There is no "absolute" point at which to take these measurements. Any point will be coloured (colored) by its relative Internet location with respect to amounts of local detail and aggregated distant detail. Thanks, Geoff Tony said:
Well I did a bit of digging at the sudden difference we are seeing. A lot of this points to different levels of aggregation at different points in the Internet. I took a dump from Bill's box at ?LA? and from the xara.net router I use at Mae-east and saw one immediate example of this. Here's the prefixes announced out of the 206.16/16 (belongs to CERFNET) block.
There is no "absolute" point at which to take these measurements. Any point will be coloured (colored) by its relative Internet location with respect to amounts of local detail and aggregated distant detail.
Yes. Dave Meyer is trying to overcome this with his new route viewer (route-views.uoregon.edu), analogous to Pushpendra's. He is getting multi-hop BGP from Europe (thanks RIPE), Japan (thanks IIJ), and MAE-West (no, LA would probably not be an interesting addition:-) to get widely disparate and hence interesting views of the infrastructure. It would be prettier if he could ip as-path access-list 142 permit ^NAS_ route-map peerN-in permit 1 match as-path 142 sed-path s/^NAS_// ! or maybe set as-path un-prepend NAS ... neighbor 42.666.7.11 remote-as NAS neighbor 42.666.7.11 route-map peerN-in in to clean the first AS off the path, as it looks tacky. But I suspect cisco would fear the impact of such a knob on their support folk. I do not think this approach would be overly useful for Tony's CIDR report. It is not a debugging tool, but an overall trend chart. The constant measurement point and relative measure is what I find useful. randy
There is no "absolute" point at which to take these measurements. Any point will be coloured (colored) by its relative Internet location with respect to amounts of local detail and aggregated distant detail.
Thanks,
Geoff
True enough.. (coloured is the correct spelling :) Dave Meyers is starting to put up something like this off his exchange in oregon. Perhaps it would be useful to get a number of players together and ensure that we are taking routing snaps at the same time and then persuade JQ to revamp his "Internet Weather Maps" to reflect these high/low pressure patterns across the landscape. --bill
In article <199612020252.AA02201@zephyr.isi.edu> you write:
There is no "absolute" point at which to take these measurements. Any point will be coloured (colored) by its relative Internet location with respect to amounts of local detail and aggregated distant detail.
Thanks,
Geoff
True enough.. (coloured is the correct spelling :) Dave Meyers is starting to put up something like this off his exchange in oregon.
Perhaps it would be useful to get a number of players together and ensure that we are taking routing snaps at the same time and then persuade JQ to revamp his "Internet Weather Maps" to reflect these high/low pressure patterns across the landscape.
Something like that is on the TODO list for the MIDS IWR, pending funding. Thanks, John John S. Quarterman <jsq@mids.org> Editor, Matrix Maps Quarterly and Matrix News <mids@mids.org> MIDS Internet Weather Report <URL:http://www.mids.org/weather/> President, Matrix Information and Directory Services (MIDS) http://www.mids.org, +1-512-451-7602, fax: +1-512-452-0127 1106 Clayton Lane, Suite 500W Austin, TX 78723 U.S.A.
participants (4)
-
bmanning@ISI.EDU
-
gih@telstra.net
-
John S. Quarterman
-
randy@psg.com