• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

How to accurately measure Internet connection speed?

C

Cooky

Golden Member
Lately I've been having performance issues w/ my Comcast cable modem.
I can tell by the response time I get when browsing sites I normally go.

Several speed testing sites yield my speed as 1.5Mbps / 350Kbps tops, rather than the advertised 6Mbps down. I know they're not 100% accurate but at least I get an idea what the speed is like.

So...my question is, how can one accurately measure his connection?

If I ping my default gateway w/ 1500 bytes of ICMP packets, and average round trip time is 200ms, is it correct to say the speed would be:
(1500 / 200) * 10 / 2 = 37.5 bytes / sec?
 
i think your method is sound. you've tried multiple services, and you're getting around the same speeds. give comcast a call and see if they've accidently provisioned you for the wrong plan.
 
Jack, I've used some of the sites you suggested for measurement and I'm asking for a more accurate way to measure my speed.
Also, I don't really feel like tweaking the TCP/IP stack because I have multiple hosts and some other reasons.
If the root cause is sub-optimal routing or physical error on Comcast's end, tweaking the TCP/IP stack may not help much.

nweaver, as far as I know, latency is the time it takes for a packet to traverse from point A to point B (ping result), and it can be used to calculate speed.
If it takes x amount of time to transmit y amount of data, isn't speed result of y / x? (assuming delay is non-factor)

 
Modify your stack and see what the results are.

Also you can in no way determine through put from a ping. There's too much delay in there to make an accurate measurement. Plus routers don't like answering pings as fast as they like forwarding real packets.
 
thanks everyone for your input.

so...does anyone know of a better way to measure speed other than those testing sites?
It sucks to have a residential pipe that's not running well...cause I have no SLA to chase the carrier.

I'll modify my tcp/ip stack as my last resort.
Everything was working well till two days ago...tried w/ a Pix501, Cisco871, and even bypassed all hardware and went straight into cable modem, same thing.
Everyday I pray to have FIOS in my area and still....
 
ok, all of a sudden my speed is back to normal.
I know I didn't make any changes so it's gotta be some where between my modem and Comcast...

Right now average ping time is 15 ms w/ 1500 byte size of ICMP packets to default GW, so the way I see it, my speed is:
(1500 / 15) * 10 / 2 = 500 bytes /s = 4Mbps / s, which is about the speed I was getting before.

Spidey, I know there's propagation delay and queuing delay along the line, which is why I said "assuming delay is non-factor".
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top