akaize
05-31-2001, 12:24 AM
http://icomputerservices.com/speedtest.shtml is a link showing about 50 hosting companies and thier speed.
What does this speed mean connecting to their site, or connection to their server?
How seriously should I consider this speed chart while selecting a webhost?
CrazyHostGuy
05-31-2001, 09:52 AM
Someone please correct me if I am wrong, but it looks like their ping tests were all from North America. What if you sell a hosting account to someone in the UK or Australia, or anywhere else? Wouldn't these results all be completely different?
If this is the case, then I wouldn't put too much weight in a list like this.
Anyone else?
Cheers!
Rob.
CRAVIS
05-31-2001, 12:28 PM
Take a look at :
http://www.tracert.com/
Their tracetoute report is more detailed because the server is tracerouted from many different servers located all over the world simultaneously.
Phoenix
05-31-2001, 01:15 PM
Choosing a hosting provider based on latency measurements isn't necessarily the best criteria. The difference between number one on the list and number eleven was almost 10 times the latency. But, latency is measured in ms-milliseconds. The difference between 9/1000ths of a second and 90/1000ths of a second can be measured by computers but isn't going to be noticed by a person loading a site hosted there.
Latency has nothing to do with server response time, or how fast a particular page loads on a visitors' machine. Latency is usually calculated by running a traceroute on 1 or more packets of data from border router to border router, and can be affected by a lot of variables including some providers using OSPF and others who don't use it, or providers who are not very well connected and require that traffic be routed through crowded public peering points or other less-savory routes. Also, how many hops involved affects the latency, as well as outages in one network that may slow others down as they have to route around that spot.
When it comes to latency, remember that it will also be affected by how well your customers are connected. You can plant your web server directly on the C&W backbone (that has the most direct connections to other networks) and if your site visitor is using a connection to the Internet that gets routed all the way around Robin Hood's barn before heading in the direction of a national backbone, they will have different latency than someone who is connected more directly.