[savnet] Re: Ketan Talaulikar's Discuss on draft-ietf-savnet-inter-domain-problem-statement-21: (with DISCUSS and COMMENT)

Ketan Talaulikar <ketant.ietf@gmail.com> Mon, 26 October 2026 06:03 UTC

Received: from mail-pl1-x635.google.com (mail-pl1-x635.google.com [IPv6:2607:f8b0:4864:20::635]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange x25519 server-signature ECDSA (prime256v1) server-digest SHA256) (No client certificate requested) by mx.ietf.org (Postfix) with ESMTPS id DD0DD47 for <savnet@ietf.org>; Mon, 26 Oct 2026 06:03:48 +0000 (UTC)
Authentication-Results: mx.ietf.org; dkim=pass header.d=gmail.com header.s=20251104 header.b=LPyuUxss; arc=pass ("google.com:s=arc-20260327:i=1"); spf=pass (mx.ietf.org: domain of ketant.ietf@gmail.com designates 2607:f8b0:4864:20::635 as permitted sender) smtp.mailfrom=ketant.ietf@gmail.com; dmarc=pass (policy=none) header.from=gmail.com
Received: by mail-pl1-x635.google.com with SMTP id d9443c01a7336-2ea1e7fad99so12898435ad.1 for <savnet@ietf.org>; Sun, 25 Oct 2026 23:03:48 -0700 (PDT)
ARC-Seal: i=1; a=rsa-sha256; t=1792994622; cv=none; d=google.com; s=arc-20260327; b=n1uylSwbp9/T9wbUcsE+L4HNPDF3X0iYItZMARXsCqxir7qamH5HI+PvLE6hcr/cUI KUNGJy8yy066uP/E2ei+DbQ7zFEk80OYTPl/kax+Rs5y4ywAwHjVYWwSI3FzsMe+JnI2 p4cWbQpDfeYP/L+n8IzxkVGDH46O8Fv9PPaiyYcOoGFrw8+BsSZnvOoWL8zcvRXXqp8G +7r1sNCbBJOYbYe6P7mO6OkhML2ZQ+SwXuPLwXl83hOaSH8xiZl44F/RbNtPqI8u6qOx ZHMuHsf19Rki7N0qapROkfIh/rpyPwtW3l7Ge2ax+y1/124RPeN6rOKbQrysf0RV3GrI /dQw==
ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20260327; h=cc:to:subject:message-id:date:from:in-reply-to:references :mime-version:dkim-signature; bh=NIKLKsx68gAeLpelzXZWPxGCDbYJd9PBLKAHYJjpm8k=; fh=dvl/4xNUya+Lr9flZmwMNWaGS+TfGsqL//pRXVQfOz8=; b=COytJ4TKGBr38IiSwR5SIhdT2gpG2YFqSBGdo5u8Ej4mZ0tiATrIHv59fINwmD/jyK BkRzRTCoWqV8ch9M6KKtwvw1CGM4vSpr4teAN+DC/nsnMIP0JkfvZeV5YptJ6FkeQhsD rMLdw4J/m+OSJEzkWj6PoimKFgDq980WExVw4LurSfUnvUqeXck1IytzVqGvMaCEGFDU BiWYWZEtxevkwwFyG/kCp/yLZxlaClxyjGOfWhCLf4+9Qd4siAtGZl42isqqDg+vCW6F 6CQHcCGwVIOuxFZhqH31IY/2nyOQrIQz+qgSS8AqkjcVXVe7ene5A6N+US44YR/WKOSh eilA==; darn=ietf.org
ARC-Authentication-Results: i=1; mx.google.com; arc=none
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20251104; t=1792994622; x=1793599422; darn=ietf.org; h=content-type:cc:to:subject:message-id:date:from:in-reply-to :references:mime-version:from:to:cc:subject:date:message-id:reply-to :content-type; bh=NIKLKsx68gAeLpelzXZWPxGCDbYJd9PBLKAHYJjpm8k=; b=LPyuUxssBMU6i80SgYT5X/zmT/tHboHgNAyzYkfuzZGhlZ7+unkYCvy09NtAxdb2Qg wFVe8qpLl5O8tIUeFpwHOg1OP6Rz/s/0NfMjy7bLPkaflyrCpYnaiASCs71ZSV6XBEGw uDn1jWEESoJ3lV1QE9uJ+zPcM3Kk4+2EtlU1emERQGlqyIpxJlVq/8XzeQ3G+H7NASzm bPbYlxvG0PcANquhmrEVgEOghSwQdgkLODdSEZd0AiDUD+JlaMF4nFjuXHMJPIcpRTF9 L9smzWmSirslb6GXeSIf/eNWNvi7+tXuO+QYLiceb3aCvLKo0MA9cO1BC3aoHwYBUgPL gmbg==
