The MPLS WG Archive

[Randy Bush <randy@psg.com>]

> Thanks for the info. Is it the data between two backbones over one NAP?

the picture was taken over something like the following route

randy@r00.dnvrco01.us.bb> traceroute e01.overtoom.ripe.net    
traceroute to e01.overtoom.ripe.net (193.0.0.25), 30 hops max, 40 byte packets
 1  p1-1-0-0.r01.dnvrco01.us.bb.verio.net (129.250.3.62)  0.534 ms  0.489 ms  0.457 ms
 2  p1-2-0-1.r00.kscymo01.us.bb.verio.net (129.250.2.78)  11.269 ms  11.237 ms  11.257 ms
 3  p1-6-1-1.r01.dllstx01.us.bb.verio.net (129.250.2.217)  22.848 ms  22.899 ms  22.862 ms
 4  p4-2-0.r00.dllstx01.us.bb.verio.net (129.250.3.73)  22.818 ms  23.322 ms  22.811 ms
 5  209.245.240.153 (209.245.240.153)  86.399 ms  86.287 ms  86.496 ms
 6  so-6-0-0.mp1.Dallas1.level3.net (209.247.10.101)  86.477 ms  86.573 ms  86.574 ms
 7  209.247.10.42 (209.247.10.42)  92.251 ms  92.358 ms  92.316 ms
 8  209.244.3.189 (209.244.3.189)  165.139 ms  165.511 ms  165.184 ms
 9  loopback0.mp1.London1.l3.net (212.113.2.11)  161.098 ms  160.528 ms  160.053 ms
10  212.187.128.90 (212.187.128.90)  168.302 ms  168.535 ms  168.703 ms
11  loopback0.core1.Amsterdam1.Level3.net (212.72.32.1)  168.208 ms  168.320 ms  167.960 ms
12  loopback0.ams-ix.Amsterdam1.Level3.net (212.72.32.8)  168.958 ms  168.964 ms  168.704 ms
13  Amsterdam1.ripe.net (193.148.15.68)  170.335 ms  170.670 ms  169.932 ms
14  fe20.pampus.ripe.net (193.0.0.246)  170.295 ms  170.049 ms  170.053 ms
15  s10.overtoom.ripe.net (193.0.0.53)  171.447 ms *  171.890 ms

remembering that, due to hot potato (hmmm, time for lunch), symmetry should
not be assumed.

> Is it generally the same for all inter-provider traffic?

of large providers.

> I meant AS's. I first saw the data in http://moat.nlanr.net/ASPL/. I
> chatted with one of my colleagues in Bell Labs (he had gone already) a
> couple of months ago. He was working on inter-domain network topology
> stuff, and collected bunch of more up to date traces. He told me that
> the average AS length in his data was 5-6.

i can believe that is what is in the routing tables.  i suspect that the
average number of ass traversed by a packet is less.

> My point was that even though major transit ISP's (Verio and UUnet) can
> control link congestion and transit delay within their networks, can all
> your customers (regional ISP's) provide the same level of service to end
> users? If no, what can we do about it? One solution may be to create a
> MPLS tunnel with bandwidth guarantees from one regional ISP over the
> backbones to another regional ISP...

and this will manufacture bandwidth?  and this will scale for a few thousand
non-giant isps?

randy