Still don't understand

[sevenworth]

Linux 1 - Windows 0

*duck* !

And you said everytime it's a new install of Windows ?

Well that was one of the things I tried in the beginning, a completely fresh install of windows, but alas, it didn't stop the devious background processes that are seemingly built into windows itself, and not added by the user...

 

[Prometheus]

Browsed the web on Wednesday night using hexpee for approximately 2 hours - usage totaled about 20mb. Browsed the web last night using linux for approximately 2 hours - usage totaled about 9mb. I know this is not an accurately defined test, but I know by now how much bandwidth is used during normal browsing in xp, and it just seems so much less in linux. Background applications must be at work in xp that use unnecessary bandwidth, where in linux, nothing happens without your say-so. People, this is my solution for the "heavy-bandwidth-usage" problem: switch to linux. Period.

Could also be that Linux is making use of compression by default where Windows doesn't want to even when specifically enabled.

Could have sworn I replied to the v4me issue... like I said neither the standard nor the lite version works for me.

 

[vodacom3g]

Well that was one of the things I tried in the beginning, a completely fresh install of windows, but alas, it didn't stop the devious background processes that are seemingly built into windows itself, and not added by the user...

The one thing that still worries me is the burst of activity we saw when you ran the first Google search in FF.

For all: sevenworth ran a search on Google using a PC that is LAN-connected to the system with the data card.

The first time he ran the search, we saw a 1Mb burst of data coming down. Subsequent (different) searches only downloaded the data of the search results page.

A toolbar or some other function in FF?

 

[Prometheus]

Or images, scripts and stylesheets and the other searches only downloaded the page itself.

 

[sevenworth]

@prometheus:

not sure what you're talking about, google has no scripts or major images, at least nothing that could possibly total about a meg. also, it's not the browser thats at fault. the browser does what it's told to do... it's background processes. don't blame google, they've worked hard to make a lightweight, yet highly accurate search tool.

 

[vodacom3g]

What about FF plug-ins?

 

[sevenworth]

What about FF plug-ins?

no such things involved. it was a fresh install, so not much was added at that point. <edit>I'm happy that I've found a solution in Linux. I have far better control over what's going on. it's not all cloak and daggars like in windows...</edit>

 

[Tazz_Tux]

Can you run it thru a proxy server and see what is up ? There are quite a few free proxy servers out there, install it on the same machine and set both ie and ff to use the proxy. Many apps use a fake session of ie to do the comms. The logs will show it all

Laterz !

The one thing that still worries me is the burst of activity we saw when you ran the first Google search in FF.

For all: sevenworth ran a search on Google using a PC that is LAN-connected to the system with the data card.

The first time he ran the search, we saw a 1Mb burst of data coming down. Subsequent (different) searches only downloaded the data of the search results page.

A toolbar or some other function in FF?

 

[kwmf]

Tried using Ethereal/WireShark to do a packet capture of the session to see what's going up and down?

 

[Prometheus]

@prometheus:

not sure what you're talking about, google has no scripts or major images, at least nothing that could possibly total about a meg. also, it's not the browser thats at fault. the browser does what it's told to do... it's background processes. don't blame google, they've worked hard to make a lightweight, yet highly accurate search tool.

Not blaming anyone here. All pages load more data the first time. If the high usage coincides with the loading of a page it's safe to assume background processes are not to blame. Just saying there has to be a reason you're using more on the first run. Maybe ask Netstar to track down the missing bits.

 

[sevenworth]

Not blaming anyone here. All pages load more data the first time. If the high usage coincides with the loading of a page it's safe to assume background processes are not to blame. Just saying there has to be a reason you're using more on the first run. Maybe ask Netstar to track down the missing bits.

All pages load more data the first time... well google's front page isn't made up of more than a few k. As previously mentioned, we've done the obvious tests to cancel out that the browser and requested webpage are to blame. there is something else at play here, otherwise simply, I wouldn't have come this far trying to figure out what's going on. There are 6 machines on this network. I'm not about to blame a popular website like google, especially when comparative tests have been run. Something else is happening on this network that regularly causes large chunks of data to come in at breakneck speeds. But I can guarantee you it's not google, and its not firefox.

 

[Kneebough]

ok. I cleared my cache and set google as my homepage and *bam* suddenly firefox downloads at 50kB/s for about 10 seconds. Wierd.....

