, but you have to have the error checking around it. I mean, it’s software. It doesn’t take every manufacturer having to change a standard. Unless, I guess, a false ACK comes back from the unsupported router. I guess I’ll have to read more about it. Damn you. But, hell, odds are even then the false ACK could be predictably detected.

ehrm, yeah.

[/nerd]

..damn your technie corner too.

]]>