Broadband Speed Tests

[Andrew Mawson]

Anyone any idea how BT's broadband speed test works? (

)

I was having a problem yesterday, my download speed being reported as 0.33 Mbs but the 'local connection' was about 5Mbs according to the BT test. This was with the PC plugged directly into the Home Hub ethernet port, and the phone line directly into the hub via an adsl filter eliminating all my lan wiring and internal phone wiring. I convinced myself that the Home Hub had failed, and this was what the BT test suggested. This morning speed was back to normal with me having done nothing!

How do they test the path from exchange to hub and separate the results from the hub to PC? (As I think that they got it wrong this time!)

AWEM

[Dave Liquorice]

[Dave Plowman (News)]

I get 17053 kbps here using that test - yet things still sometimes seem slow.

[Adam Funk]

Just add a few filter subscriptions: I have EasyList, Liste FR, EasyList Germany, Fanboy's Adblock List, & Fanboy's Annoyance List, & I can't remember the last time I had to add a filter myself.

[Huge]

I use AdBlock+ with Fanboy's Adblock list, Ghostery, Flashblock and a decent sized hosts file (16,548 entries) and I can't recall the last time I saw an advert on the web. And with spam filtering at my ISP, on my firewall (Smoothwall) and in the mail client and the amount of spam I see is negligible, too.

[PeterC]

Streaming speed about 2100kbps; d/l speed 4800kbps! - that's crap on a

2.7Mbps connection and no such level showing on NetMeter.

Judging by the actual playing quality, the streaming is accurate.

[PeterC]

Having a goodly Hosts file keeps out a load of rubbish as well.

[tony sayer]

Well should be around 4 to 5 Meg >>

Well that might turn up one day ...

[Dave Liquorice]

At a mile I'd expect >5 nearer 6Mbps sync rate. Our bit of string is over two miles and we get 5.5 to 6Mbps. That .35 up is suspicious, the uplink uses a lower set of carriers and almost invariably runs at the max of

448kbps sync. But is that .35 as reported by a "speed test" or the sync rate reported by the modem? A speed test will be lower due to the transport overheads.

[tony sayer]

Precisely .. we have had BT changed for someone else more than the once with better results;)..

Bit of string is the operative word. Unless your sure of the cable type and the exact run of it, how much is overhead what type it is under ground let alone the joint quality etc etc;!...

[charles]

and does the circuit go the direct route?

[Tim Lamb]

Speed test result 10am, 4738kbps down and 343kbps up. Good enough for the current use.

>

[Tim Lamb]

In message , charles writes

No!

Last time Open Reach had a look the circuit in use continued past the farm for about 200m and then returned! The actual fault was at the exchange:-(

>

[Dave Liquorice]

There is the problem, once you get a few hops into the route to any destination who is to say if it's one of those hops overloaded or a slow/overloaded server?

I'm not sure where this "50 user" contention occurs. I've never seen it but then being rural on an exchange with only 1200 subs and 300+ subs to the community wireless broadband that's not really surprising.

Having said that from what I've seen from my ISP (who pledges never to be a bottle neck) says when they spot a full connection within BT, BT are pretty quick in bringing in extra bandwidth, like less than a week. So I suspect those suffering slow downs at peak times are suffering purely due to lack of capacity in their ISP rather than BT or the internet in general.

[The Natural Philosopher]

No, but of the existing ISP is unresponsive to request to get BT out, a new one may not be.

If the line is noisy,

Sometimes you need to report it as both.

Crackly kit INSIDE te exchange cant even be accessed by the normal chap who does broadband faults.

I went through each in turn. Voice (gave me new lines back to the crossroads) Broadband(gave me new lines all the way from the crossroad to the exchange. The problem was still there, then, three days later it 'went away all by itself' I suspect a dodgy dslam was 'replaced'

[The Natural Philosopher]

Absolutely. I used to get reliable synch at 3-3.8Mbps, and struggled to get a BRAS over 3Mbps.

Currently after having hand selected cables replacing my overall link, I have 4 db less attenuation and when my old router died, the new one is now synching at 6Mbps and I get a solid 5Mbps BRAS.

actual recorded sppeds are about 4.8M down and about 376k up.

[The Natural Philosopher]

But that costs money.

Which some ISPs are unwilling to pay..

So I

It depends. traceroute can reveal where the issue is. Once it was appearing inside my ISP, so I called them 'yes, we have lost a fibre peering link and all traffic is going down the other one'

They had that fixed in a couple of hours which is pretty good really.

[Tim Streater]

Yes.

Also remembering to get your in-house phone wiring tidied up, in particular get the bell-wire removed.

And *also* remembering that chances are your targeted website won't be hosted by your ISP, but by another. Traffic on the other ISP is their responsibility, not that of your ISP, who can't do much about how it performs. If the website is in another country, or if the routing in some part of the Internet is non-optimal, the traffic can take strange paths and get slowed down even more. The only reliable figure in the game is the speed to your local exchange, as reported by your router.

[Rick Hughes]

Biggest impact is the number of joints combined with the age of them.

[polygonum]

I gave up smoking (anything) several years ago. And yes! My network speed has improved!