Rebecca Lee Bostwick <bostwick@es.net> writes:
From an operational point of view, imho, it is helpful to be in a state where the home AS/network-prefix information is correct and consistent, both in the route annoucements and in the registration. There are cases where playing games w/route preferences based on home AS is required, and if they are not reliable, what good are they?
That is exactly why the original (classic :-) RIPE routing registry insisted on "one route one (home) AS". It is with great trepidation that we abandoned this in ripe-181 to allow multiple proxy aggregation that some felt should be describable.
I'd much rather deal w/AS path decisions than network prefixes. Make sense?
Yes, it does! Dealing with prefixes will not scale. It is impractical in many situations already. I would rather deal with originating (home) AS only rather than opening the box of pan(ndor)aths at this point. It is my experience that there are not many people who can engineer consistent routing based on AS paths in the face of changing topology. ripe-181 gives some simple examples of the problems in the discussion of the "as-exclude" attribute. There is certainly a need for planning and/or simulation tools which support the designer in handling topologies too complex to keep in (one) mind. Anyone working on such things?
By the way, does anyone have a feel for how "good" the home AS registration information is/will be?
Have a look at ftp://ftp.ripe.net/ripe/as Nosing around a little will give you the answer for the RIPE RR. To sum up today's results: We have 12289 routes registered with an originating AS and the following 36 conflicts: 131.154.0.0 in AS137 (database) and anounced by AS2038 (BGP) 134.93.0.0 in AS2857 (database) and anounced by AS517 (BGP) 134.214.0.0 in AS2060 (database) and anounced by AS789 (BGP) 140.78.0.0 in AS1205 (database) and anounced by AS1853 (BGP) 141.192.0.0 in AS790 (database) and anounced by AS1342 (BGP) 144.168.0.0 in AS1290 (database) and anounced by AS2551 (BGP) 152.66.0.0 in AS2012 (database) and anounced by AS2547 (BGP) 157.193.0.0 in AS2111 (database) and anounced by AS2611 (BGP) 159.93.0.0 in AS2875 (database) and anounced by AS2118 (BGP) 192.54.104.0 in AS517 (database) and anounced by AS1324 (BGP) 192.58.67.0 in AS719 (database) and anounced by AS544 (BGP) 192.65.185.0 in AS513 (database) and anounced by AS1133 (BGP) 192.71.39.0 in AS1257 (database) and anounced by AS2116 (BGP) 192.76.243.0 in AS760 (database) and anounced by AS1853 (BGP) 192.83.19.0 in AS1759 (database) and anounced by AS544 (BGP) 192.88.23.0 in AS1205 (database) and anounced by AS1853 (BGP) 192.88.24.0 in AS1205 (database) and anounced by AS1853 (BGP) 192.107.125.0 in AS1205 (database) and anounced by AS1853 (BGP) 192.108.114.0 in AS2546 (database) and anounced by AS1241 (BGP) 192.109.57.0 in AS517 (database) and anounced by AS1270 (BGP) 192.121.154.0 in AS1755 (database) and anounced by AS2603 (BGP) 192.129.55.0 in AS517 (database) and anounced by AS2871 (BGP) 192.152.54.0 in AS760 (database) and anounced by AS1901 (BGP) 193.6.23.0 in AS2012 (database) and anounced by AS2547 (BGP) 193.10.80.0 in AS1653 (database) and anounced by AS1805 (BGP) 193.63.58.0 in AS786 (database) and anounced by AS2699 (BGP) 193.101.134.0 in AS1270 (database) and anounced by AS701 (BGP) 193.119.73.0 in AS1290 (database) and anounced by AS786 (BGP) 193.172.27.0 in AS2043 (database) and anounced by AS786 (BGP) 193.184.88.0 in AS719 (database) and anounced by AS1741 (BGP) 193.185.26.0 in AS719 (database) and anounced by AS1741 (BGP) 193.186.165.0 in AS2131 (database) and anounced by AS1853 (BGP) 193.186.167.0 in AS2131 (database) and anounced by AS1853 (BGP) 193.186.169.0 in AS2131 (database) and anounced by AS1853 (BGP) 193.186.170.0 in AS2131 (database) and anounced by AS1853 (BGP) 193.186.171.0 in AS2131 (database) and anounced by AS1853 (BGP) which represents 0.29 percent of the routes registered. Further there are 141 prefixes registered as having no connectivity but we see routes for them. We see 10 prefixes we know are originated in Europe but not registered. Together all these incosistencies represent 1.52 percent of the data registered. We do not have the staff resources yet to do a better job at cleaning these up. I consider it not too bad though! Daniel