Quote:
Originally Posted by pip08456
Perhaps you need to sleep on it as I cannot for the life of me imagine any reason you would have to remove ESET (although quite easy to do if you know how).
Still, the browser shouldn't be caching it, any browser storing said data in cache is faulty.
As it does then every browser is faulty on your basis.
|
No, my basis is any browser that caches data explicitly marked as "Do not cache" is faulty. Most browsers don't do this, I don't get where you seem to picking up that I'm implying all browsers are faulty.
My point on ESET is that if Kaspersky causes strange speed results, and ESET performs similar functions on the same layer then ESET could well be doing the same thing.
As for removing ESET - we provide Kaspersky as the supported solution, we cannot install Kaspersky if any other AV is already installed. For user coming in with ESET on their machines we either have to remove ESET and replace it with ours or user gets no support from us. Support policy, not technical reasons.
Quote:
Anything that intercepts, analyses, then decides what to do with incoming data is likely to add delays or changes to speed in which that data is delivered.
Quite correct as that's the way Kaspersky works, not ESET.(I've been using it for about 5 years so know it's little foibles.) Better add not the same version!
|
As far as I know there's only one way to do it. Intercept all incoming data and pass/do not pass it based on various rules.
Quote:
I don't actually believe AV "caching" is actually caching at all, simply redelivering it at fluctuating rates that confuses the speedtest application.
So it has to be the browser then ( except i the case of Kaspersky). How does the speed test site become confused on downloads and by what?(I have a good I deas of what you are going to say here but that is upload AKS)
|
No, it can't be the browser (unless a faulty browser). The site does not become confused. The tester application on the client PC gets confused. It uses some sort of weighting and does not measure average speed. Perhaps weighted peak speed over a certain timeframe or something. Certainly not the average speed over the whole test.
So if the AV/firewall delays data for one second, buffers it, and delivers twice as much in the next second then the app measures double the speed in the second second. The AV/firewall delivers data at the same overall average speed but in a much burstier manner giving higher instantaneous speeds followed by gaps. As the speedtest does not measure the average, but some sort of weighted peak or percentile, it'll display a much higher speed than actually delivered down the line. This isn't caused by the browser, last time I used a proxy cache on the network it gave me 800mbps, a browser cache would be even faster.
Quote:
I always see the speedtest data correctly being delivered through the network interface at the correct speeds, then delivered to the browser at strange ones,
So if you've ruled out the AV (notwithstanding Kaspersky) and it is not the browser what is it delivering the speed test at strange speeds?
|
The above is *only* in the case of Kaspersky AV. When I rule out the AV, the speeds delivered are fine, therefore it's the AV causing it. I've never seen a case of anomalous speeds on any machine not running an AV/firewall/filter, and on the ones that are, the anomalous speeds always disappear upon disabling said security software. So it isn't the browser.
Quote:
BTW ESET was running as normal during the tests I did and only clearing the browser cache returned true results later checked by gamefiles downloads.
I don't actually believe AV "caching" is actually caching at all, simply redelivering it at fluctuating rates The big "Give away" to redeliver at fluctuating rates there must by definition be a cache.
|
No. Cache is storing data to redeliver it faster on second access. This is buffering data and delivering it at a fluctuating speed the first time. There does not need to be a cache to do this.
Say you download a large file at peak time in about 10 minutes, when your modem config is not the limiting factor in your speeds, you'll get a fluctuating download rate with no cache involved at all. Some seconds you may get 500KB delivered, the next second you'll get 1.5MB simply due to congestion. The fluctuating speeds caused by AV programs are confusing the speedtest in this manner but on a much bigger scale.