[Outages] Theoretical vs real-world speed limit of Ping

Peter Dambier peter
Sun Jun 3 07:19:11 EDT 2007


Let us add analog and digital times:

My DSL-modem says 1 Meg uplink, 5 Meg downlink. Real speed is slower but
that is mostly artificial.

My router needs to assemble a complete packet before routing.

uplink   1480 bits 0.676 milliseconds
downlink 1480 bits 0.135 milliseconds

The DSLAM side should not add to this timing because it is the same
analog circuit. But my own switch should add 0.067 milliseconds for
10 Meg or 0.006 milliseconds for 100 Meg ports

Here is traceroute

traceroute iason.site.voila.fr
traceroute to star.site.voila.fr.wanadoo.fr (193.252.158.3), 30 hops max, 40 byte packets

  1  krzach.peter-dambier.de (192.168.48.2)  3 ms  2 ms  2 ms

my router and workstation add above the expected 0.9 ms

  2  * * *

DSLAM does not ICMP?

  3  217.0.81.118 (217.0.81.118)  46 ms (ttl=252!)  48 ms (ttl=252!)  47 ms (ttl=252!)

more than 1000 kilometers? I guess it should be some 160 kilometers over there and back.

  4  f-eb7.F.DE.net.DTAG.DE (62.154.16.222)  49 ms (ttl=250!)
     f-eb7.F.DE.net.DTAG.DE (62.154.16.218)  50 ms (ttl=250!)
     f-eb7.F.DE.net.DTAG.DE (62.154.16.222)  49 ms (ttl=250!)

  5  * * *
  6  * * *

  7  po6-0.ntaub201.Aubervilliers.francetelecom.net (193.251.126.153)  212 ms  202 ms  166 ms

That suggests a link via a geosyncronous satellite (72 000 kilometers including return path) :)


Well looking up routing tables in memory does take time.
Interleaving packets to stabilize medium speed for VoIP might
add another constant.

Both (3) and (4) are rather big routers with big tables.

Maybe I should get and try paris traceroute.


Cheers
Peter and Karin


virendra rode // wrote:
> I thought some of you might be interested in this stats.
> 
> 
> credit: pingdom
> 
> 
> Theoretical response time limits of Ping
> 
> So, if we assume that a ?ping? travels with the speed of light, which
> are the best possible response times we can get?
> 
> 10 km -> 0,067 milliseconds
> 100 km -> 0,67 milliseconds
> 1,000 km -> 6,7 milliseconds
> 10,000 km -> 67 milliseconds
> 
> (Note that Ping shows the time for a roundtrip, i.e. 2 x the distance.)
> Pinging between Europe and USA
> 
> The distance between New York and Paris is 5,839 km. For light to travel
>>from New York to Paris and then back again would take 40 milliseconds.
> 
> Maximum distance between two places on Earth
> 
> If you choose the shortest route, the maximum distance between two
> locations will never be more than halfway around the planet. Halfway
> around Earth is about 20,000 km.
> 
> Considering that Ping goes to a destination and then back again, the
> packet sent by Ping would travel 40,000 km, the equivalent of a trip
> around Earth.
> 
> That is 133 milliseconds.
> 
> Anyone who has tried to ping various servers across the world will know
> that this is a way better response time than what you can realistically
> get. So why is ping so slow?
> 
> Why Ping is slower in the real world
> 
> The examples above take place under ideal conditions that don?t exist in
> the real world.
> 
>    1. The actual distance traveled will be longer, more like zig zag
> than a straight line.
>    2. Repeaters, switches and routers will slow down transfer speeds.
> The more equipment the signal has to pass through (for example routers),
> the longer it will take to reach its target.
>    3. The actual speed of the signal will never quite match the speed of
> light. Even with fiber optics (glass) the speed of light is about 30%
> slower than through vacuum or air, and most of the distance covered will
> be through fiber.
>    4. A beam of light bounces around like a pinball when it travels
> along the thin (admittedly extremely thin) fiber, and that will of
> course make the distance traveled even longer.
> 
> With all this in mind, you should probably double the ?ideal? response
> times shown above for a more realistic target to aim at. It?s useful to
> know when there is room to push for better network performance, and when
> the actual physical limits set in.

-- 
Peter and Karin Dambier
Cesidian Root - Radice Cesidiana
Rimbacher Strasse 16
D-69509 Moerlenbach-Bonsweiher
+49(6209)795-816 (Telekom)
+49(6252)750-308 (VoIP: sipgate.de)
mail: peter at peter-dambier.de
mail: peter at echnaton.arl.pirates
http://iason.site.voila.fr/
https://sourceforge.net/projects/iason/
http://www.cesidianroot.com/




More information about the Outages mailing list