X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20260707; t=1792994622; x=1793599422; h=content-type:cc:to:subject:message-id:date:from:in-reply-to :references:mime-version:x-gm-gg:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to:content-type; bh=NIKLKsx68gAeLpelzXZWPxGCDbYJd9PBLKAHYJjpm8k=; b=CrXdHP0qv6zuEARh2B/1IJa1KD+FlgpNPlw5CzChsulCABKL1ZikxhC3caUhXwfycn hw+SWe0PTnDhvr5uOhHXQGjx2GtTdqp2/71+s18nqpnFCR12zxckPjCx6Nz1lZoolaGa QsM7Cat8cmRo0oAmZ3x/wKNJLDN4bf/lMwsuUUMGfbmCZ338D6dLL/16fKMR7xRzCmpQ 0m6DcLavmjyxyJfigNSrEY9ZOJnvvbbNtlwrCfKZLEMuCQGx4z0NC68deTRvFDZG2mVg +tQWzYaEwPC9Ey4rQTJ+VTgBz/G4vAY2kZMoEtr+UV/8XeBf1GeGX7ynGn9nyrsmxJmF jgQQ==
X-Forwarded-Encrypted: i=1; AKwUvBxwWQlbyJBgJNio3rjnzTX5lB2WhJHYlKfRSk2tDvzAP7erjs1Tf/kGFaPEPmFFbbLoGX4BJdY=@ietf.org
X-Gm-Message-State: AFq9FYKzhH2ljYUJvIXWOLzDPDbHEFg+o9iG0dOCUcrMAzH6233HDjsU 4QiCVVcaamw92B3vqgipMmBhUDsL8Zvl63OBLFKxhAmQmo19s+N2vjw9PLEGvYl2FvTf3NyrpI9 PL484lZrKkdyKqiMWyts0uTXsTKD4Zm4=
X-Gm-Gg: AYBFou3aFdKS131dGTl8+YaBY2wC+4W2Tl4ev+LhII2J+u1dJLmPHB7LAnXGpwJhUYx v6++Km7zB/Sjqpgu32aPFZinauRN/DfHyBzObGzt9IIxL6gfgLxhhwCodBVjDpdlbqVztpGj2+l CDvbvddqhiaUG7x7sEmdAsCTtuC1lswiRSaIMuJU0qrx5vZKuK/WNmr7KQv+Ipb1eeR0ZkGscir dV31P6/ey5TAltW87kRMaKorhIouPEuockh26C+fFG3n+ctrBOYH+kMWSHfIPa8izGBzEnAw0bt bpdQl+OAZC/jm4HHD8nxv09bxuzKzvpbNEiKwbsOkfFdS7dwpUwfs94NIVhK9Yaufa1XhoL8x1F WvX29nexCYkLyBVXYHeBB2w/qoNCnMSGvTGT3Tpg0nKva8FEHzcrn5oILl6XwdP8BLNXVSyBTfX z5Z3d0bWR4K3YtNh/OVQoBUlDkCnjyUVx++kc/R43SuS3u4pnhK1vIfY2uzFmhLgz6Gi8o6+hGy OU=
X-Received: by 2002:a17:902:d50a:b0:2e9:6fe3:613 with SMTP id d9443c01a7336-2ea9c5e3542mr47333655ad.20.1792994621510; Sun, 25 Oct 2026 23:03:41 -0700 (PDT)
MIME-Version: 1.0
References: <179260392240.567770.14882588021395898971@dt-datatracker-7457c9c988-pqlqx> <768c2fc1.9e30.1a1695ddb17.Coremail.liulb@zgclab.edu.cn>
In-Reply-To: <768c2fc1.9e30.1a1695ddb17.Coremail.liulb@zgclab.edu.cn>
From: Ketan Talaulikar <ketant.ietf@gmail.com>
Date: Mon, 26 Oct 2026 11:33:29 +0530
X-Gm-Features: AclHuK9XxP-zP8mAxfQL5UOmTLFV8EPfuea8_0RI43yp3sLmfl02J1gikQ4GKFY
Message-ID: <CAH6gdPzfkK1q2WTpztOyfEn6qTvQh4k2gc3rmTc_P9t3iGCF_A@mail.gmail.com>
To: Libin Liu <liulb@zgclab.edu.cn>
Content-Type: multipart/alternative; boundary="0000000000007fd993065eb81759"
X-Spamd-Bar: --------------------
Message-ID-Hash: VUV32SBJL4KWE2IJIG4BNBHPJCQHF64A
X-Message-ID-Hash: VUV32SBJL4KWE2IJIG4BNBHPJCQHF64A
X-MailFrom: ketant.ietf@gmail.com
X-Mailman-Rule-Misses: dmarc-mitigation; no-senders; approved; loop; banned-address; emergency; member-moderation; nonmember-moderation; administrivia; implicit-dest; max-recipients; max-size; news-moderation; no-subject; digests; suspicious-header