 

[Tazz_Tux]

sevenworth,

Just re-explain the network layout...I might have an appy that can solve the ever lasting mystery...

All pages load more data the first time... well google's front page isn't made up of more than a few k. As previously mentioned, we've done the obvious tests to cancel out that the browser and requested webpage are to blame. there is something else at play here, otherwise simply, I wouldn't have come this far trying to figure out what's going on. There are 6 machines on this network. I'm not about to blame a popular website like google, especially when comparative tests have been run. Something else is happening on this network that regularly causes large chunks of data to come in at breakneck speeds. But I can guarantee you it's not google, and its not firefox.

 

[Tazz_Tux]

ok. I cleared my cache and set google as my homepage and *bam* suddenly firefox downloads at 50kB/s for about 10 seconds. Wierd.....

What about RSS feeds, what browser did you use ?

 

[sevenworth]

sevenworth,

Just re-explain the network layout...I might have an appy that can solve the ever lasting mystery...

Option Nozomi in PCMCIA to PCI bridge - desktop pc win/linux - wired to switch.
Windows box wired to switch
Windows box wireless to AP
2 x Notebook PC's using windows wireless to AP.

Yet I still get much lower usage when I dial up from linux, even with all the windows machines using the connection...

 

[Tazz_Tux]

So the let's say you open www.google.co.za on the Windows box connected to the switch...do you still see higher usage in Windows compared to Linux...

Option Nozomi in PCMCIA to PCI bridge - desktop pc win/linux - wired to switch.
Windows box wired to switch
Windows box wireless to AP
2 x Notebook PC's using windows wireless to AP.

Yet I still get much lower usage when I dial up from linux, even with all the windows machines using the connection...

 

[sevenworth]

So the let's say you open www.google.co.za on the Windows box connected to the switch...do you still see higher usage in Windows compared to Linux...

At work now, will check this evening. Although, from what I remember in previous testing is, yes, usage is higher from windows all-round. The windows box on the switch also has a linux partition but spends 99% of it's time in windows. When in linux on that particular pc, usage is quite drastically reduced. All the while I'm wondering if linux is giving me accurate figures... . On that note, I take it that when monitoring pppo, Network monitor does capture the entire networks usage of the connection...? Must do... I have watched usage while my linux machine was idle.

Also found that Media Player is responsible for 'usage you don't realize'. It searches and collects content and information on your music. Say your collection is made up of 200 albums, information on each album and each track is collected from amazon or wherever. Not a major culprit, but one of them in any case.

I've also noted that some months, overall usage is higher than other months, when what I actually do with my internet connection on a daily basis remains generally the same.

 

[Kneebough]

What about RSS feeds, what browser did you use ?

Base Firefox 2.0 (no rss feeds running) and new XP installation. Never had this happening on my lappy but now that I'm connecting using my desktop, it does this. Interesting.

 

[Tazz_Tux]

Ok - here is the master plan

Your Gateway - we will call GW for now. Your Windows box connect to the switch is called D1 and the wireless ones is called W1 and W2.

Ok - test to try: (Keep the other machines off)

GW in Windows and connected. D1 in Windows. Download www.google.co.za
GW in Windows and connected. D1 in Linux. Download www.google.co.za
GW in Linux and connected. D1 in Windows. Download www.google.co.za
GW in Linux and connected. D1 in Linux. Download www.google.co.za

(Notice how the Linux on Linux is shorter....that must tell you something )

You can record (on the ppp0 int) the traffic with tcpdump (Windows and Linux util). Then we can use Ethereal/Wireshark to tell what is going on.

Let solve this one once and for all

Laterz !

At work now, will check this evening. Although, from what I remember in previous testing is, yes, usage is higher from windows all-round. The windows box on the switch also has a linux partition but spends 99% of it's time in windows. When in linux on that particular pc, usage is quite drastically reduced. All the while I'm wondering if linux is giving me accurate figures... . On that note, I take it that when monitoring pppo, Network monitor does capture the entire networks usage of the connection...? Must do... I have watched usage while my linux machine was idle.

Also found that Media Player is responsible for 'usage you don't realize'. It searches and collects content and information on your music. Say your collection is made up of 200 albums, information on each album and each track is collected from amazon or wherever. Not a major culprit, but one of them in any case.

I've also noted that some months, overall usage is higher than other months, when what I actually do with my internet connection on a daily basis remains generally the same.

 

[sevenworth]

Will do.