I noticed something interesting the other day... /24's seem to be contributing to the majority of the internet routing table today. Here is the breakdown right now as viewed through GlobalCrossing transit (the view is consistent from many other sources): Length /8: 23 Length /9: 4 Length /10: 5 Length /11: 10 Length /12: 32 Length /13: 59 Length /14: 194 Length /15: 304 Length /16: 6751 Length /17: 1003 Length /18: 2017 Length /19: 6275 Length /20: 4451 Length /21: 4316 Length /22: 6457 Length /23: 8429 Length /24: 58763 Length /25: 224 Length /26: 351 I was rather shocked by the number of /24's so I ran some further stats. Here is the top 25 offenders by origin AS: ASN: 701 routes: 1906 UUNet ASN: 1221 routes: 1775 ?????? ASN: 1 routes: 606 BBN Planet ASN: 2711 routes: 603 Rock Hill Telephone Company ASN: 7018 routes: 599 AT&T ASN: 7046 routes: 587 UUNet ASN: 1785 routes: 546 Applied Theory ASN: 705 routes: 537 UUNet ASN: 174 routes: 508 PSInet ASN: 8013 routes: 494 PSINet Canada ASN: 2551 routes: 463 ICG NetAhead ASN: 1239 routes: 453 Sprint ASN: 3561 routes: 365 Cable and Wireless ASN: 4293 routes: 355 Cable and Wireless ASN: 2548 routes: 346 Digex ASN: 2764 routes: 337 connect.com.au ASN: 3602 routes: 313 Sprint Canada ASN: 703 routes: 299 UUNet (Asia) ASN: 209 routes: 292 Qwest ASN: 2914 routes: 291 Verio ASN: 3301 routes: 271 RIPE ASN: 577 routes: 268 Bell Canada ASN: 3908 routes: 261 Supernet, Inc. ASN: 271 routes: 261 University of British Columbia ASN: 724 routes: 258 DLA Systems Automation Center Most of this waste seems to be coming from the Tier 2's, and for some reason a large number of people doing things like 100/24, 102/24, 104/24, etc, out of their larger netblocks. So here is a list considering the entire AS path (note I can't give an accurate number for gblx), which indicates those transiting the biggest offenders: ASN: 701 routes: 17867 UUNet ASN: 3967 routes: 8170 Exodus Communications ASN: 10910 routes: 5948 InterNap ASN: 1239 routes: 4967 Sprint ASN: 3561 routes: 3095 Cable and Wireless ASN: 209 routes: 3061 Qwest ASN: 6461 routes: 2383 AboveNet Communications ASN: 174 routes: 2347 PSInet ASN: 1221 routes: 2343 ?????? ASN: 7018 routes: 2260 AT&T ASN: 6453 routes: 2014 Teleglobe ASN: 4355 routes: 1894 EarthLink ASN: 702 routes: 1617 UUNet (Europe) ASN: 5511 routes: 1480 Europe ASN: 2548 routes: 1403 Digex ASN: 1 routes: 1101 BBN Planet ASN: 703 routes: 1095 UUNet (Asia) ASN: 577 routes: 862 Bell Canada ASN: 2711 routes: 814 TRW Huntsville ASN: 2914 routes: 785 Verio ASN: 3356 routes: 755 Level 3 Communications ASN: 6347 routes: 717 SAVVIS Communications ASN: 1785 routes: 715 Applied Theory ASN: 7474 routes: 661 APNIC ASN: 690 routes: 653 Merit Network <-- bad merit, bad ASN: 7046 routes: 588 UUNet ASN: 4000 routes: 564 Sprint International ASN: 8297 routes: 560 Europe ASN: 4648 routes: 546 APNIC ASN: 705 routes: 537 UUNet ASN: 2828 routes: 506 Concentric ASN: 8013 routes: 497 PSINet Canada ASN: 3908 routes: 483 Supernet, Inc. ASN: 724 routes: 476 DLA Systems Automation Center ASN: 2551 routes: 472 ICG NetAhead ASN: 286 routes: 457 EUNet ASN: 4565 routes: 456 Epoch Networks ASN: 7473 routes: 438 APNIC ASN: 5650 routes: 438 Electric Lightwave ASN: 2764 routes: 430 connect.com.au ASN: 1913 routes: 409 Defense Logistics Agency ASN: 3602 routes: 408 Sprint Canada ASN: 2685 routes: 403 IBM ASN: 816 routes: 401 UUNet Canada BTW who is AS 1221? That block is listed as allocated by ARIN, but ARIN has nothing in the whois for them. I can produce a full list if anyone is interested. I'll see if I can get a more accurate view for the aspath stats as well. -- Richard A Steenbergen <ras@e-gerbil.net> http://www.e-gerbil.net/ras PGP Key ID: 0x138EA177 (67 29 D7 BC E8 18 3E DA B2 46 B3 D8 14 36 FE B6)
On Tue, 6 Feb 2001, Richard A. Steenbergen wrote:
BTW who is AS 1221? That block is listed as allocated by ARIN, but ARIN has nothing in the whois for them.