CC: The IESG <iesg@ietf.org>, draft-ietf-savnet-inter-domain-problem-statement@ietf.org, gengnan@huawei.com, savnet-chairs@ietf.org, savnet@ietf.org
X-Mailman-Version: 3.3.10
Precedence: list
Subject: [savnet] Re: Ketan Talaulikar's Discuss on draft-ietf-savnet-inter-domain-problem-statement-21: (with DISCUSS and COMMENT)
List-Id: Source Address Validation in Intra-domain and Inter-domain Networks <savnet.ietf.org>
Archived-At: <https://mailarchive.ietf.org/arch/msg/savnet/BEeV8HTnSeqnfgwHYyEKLXNirnw>
List-Archive: <https://mailarchive.ietf.org/arch/browse/savnet>
List-Help: <mailto:savnet-request@ietf.org?subject=help>
List-Owner: <mailto:savnet-owner@ietf.org>
List-Post: <mailto:savnet@ietf.org>
List-Subscribe: <mailto:savnet-join@ietf.org>
List-Unsubscribe: <mailto:savnet-leave@ietf.org>

Thanks for the update, Libin. Please let me know when you have something
for me to review - either an update posted to datatracker (preferred) or a
PR.

Thanks,
Ketan


On Fri, Oct 23, 2026 at 8:55 PM Libin Liu <liulb=
40zgclab.edu.cn@dmarc.ietf.org> wrote:

> Hi Ketan,
>
> Thank you for your detailed review and valuable comments.
>
> We have prepared an updated version of the draft to address your comments
> and those from other IESG reviewers. The changes are currently under
> discussion and review among the authors.
>
> We will provide detailed responses to your comments soon. Thanks again for
> your time and helpful feedback!
>
>
> Best,
> Libin and co-authors
>
>
> > -----Original Messages-----
> > From: "Ketan Talaulikar via Datatracker" <noreply@ietf.org>
> > Send time:Thursday, 22/10/2026 01:32:02
> > To: "The IESG" <iesg@ietf.org>
> > Cc: draft-ietf-savnet-inter-domain-problem-statement@ietf.org,
> gengnan@huawei.com, savnet-chairs@ietf.org, savnet@ietf.org
> > Subject: Ketan Talaulikar's Discuss on
> draft-ietf-savnet-inter-domain-problem-statement-21: (with DISCUSS and
> COMMENT)
> >
> > Ketan Talaulikar has entered the following ballot position for
> > draft-ietf-savnet-inter-domain-problem-statement-21: Discuss
> >
> > When responding, please keep the subject line intact and reply to all
> > email addresses included in the To and CC lines. (Feel free to cut this
> > introductory paragraph, however.)
> >
> >
> > Please refer to
> https://www.ietf.org/about/groups/iesg/statements/handling-ballot-positions/
> > for more information about how to handle DISCUSS and COMMENT positions.
> >
> >
> > The document, along with other ballot positions, can be found here:
> >
> https://datatracker.ietf.org/doc/draft-ietf-savnet-inter-domain-problem-statement/
> >
> >
> >
> > ----------------------------------------------------------------------
> > DISCUSS:
> > ----------------------------------------------------------------------
> >
> > Thanks to the authors and the WG for their work on this document.
