[qasdfdsaq]

Originally Posted by pip08456

So correct me if I'm wrong here.

You've put the case forward that it cannot be the browser unless it's faulty.(I can believe that with IE)

Yes, this is the case I'm putting forwards. A web browser that delivers data from cache for a page that was originally marked "do not cache", or delivers data from a page cache of one page when a different page is requested is seriously broken. Web browser cache should never affect speedtest.net results because speedtest.net results should never have been stored in cache in the first place. Even IE isn't that broken.

We all know that speeds fluctuate by the second and that's why speed test site send multiple data packets and average the overall speed.

Some speed tests average the overall speed. Speedtest.net does NOT average the overall speed. Speeds do fluctuate, but speeds delivered over a given line will never exceed the line's max speed. Local buffering *may* cause data speeds to exceed line speed, confusing the speedtest program. Perhaps they didn't take this (the interaction of AV software) into account.

Kaspersky confuses the speed test on a much bigger scale.

Yes.

You have no evidence or proof that any other AV prog does the same (check out one of the kaspersky forums sometime)http://forum.kaspersky.com/lofiversi...p/t158618.html

I do not, but again other firewall/internet security programs perform similar functions at similar levels in the OS/networking stack. Unless there is evidence to the contrary, I see no reason why any other program could be any less affected.

Your assumptions and reasons are based on what you know about Kaspersky as that is the only AV prog both you and your company support.

My assumptions are based on the knowledge that to do anything useful with incoming data, any AV, firewall, or antispam software, including KAV and ESET have to buffer and analyse it before delivering it to the application layer.

I know Kaspersky buffers the data, and something to do with the buffering affects Speedtest.net's results. I know ESET also buffers the data. Without any evidence to the contrary, one would logically assume that because both applications perform the action that is known to be causing the problem, both could potentially affect speedtest data in the same way. Unless you have proof otherwise, I maintain any antivirus/firewall software could cause this effect. For some programs it may be un-noticeable except under extreme or unusual load circumstances, but nothing is immune to it.

Your reasoning also has another flaw. Buffering has to be held on a PC somewhere, it has to exist somewhere to be retrieved. This is exactly what Kaspersky does with a speed test files, It puts them in a buffer or cache to be called on when required.

Over to you.

No, it buffers data, it does not cache data. Buffers and cache are completely different things. Yes, data has to be buffered somewhere and is frequently buffered over and over in many different places. I suspect Kaspersky buffers data and delivers it in a bursty manner sufficiently so that it confuses speedtest.net's speed test app. The posts on the Kaspersky forums agree with my analysis. Buffering not caching.

A cache is completely different from a buffer. Once a data is delivered out of a buffer it is deleted from the buffer. It cannot redeliver the data later at a higher speed because it is not there. A cache stores data in order to speed up later accesses. AV programs/firewalls like any other program WILL buffer data but should NOT be caching (web) data. So when people complain about Kaspersky "cache" on these forums the effect in question is actually buffering.

So my main point is Kaspersky buffers data, so does ESET. Buffering may affect speetest results, and both programs do it. We have proof Kaspersky's buffering affects speedtest, we have no concrete proof ESET's does not.

Neither Kaspersky, ESET or the web browser should be caching the speedtest data. So cache, regardless of where, should not affect your results, ever (except with a very broken web browser).

To understand this you need to know buffering is not caching, they are different things used in different places for different purposes. You seem to imply in your post they are one and the same. They are not.