|
Post by BH Oblt. Firecage on Nov 16, 2011 23:21:16 GMT -6
Bandwidth usege on the Event Nov !6th ... barely touched the server on usage. 3.04 MBS isn't much at all. Problem has to be in the mission or ROF itsef. S! FC
|
|
|
Post by BH Oblt. Firecage on Nov 17, 2011 0:19:38 GMT -6
To add this..... THis is from the place we rent the server space at.
_______________________________________________________
Our VDSes come on mostly-unused gigE (1000mbps) connections, which means that they far outstrip the available capacity at speed-testing sites, which are sometimes overloaded and limited to much lower speeds. These sites are meant for testing fairly slow home broadband connections that are in the 10mbps range, and they are tweaked and configured with those in mind.
To give an example, you might connect from your VDS in Dallas to a speed site that has a shared 100mbps connection to the internet. The ISP running the test site allows 10 clients to run tests at once, since they realize that most broadband lines are less than that, and the few that are more will generally fit into the remainder (they also know that they could allocate more bandwidth, but bandwidth is expensive). When you connect, 9 other users might be running tests, leaving only 10mbps left out of their small link -- and making it look like your VDS can only handle 10mbps!
In reality, your server, being on a barely-used 1000mbps line, would completely dwarf even what they could test with nobody else connected (100mbps).
As another test of speed, you could download a large file from a nearby, known-fast download host, such as a close university with an underloaded FTP, or from a mirror of kernel.org. You will likely see bandwidth speeds in the 5-10 megabyte per second (40-80mbps) range if you do this. The reasons for the realized speed still being less than 1000mbps are twofold: One, you're probably still faster than their available capacity; and two, TCP connections across the internet are limited by what's called the "bandwidth delay product" and TCP Window size.
_____________________________________________________ Just FYI for info.
S!
FC
|
|
|
Post by BH Oblt. Firecage on Nov 18, 2011 11:05:35 GMT -6
We are very much on Par with them right now since we upgraded our server to a 4 core system. For what we have now it should be able to run a crap load of players. Butzzell has redone the missions and has stripped out quite a bit of stuff and changed alot of how the mission works to help reduce the warping we saw this week. The server did not have any points where it was even close to a max out on horsepower or bandwidth. Butzzell is doing his best to get things working correct and we just need to be patient and help him out as we can. This is also a great learning experience for us to see what problems a mission can create. Something to know as well..... I have been on the server doing speed to tests to different areas of the USA and its amazing to see that some providers seem to have limiters on their flow rates so you can only get a low max through put. example : our server to Jacksonville FL. To Greensboro SC (about the same distance) To Houston Texas To NY City/ NJ area As you can see ... depending on where your at and what you provider is doing you may have bandwidth way less than you think. Now what we had in the mission was not this problem. I am just saying there are a ton of routes the data takes and everyone does not have the same size of feed pipe. As you can see on the houston test the server can crank out some serious bandwidth if asked to. Hope this helps see a bit more of the picture. S! FC
|
|
|
Post by BH Lt. von Schmidlak on Nov 18, 2011 19:00:41 GMT -6
hahahahaha! ;D
|
|