> >
> > The review is inline in the idnits output of v21 of this document.
> Please look for
> > the tag <EoRv21> to ensure you have the full review.
> >
> > Thanks for aligning the terminology and the document contents with the
> intra-domain
> > document.
> >
> > I have one concern that I would like to discuss. It has also come up in
> > other ballots, so I am perhaps not the only one stumbling on this but
> please
> > correct me if I am wrong.
> >
> > <discuss-1> EFP-uRPF Alg-B and the LPP gap (Section 4.1.1)
> >
> > 483      block when performing SAV with existing mechanisms.  These
> mechanisms
> > 484      include EFP-uRPF Alg-A, which is the focus of the following
> analysis,
> > 485      while it also applies to Strict uRPF and FP-uRPF.  All these
> > 486      mechanisms suffer from the same problem of improper block in
> this
> > 487      scenario.
> >
> > Section 4.1.1 analyses EFP-uRPF Alg-A, but RFC 8704 recommends Alg-B on
> > customer interfaces, and its Section 3.3 is essentially Figure 2: Alg-A
> fails
> > it and Alg-B passes it. In Figure 2, Alg-B at AS 4 accepts AS 1's
> prefixes on
> > the AS 2 interface, because AS 1's routes arrive on another customer
> interface
> > of AS 4. For Alg-B the LPP gap is narrower, and the tolerance is paid
> for in
> > directionality (Section 4.1.3). Suggest:
> >
> > CURRENT
> >    These mechanisms
> >    include EFP-uRPF Alg-A, which is the focus of the following analysis,
> >    while it also applies to Strict uRPF and FP-uRPF.  All these
> >    mechanisms suffer from the same problem of improper block in this
> >    scenario.
> > SUGGEST
> >    These mechanisms
> >    include Strict uRPF, FP-uRPF, and EFP-uRPF Alg-A; the following
> >    analysis uses EFP-uRPF Alg-A.  EFP-uRPF Alg-B, which [RFC8704]
> >    recommends for customer interfaces, avoids the improper block in
> >    Figure 2 because AS 1, the origin AS of those prefixes, is seen on
> >    another customer interface of AS 4.  Alg-B still blocks when the
> origin AS
> >    of the prefix is not seen on any customer interface of the
> >    validating AS, and its tolerance comes at the cost of
> >    directionality (Section 4.1.3).
> >
> > and in the improper-block bullets of Sections 1 and 4.4:
> >
> > s/mechanisms suffer from/mechanisms (Alg-B less so) suffer from/
> >
> >
> > ----------------------------------------------------------------------
> > COMMENT:
> > ----------------------------------------------------------------------
> >
> > I also have some comments and suggestions to offer:
> >
> > 116      [RFC9234].  Further, [RFC9234] mentions RS and RS-client.  An
> RS-to-
> > 117      RS-client interface is akin to the customer interface.  For the
> > 118      purposes of SAV, an RS-client-to-RS interface may be treated
> (1) like
> > 119      a provider interface for simplicity, or (2) like a union of
> lateral
> > 120      peers considering all the ASes the RS-client chose to peer with
> at
> > 121      the IXP RS.
> >
> > <minor> A route server does not forward traffic (RFC 7947, Section 1),
> so an
> > RS-to-RS-client interface carries no data to validate; on the RS-client
> side
> > the traffic comes from the member ASes, often over one IXP interface
> shared
> > with bilateral peers. Perhaps, with RFC 7947 as an informative reference:
> >
> > CURRENT
> >    Further, [RFC9234] mentions RS and RS-client.  An RS-to-
> >    RS-client interface is akin to the customer interface.  For the
> >    purposes of SAV, an RS-client-to-RS interface may be treated (1) like
> >    a provider interface for simplicity, or (2) like a union of lateral
> >    peers considering all the ASes the RS-client chose to peer with at
> >    the IXP RS.