[whois.radb.net] aut-num: AS1221 as-name: TELSTRA-AS descr: TELSTRA-AS admin-c: GIH105 tech-c: DW187 notify: as1221@telstra.net mnt-by: MAINT-AS1221 changed: David.Woodgate@telstra.net 19990506 source: RADB aut-num: AS1221 as-name: ASN-TELSTRA descr: Telstra Pty Ltd Locked Bag No. 5744 GPO, Canberra, ACT, 2601 admin-c: GH105-AP tech-c: DW187-AP remarks: AS assigned by the former InterNIC remarks: Object automatically converted from the APNIC(RIPE181) registry Seems some special odd transfer between ARIN and APNIC. HTH. HAND. Pi
Length /24: 58763
ASN: 1221 routes: 1775 ??????
Now, this is interesting; AS 1221 is not in ARIN or anywhere. helium:~# whois -h whois.arin.net 1221 No match for "1221". Iny my view of the world, 1221 is wholly transited by Exodus: *>i210.9.34.0 209.67.40.13 100 0 3967 1221 4736 i They have lots of AS's behind them, including 9443, 4730, 4736, etc. ARIN: What about people who use AS's that aren't assigned to them? Community: Should one filter these announcements? Exodus: Don't you care?
Date: Tue, 6 Feb 2001 16:10:28 -0500 (Eastern Standard Time) From: Alex Rubenstein <alex@nac.net> Sender: owner-nanog@merit.edu
ARIN: What about people who use AS's that aren't assigned to them?
ARIN is only responsible for the Americas and, maybe, parts of Africa. Lots of ASes are not in ARIN. Try RIPE and APNIC for more complete information. Or try the IRR. In this case, both APNIC and the IRR show this AS. aut-num: AS1221 as-name: ASN-TELSTRA descr: Telstra Pty Ltd Locked Bag No. 5744 GPO, Canberra, ACT, 2601 admin-c: GH105-AP tech-c: DW187-AP remarks: AS assigned by the former InterNIC remarks: Object automatically converted from the APNIC(RIPE181) registry mnt-by: MAINT-AS1221 changed: hostmaster@apnic.net 20000131 changed: auto-rpsl@rpsl.apnic.net 20001213 source: APNIC
Community: Should one filter these announcements?
Why?
Exodus: Don't you care?
Why should they? R. Kevin Oberman, Network Engineer Energy Sciences Network (ESnet) Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab) E-mail: oberman@es.net Phone: +1 510 486-8634
On Tue, 06 Feb 2001 13:24:45 PST, Kevin Oberman said:
Community: Should one filter these announcements? Why?
Umm... because if you filter the /24s you will remove 58K of 95K, or some 61%. I'm *sure* that having a routing table 40% of the original size will help the next time you have a BGP flap. Your mileage may vary, depending on the actual amount of improvement obtainable by proper aggregation and/or filtering.
Exodus: Don't you care? Why should they?
Umm.. because urinating in the swimming pool tends to irritate the other people at the pool party? Or because if *everybody* acts that way, you get a Really Dirty swimming pool? -- Valdis Kletnieks Operating Systems Analyst Virginia Tech
On Tue, 6 Feb 2001 Valdis.Kletnieks@vt.edu wrote:
Exodus: Don't you care? Why should they?
Umm.. because urinating in the swimming pool tends to irritate the other people at the pool party? Or because if *everybody* acts that way, you get a Really Dirty swimming pool?
See Also: Tragedy of the commons. Almost time to bring back the satanic filters. /vijay
I was, of course, being pedantic. I was a bit irritated at the parochialism. If an AS is not registered in ARIN, too often people assume that it's not registered anywhere. R. Kevin Oberman, Network Engineer Energy Sciences Network (ESnet) Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab) E-mail: oberman@es.net Phone: +1 510 486-8634
On Tue, 6 Feb 2001, Alex Rubenstein wrote:
Iny my view of the world, 1221 is wholly transited by Exodus:
*>i210.9.34.0 209.67.40.13 100 0 3967 1221 4736 i
They have lots of AS's behind them, including 9443, 4730, 4736, etc.
ARIN: What about people who use AS's that aren't assigned to them?
Community: Should one filter these announcements?
Exodus: Don't you care?
Exodus does not sell bandwidth to Telstra. That is a 'peer' route. /24s are seen from peers in the Exodus network. (check out the route-server for more information.) thanks Christian
On Tue, 6 Feb 2001, Christian Nielsen wrote:
On Tue, 6 Feb 2001, Richard A. Steenbergen wrote:
ASN: 3967 routes: 8170 Exodus Communications
The total number of routes for 3967 is under 1400 routes.
1316 at my last count.
To be fair to Exodus, these numbers for transit projections came from Gblx+Nac.net transit, and NAC in turn gets transit from Exodus and others, so the "worldview" is quite biased. Thats the problem with transit views, in order to be unbiased you need a transit-free extensive-peering view. I already asked 6461 (AboveNet) for such a view for research purposes and my request was denied, perhaps another tier-1 would like to help? -- Richard A Steenbergen <ras@e-gerbil.net> http://www.e-gerbil.net/ras PGP Key ID: 0x138EA177 (67 29 D7 BC E8 18 3E DA B2 46 B3 D8 14 36 FE B6)
participants (8)
-
Alex Rubenstein -
Christian Nielsen -
Kevin Oberman -
Pim van Riezen -
Randy Bush -
Richard A. Steenbergen -
Valdis.Kletnieks＠vt.edu -
Vijay Gill