This mailing list is no longer active and has been transitioned to email@example.com. Members of the I-coordination mailing list have been moved to the new mailing list. To learn more, visit 1net.org.
[I-coordination] A different model
Lee at asgard.org
Mon Dec 16 16:02:54 CET 2013
>> Related: each network operator decides what routes to carry, and how to
>> decide what routes they will carry. I don't see how (or why) the
>> government of one country can affect the routing tables of network
>> operators in other countries.
>Yes and no, in some cases you may have agreements that as an operator
>makes you carry prefixes that do not belong to you (transit is a clear
>example) and here is where some fat fingers screwed up and continue to
>screw up, announcing (some times intentionally) wrong prefixes that get
>propagated to some portion of the overall network table through operators
>with lax routing policies and management.
Those agreements are made (or not) by the entity running the network, for
its own reasons. It may enter into contracts (e.g., for transit), and may
influence the terms. In any contract, there is some form of
consideration, which the parties like or they do not contract.
Sure, fat-fingering happens. It's quite different than state-mandated
carriage of routes originated from elsewhere, which was the message to
which I was responding.
>There is currently a lack of good/strong authentication for accepting
>routing announcements as valid with a trusted origin when they might not
>be, for whatever reason, once again some times intentionally.
>Actually is not very hard to exploit this weakness in the current routing
>architecture and protocols to force prefix hijacking, hence traffic
An RPKI would be helpful, though I have some technical concerns about it.
>I-coordination mailing list
>I-coordination at nro.net
More information about the I-coordination