> > SUGGEST
> >    Further, [RFC9234] mentions RS and RS-client.  A route server does
> >    not forward data traffic [RFC7947]; the traffic an RS-client
> >    receives at an IXP comes from the member ASes, often over one
> >    interface shared by route-server and bilateral peers.  For the
> >    purposes of SAV, that interface may be treated (1) like a provider
> >    interface for simplicity, or (2) like a union of lateral peers
> >    considering all the ASes the RS-client exchanges routes with at the
> >    IXP, which requires distinguishing the member AS that sent each
> >    packet.
> >
> >
> > 137      The following summarizes the fundamental problems with existing
> SAV
> > 138      mechanisms, as analyzed in Section 4.
> >
> > <major-editorial> The requirements and the scope are spread across the
> > document. Sections 6 and 7 are requirements outside Section 5. Section 8
> > bounds the requirements but comes after them, and repeats the
> > packet-modification exclusion of Section 1. Section 1 repeats the gap
> list of
> > Section 4.4 and the requirements of Section 5, the latter with drift
> (Section
> > 5.1). Perhaps make Sections 6 and 7 subsections of Section 5, move
> Section 8
> > ahead of the gap analysis, and replace the two bulleted lists in Section
> 1
> > with a pointer. I know this is a lot of editorial changes and too late in
> > the document cycle, but I still wanted to offer this for the authors'
> > consideration.
> >
> > 168         It must achieve zero improper blocking (i.e., avoid false
> > 169         positives) in certain scenarios of interest (Section 4).
> Further,
> >
> > <major> This bullet says "must achieve", while the requirement itself in
> > Section 5.1 says "must seek to achieve", and "certain scenarios of
> interest"
> > is as open here as there; see my comments on line 882 for suggestions.
> >
> > 256      Configuration Information:
> > 257         This information is configured locally by the AS operator.
> For
> > 258         example, an AS provisions (suballocates) prefixes p1 and p2
> for a
> > 259         non-BGP customer network, which also owns an RIR-allocated
> prefix
> > 260         p3.  The customer instructs the AS to advertise p1 via eBGP
> on the
> > 261         public Internet, restrict p2 strictly to internal use (in the
> > 262         customer network), and refrain from advertising p3 while
> still
> > 263         allowing it to source outbound traffic.  The AS locally
> configures
> > 264         these prefixes accordingly.  This configuration information
> is
> > 265         valuable for both intra-domain SAV to permit expected
> prefixes and
> > 266         inter-domain SAV at other interfaces to block unexpected
> prefixes.
> >
> > <minor> Local configuration appears in both the "SAV-related
> Information" and
> > "SAV-specific Information" definitions, and Section 5.4 sets different
> > protection requirements for each, so it is unclear which applies to it.
> > Perhaps:
> >
> > NEW (end of the "Configuration Information" definition)
> >    Configuration that populates the RIB or FIB (e.g., static routes)
> >    is SAV-related information; configuration made for SAV (e.g., p3
> >    above) is SAV-specific information.
> >
> > 338         -  Loose uRPF [RFC3704]: Loose uRPF verifies that the source
> > 339            address of a packet is routable on the Internet by
> matching it
> > 340            with one or more prefixes in the FIB, regardless of the
> > 341            interface on which the packet arrives.  If the source
> address
> > 342            is not routable, Loose uRPF discards the packet.  Loose
> uRPF is
> > 343            typically deployed at the provider interfaces of an AS to
> block
> > 344            packets with source addresses from prefixes that are not
> routed
> > 345            on the global Internet (e.g., IANA-allocated private-use
> > 346            addresses, unallocated IPv4/IPv6 addresses, multicast
> > 347            addresses, etc.).
> >
> > <minor> RFC 3704 Section 2.4 notes that, unless the implementation checks
> > where the default route points, a default route nullifies loose uRPF;
> Section
> > 2.5 describes loose uRPF ignoring default routes. Many ASes that buy
> transit
> > carry a default, so perhaps:
> >
> > NEW (end of the Loose uRPF bullet in Section 3)
> >       If the FIB contains a default route, Loose uRPF permits all
> >       source addresses unless the implementation ignores the default
> >       route ([RFC3704], Sections 2.4 and 2.5).
