Re: [radext] Confirmation of discussions on RADIUS/(D)TLS at IETF117

Heikki Vatiainen <hvn@radiatorsoftware.com> Wed, 02 August 2023 20:38 UTC

Return-Path: <hvn@radiatorsoftware.com>
X-Original-To: radext@ietfa.amsl.com
Delivered-To: radext@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id EE339C151534 for <radext@ietfa.amsl.com>; Wed, 2 Aug 2023 13:38:44 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.906
X-Spam-Level:
X-Spam-Status: No, score=-6.906 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_HI=-5, RCVD_IN_ZEN_BLOCKED_OPENDNS=0.001, SPF_HELO_NONE=0.001, SPF_PASS=-0.001, T_SCC_BODY_TEXT_LINE=-0.01, URIBL_DBL_BLOCKED_OPENDNS=0.001, URIBL_ZEN_BLOCKED_OPENDNS=0.001] autolearn=ham autolearn_force=no
Authentication-Results: ietfa.amsl.com (amavisd-new); dkim=pass (2048-bit key) header.d=radiatorsoftware-com.20221208.gappssmtp.com
Received: from mail.ietf.org ([50.223.129.194]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id y-z0nB9gib0E for <radext@ietfa.amsl.com>; Wed, 2 Aug 2023 13:38:41 -0700 (PDT)
Received: from mail-wr1-x430.google.com (mail-wr1-x430.google.com [IPv6:2a00:1450:4864:20::430]) (using TLSv1.3 with cipher TLS_AES_128_GCM_SHA256 (128/128 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits) server-digest SHA256) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 14489C14CE36 for <radext@ietf.org>; Wed, 2 Aug 2023 13:38:40 -0700 (PDT)
Received: by mail-wr1-x430.google.com with SMTP id ffacd0b85a97d-31759e6a4a1so174632f8f.3 for <radext@ietf.org>; Wed, 02 Aug 2023 13:38:40 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=radiatorsoftware-com.20221208.gappssmtp.com; s=20221208; t=1691008718; x=1691613518; h=to:subject:message-id:date:from:in-reply-to:references:mime-version :from:to:cc:subject:date:message-id:reply-to; bh=9M9p6hYCSGTYqYe50pUMbaThLXSpTHbY4X09NFGyUxU=; b=RXcemkLM53jRCuw0Ml1OtDJqtmXxvLXCr6Uo0jPRJKvZa8Fj5Oljr4WlKsGos9Xyzx D2ESv8SHM6V8jYbYnVWzyPr4kv2Agx4Miw/Apn1qchxcxKkQ2lVvUXHIuTNapRzjyuaD pRAZEcGCZXbJVwkLtIbXViEV0ZzVAeOzPmvk1rJzPD9Z9Vbm5kY8eyx0rvggSj+DprVg At7ICBnZ2eTZPxchxkajMOP66cIuWIfuePPsEsr5L/P/Y7nISTf907s0vlfOhiPftePx YI5dYDR9+n9e6BndoUa6xGg+fu/C07qRXJEG2K8wRAonMB7WtsWg39/NhCJNb5oqYLxT I59w==
