How to test the speed of a server?

[John]

Testing on a single PC doesn't really help. And their statement of how many backbone doesn't show the quality either.

Is there anyway to have some info on how fast they are across continents or the world?

[webhost17]

Try www.alertra.com, it should tell you results from all around the world.

[Timmy00]

Another site is www.hostingspeeds.com -- currently they only test from one location though. Don't know if they have plans otherwise or not.

[hostmaster]

If you have an account there you can use websitepulse...
They're currently giving away free beta accounts- http://beta.websitepulse.com

[HostingGuy]

Just test out 10 - a dozen web hosting. Tested on both clients and their main homepage .......... Some interesting Finding.

Most of the less expensive web hosting ( $4 per month ) are slower in comparsion to the more expensive one. ( tested on client we

But all these less expensive web hosting actually have the fastest server for their own homepage. ^^"

Europe access are averge 50 + % slower than USA to USA server.

None of the hosting managing to get more than 100Kbps to Hong Kong. And it gets worst. Some less expensive hosting have a mere 10Kbps access rate.!

By the way, for thoe who dont know much about internet these test doesn't actually mean anything. Since it depends on the time of day at the server's location, the testing location. And many other factors...... But it will still give you some general idea how it perform compare to others.

[Jarrid]

Originally Posted by HostingGuy

By the way, for thoe who dont know much about internet these test doesn't actually mean anything. Since it depends on the time of day at the server's location, the testing location. And many other factors...... But it will still give you some general idea how it perform compare to others.

That is very true. Some times of the day you will have different speeds than other times. Really depends on the load of the network.

[Tony]

You can try to use ping or tracert commands to server and observe the results..

[Hostbiz]

dnsstuff is good one too

[Matt]

Originally Posted by Timmy00

Another site is www.hostingspeeds.com -- currently they only test from one location though. Don't know if they have plans otherwise or not.

Testing is done from many locations everynight by a few moderators & the admin this gives a better overall rating.

We already have locations in: Australia, 2 in Canada Canada (Calgary and Ontario), Germany (Cologne), Mexico, Norway, 2 in The Netherlands, 2 in United Kingdom (London Harbour Exchange and London Meridian Gate) and 6 hosts in USA (Dallas-TX, Denver-CO, Ft Lauderdale-FL, Parsippany-NJ, Pennsylvania and Green Bay-WI)

[-DataCenter-]

I also agree this very true - and a very common area for confusion. Here's everything you'll want to look at to understand the scenario:

- Ping Response - low pings are good, but remember they'll be consistently higher the further from the server you test. This does make a difference, but there are more important things to consider.

- Packet Loss - when you ping the server, do you see any packet loss? Packet loss is very bad, and the lost effeciency can very quickly outweigh all other traits of the network.

- Bandwidth - In dedicated sales, companies typically put up sizable test files for testing bandwidth, to be downloaded from a server at a different location. By doing this, you are able to see how much of the 'pipe' you can use. If you have broadband at home, you can at least test if it's absolutely aweful, by testing your home connection's abilities at speedtest.com - and then watching what speeds do as you download a file from that server. You might want to test this at various times of the day.

- Network Mix - If the host advertises several carriers, it can mean a few things.

a.) One is used for outbound, or at least very heavily weighted on outbound, and the rest are all for inbound. This is a very, very, very common practice (especially for the very cheap facilities), and if their one outbound carrier blows a router or has a line cut, you can be in big trouble.

b.) They could have a mix of carriers set up right, which preferably chooses an ideal route. If you see the abbreviation "BGP", you're probably in good shape.

You won't be able to test this from a shared hosting account, but the host's website should tell you. Or, if not, they should be happy to tell you if you ask.

- Data Center Location - This isn't talked about a lot, but the relationship to the data center to the carriers' networks is also very important. For example, a data center that has all the carrier's based nearby, or in the same building, will perform great. A data center in a corn field that sends everything 100 miles before it can go anywhere else, will be neither efficient or gauranteed reliable.

- Lastly, you have the server itself to think about, because all of the items above actually only test what the server is plugged into. Unfortunately, you won't be able to do much to test this until you see it's performance yourself (assuming the host has a secured machine)