ARIN removed 154 as-set's from it's IRR a week+ ago

2020-06-24 Thread Martin J. Levy
This is a quick note about as-set's.

First some background. For those that followed the ARIN to ARIN &
ARIN-NONAUTH split a few weeks ago, you'll know that ARIN moved around
56% of all their IRR route/route6 objects into the ARIN-NONAUTH source
database. This was based on objects that are not owned within ARIN
managed space. For more on this; review the ARIN mailing lists and the
video from their meeting a few weeks ago.

But what about all those as-set objects within ARIN IRR? This wasn't
really reviewed heavily and so I thought, I'd dump some info
everyone's-way. Yes there are objects and yes, some moved to
ARIN-NONAUTH (and hence maybe aren't being used for route filtering
anymore).

First off - this data is based on the ARIN IRR database dump of
24/June/2020. You can find the current IRR data at
ftp://ftp.arin.net/pub/rr/ - it's easy to download the dataset. Here's
the raw numbers:

$ gunzip < arin-nonauth.db.gz | egrep -i '^as-set:' | wc -l
154
$ gunzip < arin.db.gz | egrep -i '^as-set:' | wc -l
1413
$

Only around 154 as-set objects were placed into ARIN-NONAUTH. Around
9.8% of the total number of objects.

Running a quick script against the RADB combined IRR database shows
that of those 154 as-set objects that ARIN moved; there're 133 as-set
objects that are only in ARIN-NONAUTH. An object with that name
doesn't exist in any other IRR (like ARIN, ALTDB, RADB, etc.)

Of the remaining 21 as-set objects, the object names (but maybe not
the same content???) exist in various other IRR registries. As
follows:

AS-RAWBW   ||ALTDB|||
AS-NFCR||ALTDB|RADB||
AS-CTA |ARIN| ||RIPE|
AS-27299   || |RADB||
AS-33251-TRANSIT-CUSTOMERS || |RADB||
AS-46844   || |RADB||
AS-7349-TRANSIT-CUSTOMERS  || |RADB||
AS-ELI || |RADB||
AS-MEEBO   || |RADB||
AS-MFT || |RADB||
AS-28140   || ||RIPE|
AS-FUZENET || ||RIPE|
AS-MM  || ||RIPE|
AS-NETLOGIC|| ||RIPE|
AS-PCNET   || ||RIPE|
AS-SVINE   || ||RIPE|
AS-BLUENET || |RADB|RIPE|
AS-LIUXYON ||ALTDB||RIPE|
AS-NITAET  ||ALTDB||RIPE|
AS-NITAETv6||ALTDB||RIPE|
AS-VIATEL  || |RADB|RIPE|

I didn't try to check the content of these - but then again; comparing
as-set's across IRRs is a thankless task - even in the best of days.

So where does this leave us? We have various networks that own
as-set's that are now potentially unavailable to upstreams or transit
providers for creating customer filters. Keep in mind that
ARIN-NONAUTH exists within RADBs IRR datasets - so it's possible that
some upstream transits may use that data. Some may not. YMMV.

Now onto the next step - answering that question.

Of the 154 as-set objects, only 29 names appear in other as-set
objects. (BTW: Thanks to "irrexplorer" for that information!). It does
not mean that the 125 other entries are unused - they may be used; but
not referenced by their transits (doubly so if the transit is a
tier1).

Focusing on the 29 that are referenced somewhere else; you end up with
this listing showing the as-set object name that they appear in. If
they appear in ARIN-NONAUTH that's kinda zero sum game. If they show
as being referenced in other IRRs then there's hope - but not really,
as they are ARIN-NONAUTH objects. Basically - however, you look at it;
lesser filters are being created and maybe routing is affected.

AS-23016:AS-BACKBONE
  ARIN-NONAUTH mentioned-in  AS-23016
AS-23073
  ARIN-NONAUTH mentioned-in  AS-23073
AS-6653
  RADB mentioned-in  AS-COMCAST-IBONE
AS-AHS
  LEVEL3   mentioned-in  AS-BANDCON
  LEVEL3   mentioned-in  AS-ISOMEDIA
  RADB mentioned-in  AS-ISOFUSION
AS-AS36412-CUSTOMERS
  ARIN-NONAUTH mentioned-in  AS-AS36412-ALL
AS-CBSI
  RIPE mentioned-in  AS-TELIANETNA
  RIPE mentioned-in  AS-TELIANETNA-V6
AS-CDSTEPHENS
  RIPE mentioned-in  AS-4IXP
AS-CRM-AR
  RIPE mentioned-in  AS-TDATANETSA
AS-FCH-Customers
  ARIN-NONAUTH mentioned-in  AS-FCH
AS-IMGIX
  NTTCOM   mentioned-in  AS2914:AS-US
AS-IRONP-1
  ALTDBmentioned-in  AS-NLAYER-CUSTOMERS
AS-LOGIN
  ARIN mentioned-in  AS-DECIX-DFW
  ARIN mentioned-in  AS-DECIX-DFW-V6
AS-MCBB-INTERNAL
  ARIN-NONAUTH mentioned-in  AS-MCBB
AS-MCBB-TRANSIT
  ARIN-NONAUTH mentioned-in  AS-MCBB
AS-MEDIA-HOSTS
  ALTDBmentioned-in  AS-NETELLIGENT
AS-MQCUSTOMERS
  ARIN-NONAUTH mentioned-in  AS-MARQUISNET
  ARIN-NONAUTH mentioned-in  AS35937:AS-MARQUISNET
AS-PEACHNET
  RADB   

lol reddit

2020-06-24 Thread Michael Thomas



Apparently they've completely melted down. But them crapping out is 
almost a daily occurrence anyway.


Does anybody have any insight as to why they are unable to keep the 
lights on? I mean from the outside what they do is not particularly 
complicated. At least Usenet had an inter-domain flood fill problem to 
contend with.


Mike



Re: why am i in this handbasket? (was Devil's Advocate - Segment Routing, Why?)

2020-06-24 Thread Masataka Ohta

Owen DeLong wrote:


Saying /16 is ambiguous depends on IP version.


Not really… A /16 in IPv6 is a lot more addresses, but its still
using the first 16 bits to specify the prefix, same as IPv4.


As I wrote:

: But, it should be noted that a single class B routing table entry
: often serves for an organization with 1s of users, which is
: at least our case here at titech.ac.jp.

the number of remaining bits save the first 16 matters, which depends
on IP version.

Masataka Ohta