> >
> >
> > 863      The limitations of existing uRPF-based mechanisms are due to
> their
> > 864      exclusive reliance on BGP data.  Although the algorithms
> themselves
> > 865      have evolved (e.g., [RFC8704]), the underlying input has
> remained
> > 866      unchanged, inherently constraining their accuracy in scenarios
> such
> > 867      as LPP and HP.  With the availability of authoritative
> SAV-related
> >
> > <minor> RFC 8704 Section 3.5 already augments the RPF lists with ROA and
> IRR
> > data, and Section 3.7 recommends it, so "exclusive reliance on BGP data"
> and
> > "the underlying input has remained unchanged" do not describe the
> published
> > method. The conclusion holds, since the augmentation only adds prefixes
> of
> > origin ASes already seen in BGP. Am I correct? Perhaps:
> >
> > CURRENT
> >    The limitations of existing uRPF-based mechanisms are due to their
> >    exclusive reliance on BGP data.  Although the algorithms themselves
> >    have evolved (e.g., [RFC8704]), the underlying input has remained
> >    unchanged, inherently constraining their accuracy in scenarios such
> >    as LPP and HP.
> > SUGGEST
> >    The limitations of existing uRPF-based mechanisms are due to their
> >    reliance on BGP data.  [RFC8704] (Section 3.5) augments it with ROA
> >    and IRR data, but only for origin ASes already seen in BGP, which
> >    leaves their accuracy constrained in scenarios such as LPP and HP.
> >
> > 868      information, plus the potential for SAV-specific information
> > 869      (Section 2), it would be possible to develop comprehensive new
> SAV
> > 870      algorithms or mechanisms to overcome the existing gaps.
> >
> > <minor> SIDROPS is working on draft-ietf-sidrops-bar-sav, which uses BGP
> > UPDATEs, ASPA and ROA objects to address the improper block identified
> here
> > and proposes to update RFC 8704. A pointer would help readers coming
> from the
> > RPKI side. Perhaps:
> >
> > NEW (after the last paragraph of Section 4.4)
> >    [I-D.ietf-sidrops-bar-sav] is an example of such work: it uses
> >    ASPA and ROA objects in addition to BGP UPDATE messages.
> >
> >
> > 882      mechanisms.  It must seek to achieve zero improper blocking
> (i.e.,
> > 883      avoid false positives) in certain scenarios of interest
> (Section 4).
> >
> > <major> Section 1 says a new mechanism "must achieve zero improper
> blocking";
> > Section 5.1, the requirement itself, says it "must seek to achieve" it.
> Please
> > align the two, and perhaps name the scenarios: Section 4 also covers
> prefixes
> > never advertised in BGP, where zero blocking depends on SAV-specific
> > information that Section 5.3 says cannot be assumed. If the stronger
> form is
> > intended:
> >
> > CURRENT
> >    It must seek to achieve zero improper blocking (i.e.,
> >    avoid false positives) in certain scenarios of interest (Section 4).
> > SUGGEST
> >    It must achieve zero improper blocking (i.e., avoid false
> >    positives) in the Limited Propagation of a Prefix and Hidden Prefix
> >    scenarios (Section 4.1) where the information it relies on is
> >    available.
> >
> > with the same wording in the corresponding bullet of Section 1.
> >
> > 899      Any new inter-domain SAV mechanism must not assume pervasive
> adoption
> > 900      of the SAV method, the SAV-related information, or the
> SAV-specific
> > 901      information.  It should benefit early adopters by providing
> effective
> > 902      protection from spoofing of source addresses even in partial
> > 903      deployment.
> >
> > <major> Section 5.3 asks that early adopters benefit, but nothing asks
> that a
> > new mechanism do no harm where its information is missing: for an AS
> that has
> > not adopted it, or a prefix with no SAV-specific information published. A
> > mechanism that blocked more legitimate traffic than RFC 8704 in that case
> > would struggle in partial deployment, and the requirement would also give
> > "over existing mechanisms" (Section 5.1) a baseline. Perhaps:
> >
> > NEW (end of Section 5.3)
> >    Where the SAV-related or SAV-specific information a new mechanism
> >    relies on is absent for a prefix or an AS, the mechanism should not
> >    improperly block more traffic than the existing mechanism it
> >    replaces (e.g., EFP-uRPF Alg-B [RFC8704] on customer interfaces).
> >
> > I will leave the wording, and its strength, to the authors and the WG.
