[asterisk-biz] Experimental/new VoIP rate search engine.

[Kristian Kielhofner]

On 1/5/09, Nitzan Kon <nk3569 at yahoo.com> wrote:
> --- On Sun, 1/4/09, Kristian Kielhofner <kristian.kielhofner at gmail.com> wrote:
>
>  > However, if the carrier has IP endpoints in Amsterdam, why
>  > not test it from various points in Europe?
>
>
> This is a little pointless IMHO because routing inside Europe
>  is usually done via insanely high bandwidth European rings
>  where you literally get no latency. I just tested from our
>  Amsterdam server and latency to Germany was 8 ms, London
>  was 7 ms, and a "high" 20 ms to Nice, France.
>
>  In other words, if a provider tests well from one point in
>  Europe - chances are they'll be OK from anywhere within the
>  EU.

  Thanks for the information.  Now we know we'll only need one server
in Europe. ;)

>  The US is a totally different story. A west cost user may
>  have 20 ms to a California server, but 100 ms to an east
>  coast server. And even if two users have the exact same
>  latency they can and probably will travel over different
>  IP transit carriers (one will travel on L3, the other on
>  Cogent, for example). It's quite a mess when you really
>  look at the traceroutes. :)

  I think you are forgetting just how big the US is :).  For
comparison I'll use Google driving directions which of course isn't
exact but will serve for this example:

Amsterdam to Berlin - 410 miles
Amsterdam to Nice - 865 miles
Amsterdam to London - 331 miles
Stockholm to Rome - 1,585 miles
Los Angeles to New York - 2,792 miles
Los Angeles to Miami - 2,731 miles
New York to Tampa - 1,135 miles

  Sorry about the units ;).

  I just pinged New York to Tampa in 36ms via Level(3).  Considering
the distance these numbers are inline with what you are describing in
Europe.  Granted you are currently at least 6 hours (probably more
like 8) ahead of me and the traffic level is likely quite a bit higher
but I'd say this number (36ms) seems to be about what I would expect
during the day.  Of course I can run this again at 10am here...

  100ms really isn't that bad considering you are (quite possibly)
traveling almost twice the distance of any geographical path in
"Western" Europe.  That's also just about the worst possible case (as
far as geography) in the US 48.  The only thing worse would be
something like Seattle to Miami.

-- 
Kristian Kielhofner
http://blog.krisk.org
http://www.submityoursip.com
http://www.astlinux.org
http://www.star2star.com