Virgin to give new Superhub out.

[qasdfdsaq]

Sorry misread. My brain's been fried of late.

My previous understanding was speedtest.net's speedtests didn't use normal HTTP requests to fetch data and had something proprietary. Either I was misinformed or things have changed, so now it uses standard browser requests.

Still, the browser shouldn't be caching it, any browser storing said data in cache is faulty.

Not familiar enough with ESET to know what it does - all the involvement I've had through my job has been removing it - but as it has a firewall, filtering, and interception capabilities I'd hazard it could contribute in the same way as Kaspersky. 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.

I don't actually believe AV "caching" is actually caching at all, simply redelivering it at fluctuating rates that confuses the speedtest application. 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, but the data definitely gets downloaded over the internet, not delivered from cache. Besides, cache would give you several hundred megabits to gigabits per second, not 50-200 meg like we're commonly seeing.

[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.

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!

BTW ESET was fully up and running during the tests I did and only clearing the browser cache returned a true result.

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)

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?

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.

[qasdfdsaq]

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.

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.

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.

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.

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.

[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)

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.

Kaspersky confuses the speed test on a much bigger scale.

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

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.

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.

[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.

[pip08456]

I'll grant you your last paragraph.

I'll argue on Kaspresky's buffering and how ESET does it though.

That though is a totally different argument and "off topic" so best left to another day, besides I'm kna****ed!

[qasdfdsaq]

Agreed.

I don't know the details of Speedtest's algorithm or ESET's, but I do know buffering affects the former.

I'll admit in some cases clearing the browser cache *may* help results such as in yours, but I really can't explain why, as caching should (in theory) not play a part at all.

[Ignitionnet]

The Ookla tester measures the download during several periods and barring some normalisation gives as its' result the fastest period. Any buffering, such as that from a virus scanner which buffers data to scan it before delivering it to higher level apps, will confuse it quite handily.

[starfish]

Originally Posted by whizzard

That KB info needs to be updated really. As stated earlier in the thread, as of today the superhub will be for all new customers on XL, XXL and XXXL, or those upgrading to XL from lower tiers.

I am an existing customer and I called up earlier today to ask for a V+ HD Box. Later on, I found a message from VM on my answering machine, saying that with the V+ Box they will be dispatching, a Superhub will be included as well - apparently, without including the Superhub the guy could not complete my order. He also said I don't have to use it and I can just put it in the cupboard for a possible later use! I am on 20MB and was not upgrading broadband on this occasion.

[jb66]

Originally Posted by starfish

I am an existing customer and I called up earlier today to ask for a V+ HD Box. Later on, I found a message from VM on my answering machine, saying that with the V+ Box they will be dispatching, a Superhub will be included as well - apparently, without including the Superhub the guy could not complete my order. He also said I don't have to use it and I can just put it in the cupboard for a possible later use! I am on 20MB and was not upgrading broadband on this occasion.

Must be a computer glitch

[MakaveLee]

My upgrade 50mb is being installed today so we'll see if i get a superhub or not.

[starfish]

Originally Posted by jb66

Must be a computer glitch

I guess, or a new unofficial policy to tempt customers to upgrade their broadband!

[matty1968]

Does anyone know if the new superhub will work with existing wireless routers. I'm on L BB, thinking about upgrading but dont want to ditch my existing router. Can the wireless be disabled in them? Will I need a x-over cable?

[chill_fm]

Originally Posted by Media Boy

An Virgin insider has told me that Virgin Media are to give ''new superhub'' to customers on XL, XXL and 100Meg.

But it will be for new customers or customers upgrading.

just upgraded mines to xl not receive or heard anything of this

[daveybm]

Son in law left Sky and got VIP50 package installed yesterday separate modem and router, asked for a dongle and was told "we aren't supposed to give these out but you can have this one" cheeky git! is this how they end up on ebay? also did not activate V-HD box, son in law not impressed especially as I told him it's a good outfit.

Dave