> >
> > 911      diligent about this).  If any new inter-domain SAV mechanism
> > 912      introduces or uses SAV-specific information, security
> mechanisms must
> > 913      exist to prevent malicious injection or alteration of the SAV-
> > 914      specific information.
> >
> > <major> This covers injection and alteration of SAV-specific
> information, but
> > not stale or replayed information, nor limiting its distribution to the
> ASes
> > that need it. I would like to propose text that includes a sentence for
> the end
> > of this section; see my comments on line 1005.
> >
> > 952      false negatives (improper permits).  Consequently,
> troubleshooting
> > 953      these anomalies will rely on operational feedback loops, where
> > 954      network operators receive, investigate, and resolve routing and
> > 955      reachability complaints from affected downstream or peering
> parties.
> >
> > <minor> An ASBR cannot judge its own drops, but operators find improper
> blocks
> > with per-interface drop counters, exported samples of dropped packets,
> and a
> > log-only mode before enforcement, not only through complaints. Perhaps
> ask new
> > mechanisms for those:
> >
> > CURRENT
> >    Consequently, troubleshooting
> >    these anomalies will rely on operational feedback loops, where
> >    network operators receive, investigate, and resolve routing and
> >    reachability complaints from affected downstream or peering parties.
> > SUGGEST
> >    Consequently, a new mechanism should provide counters of packets
> >    dropped by SAV (per interface and, where feasible, per prefix), a
> >    means to export samples of dropped packets, and a mode that logs
> >    without dropping, so that operators can detect improper blocks
> >    before, and in addition to, complaints from affected parties.
> >
> > 983      *  Native IP forwarding: This includes both the global routing
> table
> > 984         based forwarding and Customer Edge (CE) site forwarding of
> VPN
> > 985         traffic.
> >
> > <question> Section 8 puts "Customer Edge (CE) site forwarding of VPN
> traffic"
> > in scope. Does this mean SAV at a PE's VRF interface facing an eBGP CE?
> If so,
> > the SAV-related information there is the VRF's routes, and the RPKI
> objects of
> > Section 2 and the SLURM approach of Section 7 do not describe VPN address
> > space, which is often private and overlapping. It may be worth saying
> which
> > requirements apply in that context or drop the entire VPN thing?
> >
> > 1005     The SAV table will be generated based on SAV-related
> information and/
> > 1006     or SAV-specific information.  If such information is poisoned by
> > 1007     attackers, the resulting SAV table will be inaccurate.
> Consequently,
> > 1008     legitimate traffic may be dropped improperly, or spoofing
> traffic may
> > 1009     be permitted improperly.  For SAV mechanisms that use BGP data
> as
> > 1010     input for generating SAV tables, the use of applicable BGP
> routing
> > 1011     security methods is important.  Such methods include mechanisms
> for
> > 1012     the prevention, detection, and mitigation of route hijacks,
> route
> > 1013     leaks, and AS_PATH manipulations.
> >
> > <major> SAV-specific information is a new channel between ASes, and
> Section 9
> > does not cover three of its exposures. It can disclose what an operator
> keeps
> > out of BGP on purpose: hidden or DSR prefixes, prefixes sent with
> NO_EXPORT,
> > customer relationships. Stale or replayed information causes improper
> permit
> > or improper block. And anyone able to influence another AS's SAV inputs
> can
> > get legitimate traffic dropped, so improper block is itself a
> > denial-of-service vector. The BGP security methods mentioned are also
> uncited.
> > Perhaps:
> >
> > NEW (end of Section 9)
> >    SAV-specific information may disclose prefixes, routing policy, or
> >    business relationships that an AS deliberately keeps out of BGP.
> >    Stale or replayed SAV-specific information can cause improper
> >    permit or improper block, and an attacker able to influence the SAV
> >    inputs of another AS can use improper block to deny service to
> >    legitimate traffic.  Relevant BGP security methods include route
> >    origin validation [RFC6811], BGPsec [RFC8205], and route leak
> >    prevention [RFC9234].
> >
> > NEW (end of Section 5.4)
> >    Such mechanisms should also bound the freshness of SAV-specific
> >    information and allow its distribution to be limited to the ASes
> >    that need it.
> >
> > <EoRv21>
> >
> >
> _______________________________________________
> savnet mailing list -- savnet@ietf.org
> To unsubscribe send an email to savnet-leave@ietf.org
>