said by Hank:
I know exactly what you mean about the .5 second difference in latency, I have seen a few things broken or misleading results because of it.
I do get a kick out of the revisions to the AUP, but remember those changes are for our protection.
I am still working on a line problem with by second DSL here. It will drop shortly as the outside temperature drops. Line stats look good, but it will suddenly go. Frontier CSR told me he was going have my modem replaced because of the temperature problem. Must be dust in his headset because that was his remark after I told him that have used 2 different 7500's and 2 different 7550's modems and they all exhibit the same problem. There's days I just shack my head and laugh.
I believe Aranarth must be trying to find the "good stuff" in that file. He has been rather short in his responses lately.
He won't find anything. It's made by a script that fills the file with complete garbage that just cannot be read by anything. If he wants to try to get it though, then he just needs to find the storage and connectivity to save the file and then try to open it. It's basically a data eater used for benchmarking. With that if OP has Wireshark going too (hopefully resetting it) he can see if there's any form of shaping, or if it was simply a network glitch which so far would be more probable, based on my experience with my own line.
That line problem too sounds like an integrity issue if it's dropping like that. I recently had Verizon out to repair a problem with a nearby business's line that keeps getting errors at certain times of the night. I determined some of the cause to be from street lamps, which they really can't do anything about, but a more recent problem was due to the drop between the street and the building getting a small open fault that turned from DSL errors, to DSL sync loss, to tons of static to no phone service. If it's a temperature change doing it then it's quite possible it's an imperfection somewhere along the line. A modem definitely won't fix that.
But yeah, .5 seconds definitely makes a difference especially for the slower web browsers out there or for database calls between servers and such. 500ms for example is bad enough as it is for games, but when doing HTTP fetch (Customer > Backbone > Datacenter network > server > Server processing delay > Datacenter Network > Backbone > customer) any bit of latency can definitely slow things down. Take YouTube's homepage. Page generation times are stupid fast because their site is quite optimized to run on their setup, however when it comes to loading images, if you have latency on any part of the transaction, a browser that loads one thing at a time will really see the hit. With big sites, loading speed is key in addition to quality. YouTube loads VERY fast when you're not biting latency. For some folks, all it takes is a competing site with similar content to load that much faster to lose a view. For example: MSN.com vs. Yahoo.com for home page choice.