Hi Magnus and Sharon,
Below please find replies to Sharons comments.
The replies include suggestions for clarification
that address the raised issues.
--On 16.01.07 15:54 +0100 Sharon Chisholm wrote:
I have been selected as the General Area Review Team (Gen-ART) reviewer
for this draft (for background on Gen-ART, please see
<http://www.alvestrand.no/ietf/gen/art/gen-art-FAQ.html>http://www.alves
trand.no/ietf/gen/art/gen-art-FAQ.html).
Please resolve these comments along with any other Last Call comments
you may receive.
Document: Middlebox Communications (MIDCOM) Protocol Semantic
Summary: This document is ready for publication/advancement, but has the
following nits that should be resolved.
Comments
---------
There is a lot of text in the document and I will confess I skipped over
some of it. I'm already late on the review and the document seems
reasonable (it has already been published), so I figured it was best
just to send what I had. I did note the following, where issue 5 is the
one that identifies something introduced by becoming a normative
document.
1. I didn't find where it was explained the reasoning behind being able
to send a request that reserves stuff on neither side of the middlebox.
It would seem a waste of processing.
We addresses this issue in an earlier email.
We do not think that there is an action required.
2. In section 2.1.1 it says "Asynchronous transactions allowing the
middlebox to change state without a request by an agent." but it would
seem that asynchronous transactions would allow for the reporting of the
change of state but not be able to trigger it. Or are we saying that the
asynchronous transaction does not originate from the agent but triggers
the state change?
In section 2.1.1:
OLD:
- Asynchronous transactions allowing the middlebox to change state
without a request by an agent.
NEW:
- Asynchronous transactions allowing to report state changes that
have not been requested by the agent.
3. In section 2.1.6 it says that capabilities include "Optional
interface-specific policy rule support: not supported or supported" but
it wasn't clear to be whether this was the advertisement of specific
optional capabilities or whether it was a simple yes or no that some
optional capabilities were supported.
This was a late change and phrasing is bad. The background is that
some NAT support interface-specific rules and some don't.
The capability information for this is just a 'yes' or a 'no'.
In section 2.1.6:
OLD:
- Optional interface-specific policy rule support: not
supported or supported
NEW:
- Support for interface-specific policy rules
4. In section 2.1.7, I was confused as to whether the parameter
identifying the requesting agent were specified or not. At first it says
they are but then it says 'Although, they are not necessarily passed
explicitly as parameters of the MIDCOM protocol, they might be provided
by the underlying (secure) transport protocol being used." Are they only
sent if they can't be sent via the underlying layers, because if so that
seems like a layering issue. The protocol should be independent. In
general, this section wasn't clear to me on this issue.
The semantics specification is independent of a concrete protocol
implementing the semantics. The section states that identifiers
need to be exchanged, either by the protocol or by the underlying
security mechanism.
Existing implementations of the semantics (MIDCOM MIB, SIMCO) solve
this point differently. The MIDCOM MIB uses identifiers provided
by SNMP, while SIMCO uses certificates.
5. In section 2.1.8 and at least one other place (section 3, first
paragraph) in the document it says something along the lines of
"However, an actual implementation of a middlebox may support only a
subset of these transactions." which I think is trying to say that
people shouldn't assume the middlebox supports everything but in the
context of a normative document may get misinterpreted to indicate that
the middlebox MUST NOT support everything.
OLD:
2.1.8. Conformance
The MIDCOM requirements in [MDC-REQ] demand capabilities of the
MIDCOM protocol that are met by the set of transactions specified
below. However, an actual implementation of a middlebox may support
only a subset of these transactions. The set of announced supported
NEW:
2.1.8. Conformance
The MIDCOM requirements in [MDC-REQ] demand capabilities of the
MIDCOM protocol that are met by the set of transactions specified
below. However, it is not required that an actual implementation of a
middlebox supports all these transactions. The set of announced supported
OLD:
3. Conformance Statements
A protocol definition complies with the semantics defined in section
2 if the protocol specification includes all specified transactions
with all their mandatory parameters. However, concrete
implementations of the protocol may support only some of the optional
transactions, not all of them. Which transactions are required for
compliance is different for agent and middlebox.
NEW:
3. Conformance Statements
A protocol definition complies with the semantics defined in section
2 if the protocol specification includes all specified transactions
with all their mandatory parameters. However, it is not required that
an actual implementation of a middlebox supports all these transactions.
Which transactions are required for compliance is different for agent
and middlebox.
6. Does the noauth state consume any resources. If so, doesn't this
leave the box vulnerable to attack? (section 2.2.5)
It may create state, depending on a concrete protocol specification.
However, the middlebox may at any time remove the state using an AST
transaction that returns the session to state CLOSED.
So, if an attacker creates too much state, the middlebox can easily
get rid of all sessions in state NOAUTH.
7. In section 2.3.4, it indicates "If the state or lifetime change was
requested explicitly by a request message, then the middlebox notifies
the requesting agent by returning the corresponding reply." Is this
instead or in addition to the asynchronous message?
This is explained two paragraphs below the sentence you cite:
In case of a PER, PRR, or RLC transaction, the requesting agent
receives a PER, PRR, or RLC reply, respectively. To all other agents
that can access the created, modified, or terminated policy rule (and
that participate in an open session with the middlebox) the middlebox
sends an REN message carrying the policy rule identifier (PID) and
the remaining lifetime of the policy rule.
8. In section 2.3.6, it wasn't obvious to me how I was support to
indicate a port range of lets say 200-205. The text says "A single
policy rule P containing a port range value greater than one is
equivalent to a set of policy rules containing a number n of policies
P_1, P_2, ..., P_n where n equals the value of the port range parameter.
".
We think this is well explained in the paragraph above.
It states
The port range parameter specifies a number of consecutive port
numbers. Its value is a positive integer. Like the port number
parameter, this parameter defines a set of consecutive port numbers
starting with the port number specified by the port number parameter
as the lowest port number and having as many elements as specified by
the port range parameter. A value of 1 specifies a single port
number. The port range parameter must have the same value for each
address tuple A0, A1, A2, and A3.
This means that for the range 200-205 you specify
port number: 200
port range: 6
Sharon Chisholm
Nortel
Ottawa, Ontario
Canada
Thanks a lot!
Martin and Juergen
_______________________________________________
Gen-art mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/gen-art