I gather you do not have 'internal' DNS? Latency of pings to many sites, and the download of a file will tell you much about your network. Google DNS is scaled to a very large audience, I believe someone has already mentioned they use it and it is a good fall back. Good you changed DNS, you never know how many users are on the same. But still congestion can provide bottlenecks. They measure the time to download the file. If you really want throughput, try a download test, usually 10M,100M,1G, file downloads. Internal ISP speed tests tell you -WAN - local - result, if the bottle neck is outside of their - local - WAN (ISP's 'domain') then no real info. Rerouting by these sites can improve stat's but does not tell you of your connection to remote sites. Speed test sites that push from a server to your local site can run into problems running through routers on the way to you. The question is much too general without more context and information.Ĭongestion at a site that is not 'local' to your intranet, how can you see what is going on - with another connection. Have you tried to ping the destination site to see what those results look like? Traceroute? What type of connection do you have (DSL, Fiber, Cable.etc)?Īre they downloading files and that is what is feeling slow? If so, do you have a perimeter AV or firewall that is doing some kind of file scanning or proxy that could be slowing things down? If the destination point is experiencing some kind of congestion or issue, a speed test won't reveal this but the user experience will still be much poorer than what they're used to.Īre your users complaining of slow "internet" from a particular site or from all sites? There are dozens of reasons why the "internet" may seem slow, not the least of which is the site(s) you are trying to access. It sounds like the real problem you (OP) are trying to diagnose is the slow internet when the results appear to show a connection that is operating at best speed. It's definitely true that latency will kill the throughput but I'd start with a bunch of questions to narrow down the problem first. If testing from the nearest server is not useful then neither is testing from the furthest possible server either. I'm sorry but I fail to see the point of this. Try running the test from a server 4,000 miles away and 10,000 miles away, your local internet speed doesn't change but your throughput definitely will
this is not always real world scenario because the site you are trying to access may require more hopes to reach with much higher latency.
0 kommentar(er)
