ndt-users list created

[Richard Carlson <>]

Aris;

The NDT system fills a data buffer with printable characters and then 
repeatedly sends this buffer to the receiving host.  The sending host 
keeps track of the number of bytes it sent, and the time it took to send 
them.  These 2 values are used to calculate the measured speed.

The NDT data flows consist of a repetitive pattern of characters that 
may be compressible.  I would look into the VPN package to see if it is 
using compression techniques.

Right now I am not looking at ways to randomize the data generation 
process, but would do so if this proves to be a problem for a large 
number of users.

Rich

On 4/28/2010 4:03 AM, Aris Adamantiadis wrote:
> Hello,
>
> Currently the NDT server of BELNET is only accessible from our VPN. I
> tried to connect to it from home and got this result :
> TCP/Web100 Network Diagnostic Tool v5.5.4b-belnet
> click START to begin
>
> ** Starting test 1 of 1 **
> Connected to: ndt.belnet.be  --  Using IPv4 address
> Checking for Middleboxes . . . . . . . . . . . . . . . . . .  Done
> checking for firewalls . . . . . . . . . . . . . . . . . . .  Done
> running 10s outbound test (client-to-server [C2S]) . . . . . 3.59Mb/s
> running 10s inbound test (server-to-client [S2C]) . . . . . . 11.70Mb/s
> The slowest link in the end-to-end path is a 10 Mbps Ethernet subnet
> Information: Other network traffic is congesting the link
>
> The second test showed similar results. However, here are the sync
> information from my modem:
> DSL Speed       640 Kbps Upstream
>                 6144 Kbps Downstream
>
> the detected upstream capacity is 5x the available capacity and
> downstream is 2x.
> I tried with switch's ndt server:
>
> ** Starting test 1 of 1 **
> Connecting to 'ndt.switch.ch' [ndt.switch.ch/130.59.31.2] to run test
> Connected to: ndt.switch.ch  --  Using IPv4 address
> Checking for Middleboxes . . . . . . . . . . . . . . . . . .  Done
> checking for firewalls . . . . . . . . . . . . . . . . . . .  Done
> running 10s outbound test (client-to-server [C2S]) . . . . . 3.57Mb/s
> running 10s inbound test (server-to-client [S2C]) . . . . . . 4.75Mb/s
> Server unable to determine bottleneck link type.
> Information: Other network traffic is congesting the link
>   [S2C]: Packet queuing detected
>
> Now, this gives good results for downstream but still wrong for
> upstream. I finally tried to connect to switch from my desktop which is
> not behind the VPN:
> running 10s outbound test (client-to-server [C2S]) . . . . . 493.0kb/s
> running 10s inbound test (server-to-client [S2C]) . . . . . . 4.80Mb/s
>
> which looks correct.
> Have you got any idea why the results are wrong when behind a VPN ? We
> are using openvpn, directly installed on my linux laptop.
> A colleague of mine had exactly the same kind of strange capacity
> results when connecting from home with the VPN.
>
> Will upgrading fix this ?
>
> Thanks,
>
> Aris

--
Richard Carlson
DOE ASCR  --  SC-21.1 / GTN
1000 Independence Ave. SW
Washington, DC   20585
P: (301) 903-9486