Re: [Mip4] Comments on draft-sastry-mip4-string-extension-01
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Mip4] Comments on draft-sastry-mip4-string-extension-01



Hi Kent,

On Oct 1, 2005, at 9:11 AM, Kent Leung ((kleung)) wrote:
My main question is how to guarantee interoperability. The
draft says that the message can be displayed on a UI, logged,
or handled any other way, as the implementor wishes. This
obviously creates issues in terms of interoperability and
reliability of the system.



The message is purely informational on the MN. And how this information
is conveyed to a user is implementation dependent.

OK, I understand the basic goal.. But I haven't seen many examples of protocol extensions designed to simply carry user-oriented informational messages about a low-layer protocol. Is there something here that couldn't be achieved by SNMP?


It is easy to imagine displaying or logging a message when we
have a traditional UI or syslog, and the message is the
suggested registration failure or welcome/motd. In other
circumstances it might be unpredictable what a message will
cause. This is especially an issue when interoperating
between one vendor's HA/FA and another's MN.



Typically, a MN will have some UI for configuration, monitering, and
debugging purpose, as well as some logging facility. So it's up to the
software to decide how to handle this information from the HA about the
registration response.




The Text field's definition says the text "is intended to be
human readable, and MUST NOT affect operation of the
protocol." However, I would assert that this is not the case
at all -- what is the purpose of the text if no effect is
desired? It suggests user intervention on failure, which
would somehow have to affect the operation of the protocol
(i.e. causing configuration changes) to be useful. To go even
further, the example given shows a case where the HA would
reject the Reg. Req. due to timestamp mismatch, and the MN's
implementation would shield the user from the text message,
sending another registration. How would this be possible
unless the MN somehow parsed the text message on behalf of the user?



Sorry if this example is confusing. The timestamp mismatch is part of
normal Mobile IPv4 operation to synchronize the timestamp between the MN
and HA. So actually, the MN sending another registration is in response
to the error code received and has nothing to do with the informational
message. Maybe we should come up with a different example instead.
Maybe along the line that HA sends a registration deny with a message
"Prepaid quota reached. Please contact www.paymore.com to update
balance."

Right. My point was simply that you described how a MN could suppress the informational message if it could "figure out" and fix what the problem was without user intervention (here, based on the error code). But since the message is informational, the MN has no idea that the message intended for the user was that the prepaid quota had run out! Maybe it was an unrelated informational message.


I do understand your goal, if it is that informational messages should only be displayed if (1) an error condition exists that the MN can't figure out how to rectify with another attempt, or (2) the registration was successful. But then it should be stated as such.

One final question: in the security considerations section,
what is your interpretation of the difference for HA and FA,
and what if a malicious node tries to insert the message
after the M-H auth ext?



This is the same issue brought up on FA Error Extension draft. We
recommend using MN-FA Auth Ext whenever possible. If not, then the harm
is really some annoying message. For example, if the user goes to
www.paymore.com to add more money to his account, he'll see that the
fund is sufficient already. We should recommend that the message should
not compromise any user or MN information as the text string is clear
and may be added by an unscrupulous node.

I guess it depends on the message itself. It could be used for social engineering or some sort of phishing attack that's not as innocuous as a simple error message. Of course, that would depend on what the implementors use the messages for.


Thanks for your reply.
TJ


-- Mip4 mailing list: Mip4 at ietf.org Web interface: https://www1.ietf.org/mailman/listinfo/mip4 Charter page: http://www.ietf.org/html.charters/mip4-charter.html Supplemental site: http://www.mip4.org/




Note: Messages sent to this list are the opinions of the senders and do not imply endorsement by the IETF.