X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1691008718; x=1691613518; h=to:subject:message-id:date:from:in-reply-to:references:mime-version :x-gm-message-state:from:to:cc:subject:date:message-id:reply-to; bh=9M9p6hYCSGTYqYe50pUMbaThLXSpTHbY4X09NFGyUxU=; b=eHBHMj8PbCtFvrQGmvKrBUljCAQRFQQwkWkXUlRqMS2SlwTXbq+b8TYHvit9A0Cp1C LE22F9tlNNO5V1BsUgg/KASuB6FjjyIbMcv8upcXFI7o8zkAZ4Fb60vOT7tXt/VT+L9G LxeI5zOQWCk9KNwJh7zbuTejjWouUn43JO1DIg8zBGi/3zoorkU4XpEnlT3MVktHjMEX hutpHJ07e0u0NIXwzcqZwWuMLe6fgRo/UrZVZQq94os7R/NC5ofsuVg1MXU27ZEQBXN8 m4vZDg1FERbqHZ/TueHSYP2e0QHGsdaV0gUMHJH0C++QTM5MQxdQfanMF+Z9Jpsw+8QY SsxA==
X-Gm-Message-State: ABy/qLZVZgHgfn0PiTt3w+v4H2CHPKGQIh5Qdu/ZE6bcqxVsHyx12Jab uUc4qmAQz3Uuu8idL5wDhE9CShZGn5WKfzX1RHawTmeaO0TlDUUeowo=
X-Google-Smtp-Source: APBJJlFDZvSrgVDH3cNoaBKT6U7J73ZEVwnlrm/2u1629NBn17oGQ7iRKYt9pssPl6tE7+eT/uXtctE8K0cxGsopkiY=
X-Received: by 2002:a5d:69c1:0:b0:314:c9c:bc96 with SMTP id s1-20020a5d69c1000000b003140c9cbc96mr5420601wrw.7.1691008718197; Wed, 02 Aug 2023 13:38:38 -0700 (PDT)
MIME-Version: 1.0
References: <a401c4c2-3ddd-3d93-3fbc-b3fc02a1d26c@dfn.de>
In-Reply-To: <a401c4c2-3ddd-3d93-3fbc-b3fc02a1d26c@dfn.de>
From: Heikki Vatiainen <hvn@radiatorsoftware.com>
Date: Wed, 02 Aug 2023 23:38:22 +0300
Message-ID: <CAA7Lko_dZjQ3sCWcJOVjeosqeteg2PnUN=Qeyg4ujsbqSE0CSg@mail.gmail.com>
To: radext@ietf.org
Content-Type: multipart/alternative; boundary="000000000000f5f7e20601f6a58d"
Archived-At: <https://mailarchive.ietf.org/arch/msg/radext/B_bzRwKxRSt1IWfG0Q0BVmSqP5k>
Subject: Re: [radext] Confirmation of discussions on RADIUS/(D)TLS at IETF117
X-BeenThere: radext@ietf.org
X-Mailman-Version: 2.1.39
Precedence: list
List-Id: RADIUS EXTensions working group discussion list <radext.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/radext>, <mailto:radext-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/radext/>
List-Post: <mailto:radext@ietf.org>
List-Help: <mailto:radext-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/radext>, <mailto:radext-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 02 Aug 2023 20:38:45 -0000

