Re: [NNagain] are you Bill Woodcock?

[Bill Woodcock <woody@pch.net>]

> On Jan 19, 2024, at 08:14, Sebastian Moeller <moeller0@gmx.de> wrote:
>> On 18. Jan 2024, at 23:38, Bill Woodcock via Nnagain <nnagain@lists.bufferbloat.net> wrote:
>> So, if one Internet user wants to talk to another Internet user, generally they hand off their packet to an Internet service provider, who takes it to an exchange, and hands it off to another Internet service provider, who delivers it to the second user.  When the second user wants to reply, the process is reversed, but the two Internet service providers may choose a different exchange for the hand-off: since each is economically incentivized to carry the traffic the shortest possible distance (to minimize cost, speed x distance = cost), the first ISP will always choose the IXP that’s nearest the first user, for the hand-off, leaving the second ISP a longer distance to carry the packet.  Then, when their situations are reversed, the second ISP will choose the IXP nearest the second user, leaving the first ISP to carry the packet a longer distance.
> 
> I would propose a slight modification, "each is economically incentivized to carry the traffic the shortest possible distance" is not free of assumptions... namely that the shortest path is the cheapest path, which is not universally true.

Correct.  That’s a simplification of a complex field where distance and cost are frequently intermingled, and routing decisions are typically based on latency, overridden by cost as a matter of policy.  However, in a simplified or idealized case, if speed is held constant, distance and cost scale together, so they are usually held to be interchangeable in decision-making in the general case.  Speed x distance = cost.

> My personal take is "routing follows cost" that is it is money in the end that steers routing decisions

Yes, exactly.

The primary case in which routes follow a cost that differs from distance is in the preference for distant downstream transit over nearby peers, and distant peers over nearby upstream transit.  Though it’s uncommon in networks of small geographic scale, most global-scale networks do this, and it’s the cause of many routing problems and loops.

> ...at least once we include paid peering...

That’s a marketing euphemism for transit.

> My ISP aggregates its customers in a handful of locations in Germany, Hamburg in my case while I actually live a bit closer to Frankfurt than Hamburg, so all traffic first goes to Hamburg even traffic to Frankfurt (resulting in a 500-600 Km detour), I assume they do this for economic reasons and not just out of spite ;) 

Essentially all mobile network operators do this.  It’s generally a matter of incompetence and lack of competition, rather than spite or economic reasons.

> Now, maybe the important point is, this does not involve IXPs so might be an orange to the IXP apple?

Yes, only indirectly.  Most of what you’re discussing involves non-optimal outbound IXP selection, one quarter of the round-trip path.  Very real issues, but not anything an IXP or receiving-side ISP can do much about without second-guessing routing decisions to an impractical degree.

                                -Bill