Re: [sig-policy] Requests from routing/packeting concerns

  • To: Izumi Okutani <izumi at nic dot ad dot jp>
  • Subject: Re: [sig-policy] Requests from routing/packeting concerns
  • From: Seiichi Kawamura <kawamucho at mesh dot ad dot jp>
  • Date: Tue, 17 Feb 2009 18:53:23 +0900
  • Cc: sig-policy at apnic dot net
  • Delivered-to: sig-policy at mailman dot apnic dot net
  • In-reply-to: <499A7FEE.7030403 at nic dot ad dot jp>
  • List-archive: <>
  • List-help: <>
  • List-id: APNIC SIG on resource management policy <>
  • List-post: <>
  • List-subscribe: <>, <>
  • List-unsubscribe: <>, <>
  • References: <> <> <> <> <> <> <>
  • User-agent: Thunderbird (Windows/20081209)
    • Hash: SHA1
      hi izumi
      > Another point that was mentioned that an operator wish to be aware of
      > the risks of "contaminated" space (black listed, etc) when obtaining
      > space and seeing past records help sometimes. you ofcourse have to do
      > more checks in addition.
      correct me if i'm wrong. my image of this problem.
      the following takes place inside one single ISP.
      A naughty customer-A will return the address after they have done
      all the bad stuff that needs to be done, and say the address was
      given to some good-behaving customer-B that were not effected by
      black lists (e.g the address was used for VPN puroposes).
      when cusotmer-C receives the address and gets blocked by blacklists,
      the ISP has ways to investigate to check if it was cutomer-B
      or customer-A that was doing the naughty behavior, such as
      checking customer profiles, logs, etc.
      now lets say if A was an ISP, B was one single corporate IT division,
      C was a small data-center. If C didn't know that the prefix was owned
      by A in the past, they may go and lawsuit B.
      > Hi Terry,
      > Nice to have a comment from another operator here.
      > Terry Manderson wrote:
      >> Hi Izumi,
      >> On 16/02/2009, at 10:40 PM, Izumi Okutani wrote:
      >>> These were major requests from routing/packeting concerns.
      >>> 1. To have a system that allows a third party to confirm
      >>>    "authenticity" of address space. (prove you are the right holder)
      >>>    A third party may mean an upsteam ISP, or to get internal approval
      >>>    by non-tech people within an organization to obtain IPv4 resource
      >>>    Resource Certificate may provide an answer to the first needs, but
      >>>    may be more studies are required for proving it to non-tech people.
      >> My reading of this, and do correct me if I'm wrong, is the underlying 
      >> question of:
      >> "What, if any, tools are available that allows my non-technical people 
      >> to verify that a new/existing customer has this 'new' prefix for which 
      >> they are asking me to route?"
      >> yes?
      > Not quite. (but thanks for trying to clarify)
      > the idea is that a routing engineer might need to justify within their
      > organization (manager, account department, etc) that it is an authentic
      > address worth spending the budget when they obtain a resource.
      > so it would help to have a tool/document published to do this. may be a
      > resource cert would be good enough but a concern is that it may be too
      > digital/techy for others to understand.
      > That was a comment from one of the ISPs here. I wonder how general this
      > needs would be as the region?
      >>>  2.Information from APNIC to help confirm the "cleaness" of address
      >>>    Records on the past holders of the address space (not only the
      >>>    previous, but all past holders by date) would help at the time of
      >>>    obtaining the resource/trouble shooting for transfered space.
      >>>    The public log defined in prop-050 is probably quite good overall to
      >>>    but hope we can review more on other information which may be
      >>>    required.
      >> "cleaness" is interesting. I see the value in the immediately previous 
      >> details, however due to the business climate and the way organisations 
      >> are sold/bought/wound-up I suspect that the information used for trouble 
      >> shooting, such as calling a 'long-ago' holder to get their upstream to 
      >> change a filter, may not be all that useful due to ageing of details.
      > I see. i wondered about this after your comment and asked Tomoya Yoshida
      > from OCN.
      > Apparently, sometimes the issue or the problem doesn't just lie in the
      > previous holder, but could go a few times back, e.g. to remove address
      > from black list.
      > Another point that was mentioned that an operator wish to be aware of
      > the risks of "contaminated" space (black listed, etc) when obtaining
      > space and seeing past records help sometimes. you ofcourse have to do
      > more checks in addition.
      >>> They may sound more like operational details rather than policies, but
      >>> operators here feel it's quite important that we have them ready before
      >>> implmenting this policy for the transfer policy to be work in real life.
      >> I think bring forward operational realities is important. Any policy in 
      >> the internet space that forgets the operational truths is at risk of 
      >> being half-baked.
      > right.
      >>> i'll stop here for today...but I'll have to spam some more tomorrow :-P
      >> :-)
      >> I don't consider this spam.
      > okay ~ you've encouraged me. I'll keep going.
      > izumi
      > *              sig-policy:  APNIC SIG on resource management policy           *
      > _______________________________________________
      > sig-policy mailing list
      > sig-policy at lists dot apnic dot net
      -----BEGIN PGP SIGNATURE-----
      Version: GnuPG v1.4.6 (MingW32)
      -----END PGP SIGNATURE-----