On Wed, 2 Aug 2023 at 17:02, Jan-Frederik Rieckers <rieckers@dfn.de> wrote:

There were some discussion points that were raised during the meeting
> still waiting for confirmation on this list. I don't know if the chairs
> want to issue a "formal consensus call" on the list, but I just want to
> get a feeling of the consensus to start producing text.
>

When I got ready to remotely join the meeting at midnight, the meeting had
already concluded about two hours ago. Teaches me to double check the
schedule the next time. The minutes, thanks Q, and video are a very
valuable resource.


> Arran was so kind to submit a pull request regarding the specification
> of application layer watchdogs, DTLS heartbeats, Connection closure
> etc., as agreed in the session.
> It can be found here:
> https://github.com/Janfred/draft-janfred-radext-radiusdtls-bis/pull/1
>
>
Looks good to me. There's one thing I'd say needs a fix. Line 570 in the
patched version (and the original RFC) is:

Once a DTLS session is established, a RADIUS/DTLS server MAY use DTLS
Heartbeats {{!RFC6520}} to determine connectivity between the two servers.


Should "the two servers." be "the two DTLS endpoints."?
https://github.com/Janfred/draft-janfred-radext-radiusdtls-bis/pull/1/files#diff-3f73d9747b6cff6fdf2b0cf981a80baef79fe8493f847b948e9e47bcb6a67616R570


> The other questions were:
>
> **Mandatory-to-implement Protocol**
>
> Should be either RADIUS/TLS or RADIUS/DTLS mandatory to implement or
> should both be mandatory?
>

The recent messages on the list suggest 'both'. In this case, could it be
mandatory for the servers and either for clients. For example, an Ethernet
switch may want to implement only RADIUS/DTLS (one vendor does) and some
other, possibly constrained, client device way implement just RADIUS/TLS.

One reason could be that the platform the implementation uses, supports
only one or the other. Servers are likely to run on more resourceful
platforms. Even on server side, for example, Rust doesn't have DTLS, at
least yet.


> **Single Port for auth/accounting**
>
> Current implementations use 2023/tcp and 2023/udp for both
> authentication and accounting. For load balancing/traffic engineering
> reasons it may be good to use different ports as it is with RADIUS/UDP.
>
> The rough consensus in the room was that the document should stay with
> the single port.
>

The point about separating auth and acct is a good one and often done with
RADIUS/UDP. The two types of messages often have completely separate
handling on the server side and can be handled with separate processes,
configurations, etc.

On the other hand, if I remember correctly, one of the resons eduroam
doesn't proxy accounting is that accounting was often blocked with firewall
and not treated with equal care as authentication. In other words, a single
port for all message types is easier to handle from this perspective.

Most likely different roaming consortiums also would like to see a single
port.  This especially because they use DNS based roaming where it's easier
to ensure that one port passes firewalls than to try to manage multiple
ports.

A smart client could make this configurable. For example, a broadband
network gateway (BNG) could have an option for separate ports for
RADIUS/TLS and RADIUS/DTLS similarly as it already has for RADIUS/UDP. This
would make it possible to what's already done with RADIUS/UDP, typically
within the organisation that runs the BNG, and known-to-work peer
organisations, such as mobile network operators (MNOs) and Virtual MNOs.

To summarise my suggestion: Continue with a single port but remind that
there are cases, and types of devices, which could benefit from separate
ports.

**Deletion of MIB References**
>
> The new document has currently no text regarding the MIBs. Should there
> be some text about MIBs for RADIUS/(D)TLS.
>
> There were no strong opinions in the room, since it is not sure who
> actually uses these MIBs. So if there are opinions about this on this
> list, please share :)
>

So far those who have asked about MIB support have been organisations that
have not used RadSec. MIBs are used, but their current users appear to be
different types of organisations that use RADIUS/(D)TLS. When RadSec
becomes more widely used, this may chance. It may be best to continue
keeping MIBs separate from the RADIUS/(D)TLS document.

**Watchdogs**
>
> (See PR from Arran)
> The current spec is ambiguous about the usage of watchdog mechanisms.
>
> The rough consensus was to use the same watchdog mechanism for TLS and
> DTLS, namely Status-Server.
>

I suggest using RFC 3539 in the body of section 3.4:
https://www.rfc-editor.org/rfc/rfc3539#section-3.4

Any reply from the peer is good enough, not just reply to Status-Server.
Any reply will also push the Status-Server timer forward so that it
triggers only when there haven't been any requests sent over the
connection, including UDP neighbors.

This is what Diameter does and it appears to work fine there. A bit related
to Diameter, because Diameter uses peer-to-peer model where both peers can
equally send requests, the watchdog polling is done by the both peers. This
ensures the request-answer messaging works for the both peers. When a
RadSec connection is used for reverse CoA, should Status-Server run to both
directions in this case? This would ensure that the CoA server is still
responding. This is a new idea, I just thought about it.

**Add ref to RFC9325 (TLS/DTLS BCP)**
>
> The old spec has some text about MTI cipher suites, implementation
> hints, ...
>
> Since there is RFC9325 which gives recommendations about secure use of
> TLS and DTLS, we should just reference this.
>

Sounds good!

-- 
Heikki Vatiainen
hvn@radiatorsoftware.com