cat2 performance

[Phil]

We moved our main site here some time back because the transfer rate performance of the cat2 servers was MUCH faster than our previous hosts (the truely dreadful Fasthosts). We still have some sites on Fasthosts and monitor both via Webperf.net. At the time we moved our site, cat2 was averaging around 1800 kb/sec, whilst Fasthosts was staggering along at around 600 kb/sec. In November Fasthosts obviously upgraded their systems as all of a sudden they are now averaging around 1300 kb/sec. Unfortunately at the same time cat2 seem to have slowed to around 1450 kb/sec (although it has dropped as low as 1300 kb/sec). So now there isn't that much difference in transfer rates between cat2 and Fasthosts. I hope cat2s performance doesn't decline any further - perhaps cat2 need an upgrade too?

[paulredpath]

Phil,

I think you will find in real world performance we will beat fasthosts hands down

Webperf isnt a particularaly accurate measure as its uses extremely small files and many hosts tune dedicated boxes for just that purpose. We host our webperf test file on a normal webserver, with customers on it. Hence ours gives a more accurate view than many would have you believe.

We have over 500mbit/sec of transit and run our own network, fasthosts are also based out of london so will give higher pings

[Phil]

Originally Posted by paulredpath

Phil,

I think you will find in real world performance we will beat fasthosts hands down

Webperf isnt a particularaly accurate measure as its uses extremely small files and many hosts tune dedicated boxes for just that purpose. We host our webperf test file on a normal webserver, with customers on it. Hence ours gives a more accurate view than many would have you believe.

We have over 500mbit/sec of transit and run our own network, fasthosts are also based out of london so will give higher pings

The figures I am quoting are for our own sites (that we have added to webperf), so they are not on dedicated boxes and are 'real world performance'. I hope cat2 will continue to 'beat fasthosts hands down' but it seems to me that the difference between the two (at least in terms of transfer rates) is getting worryingly smaller....

[paulredpath]

What are they listed under webperf as?

Also please remember that means you can download a small file at over 140k/bit sec, which is 1.4megbytes/second. How big are your webpages if you need faster than that

[Phil]

Originally Posted by paulredpath

What are they listed under webperf as?

Also please remember that means you can download a small file at over 140k/bit sec, which is 1.4megbytes/second. How big are your webpages if you need faster than that

It has nothing to do with the size of our web pages. The issue I am trying to bring to your attention (and you seem to be trying to avoid) is that of the reduced transfer rate performance of the 'pineapple' webserver. I would have thought that it would at least merit an investigation by you to determine exactly why this fall in performance is happening. Our site is shown on webperf is '1st Class Software'. It is interesting to compare our figures with the catalyst2 results which are MUCH faster (currently running at 2500kb/sec). So presumably there IS an issue with pineapple. If not can we be transferred to the server which hosts catalyst2?

[paulredpath]

Phil,

The real question is do you notice any speed issues with your site? Our test file is on a linux server which are usually faster at serving small items than iis. There are also many other variables to take into account, and due to the way webperf works a few milliseconds difference in response will show a much different transfer rate for a server.

I dont see why webperf results would cause an investigation by us unless clients were actually reporting issues with the server, which noone has done.