|
chiel
Guest
|
 |
« on: March 05, 2010, 01:58:49 pm » |
|
I sometimes have very low download speeds available through my HE tunnel.
Would it be possible to make some dump files available on the tunnelbroker website? Something like a 10MB, 100MB and 1000MB file. Then I can test it also over ipv4.tunnelbroker.net as well as IPv6 (or make a nice little button for IPv4 and IPv6).
I know that some IPv6 website can be accessible over IPv4 and IPv6 but then you can't be sure about the routes to these sites.
|
|
|
|
|
Logged
|
|
|
|
|
brad
|
|
« Reply #1 on: March 06, 2010, 01:34:04 pm » |
|
I sometimes have very low download speeds available through my HE tunnel.
Would it be possible to make some dump files available on the tunnelbroker website? Something like a 10MB, 100MB and 1000MB file. Then I can test it also over ipv4.tunnelbroker.net as well as IPv6 (or make a nice little button for IPv4 and IPv6).
I know that some IPv6 website can be accessible over IPv4 and IPv6 but then you can't be sure about the routes to these sites.
How about having the site be able to provide links to those files on each tunnelbroker POP for v4 or v6? Easy enough to setup lighttpd or nginx on the tunnelbroker POPs for the HTTP server. |
|
|
|
|
Logged
|
|
|
|
|
chiel
Guest
|
|
« Reply #2 on: March 16, 2010, 08:52:17 am » |
|
Yep having tests files at each POP would be the best. Can someone from HE comment on this?
|
|
|
|
|
Logged
|
|
|
|
|
snarked
|
|
« Reply #3 on: March 16, 2010, 12:35:49 pm » |
|
Do we really need more than two multiple sizes? "small" and "large" seems sufficient.
Small will tell us how much the protocol's overhead interferes.
Large will be the best measure of sustained throughput.
Therefore, pitch the 100Mb option.
File contents should be random. One doesn't want to find a 1GB file of all zeros compressed into 5 bytes (the 1 byte value "00" and a 32-bit integer of repetition count).
|
|
|
|
|
Logged
|
|
|
|
|
broquea
|
|
« Reply #4 on: March 16, 2010, 04:12:08 pm » |
|
I'll toss around the idea, and let you folks know.
|
|
|
|
|
Logged
|
|
|
|
|
hisken
|
|
« Reply #5 on: March 16, 2010, 11:34:46 pm » |
|
I think 10 MB is way too small; sizes of 32 MB, 100 and maybe 1000 MB would be the way to go. Just like they've got it here: http://speedtest.bbned.nl/ |
|
|
|
« Last Edit: March 16, 2010, 11:37:05 pm by hisken »
|
Logged
|
|
|
|
|
|
|
chiel
Guest
|
|
« Reply #7 on: June 20, 2010, 10:27:41 am » |
|
Any update on this?
|
|
|
|
|
Logged
|
|
|
|
|
broquea
|
|
« Reply #8 on: June 20, 2010, 11:34:02 am » |
|
No plans to host any.
|
|
|
|
|
Logged
|
|
|
|
|
liuxyon
|
|
« Reply #9 on: June 20, 2010, 12:09:20 pm » |
|
I try for check ipv6 speed with download 100M file. it's speed only 80KB/s. |
|
|
|
|
Logged
|
|
|
|
gawul00
Newbie
Posts: 10
|
|
« Reply #10 on: June 25, 2010, 01:43:00 pm » |
|
Remember you loose a significant amount of bandwidth with the extra headers in 6in4 tunnels. Couple this with processing delay at your and HE end of the tunnel, results in "slow" speeds. This is why tunnels should be considered a transition mechanism only.
|
|
|
|
|
Logged
|
|
|
|
|
cholzhauer
|
|
« Reply #11 on: June 25, 2010, 08:48:16 pm » |
|
Is it the extra headers that take up the traffic or is it just the latency?
I'm in Ohio and the closest server to me is Chicago, which adds a pretty large penalty to everything.
|
|
|
|
|
Logged
|
|
|
|
ngjvjRbYM
Newbie
Posts: 23
|
|
« Reply #12 on: July 02, 2010, 05:28:39 pm » |
|
Have tested http://speedtest6.tele2.net/ (forced ipv6) and shows a good speed (almost full speed). An other site with test files. The result shows not full speed (500). ftp://[2001:7b8:3:37:20e:cff:fe4d:69ac]/pub/speedtest/ a.k.a ftp://ftp.bit.nl/pub/speedtest/ |
|
|
|
« Last Edit: July 02, 2010, 05:30:56 pm by ngjvjRbYM »
|
Logged
|
|
|
|
|
|
|
