[hipl-users] Re: State machine used for UPDATE packet handling

  • From: Miika Komu <miika.komu@xxxxxxx>
  • To: hipl-users@xxxxxxxxxxxxx
  • Date: Tue, 20 Oct 2009 15:02:59 +0300

Leonardo Arraez wrote:

Hi,

Hello everyone,

First of all, congrats for the article on the Linux journal, I cant wait to get a copy to read it. I can see how hip is getting to be known by more people everyday, which will surely help it's development and evolution sooner than later.

I have a couple of questions regarding the implementation I know you can help me to understand a bit better:

- The state machine used in UPDATE and NOTIFY packets is still the same as the one defined in the draft-ietf-hip-base-06 referenced in the code (line 16 file update.c)?

No, should be RFC5206. Are you sure your using the released version or the latest one?

http://hipl.hiit.fi/index.php?index=source

I recommend getting the latest nightly tarball or even preferrably by using the version control. For the past five months, all changes have been basically bug fixes or improvements.

- Is the UPDATE packet management still uses the parameter revision (locator, echo_request, echo_response) to handle the appropriate response?

I am not sure what you meant by "revision" in this context, but I am guessing that the answer to your question is yes.

The update.c is being rewritten completely as we speak and probably this will be finished in the end of this month.

- Is it possible (as a general design of a service) to use NOTIFY packets for new services? (rfc5203 makes reference of UPDATE, I2 and R2 packets only) as I would like to create a new service and use the NOTIFY packet to inform send some information to the subscribed node and not only use the NOTIFY packet to provide feedback on errors.

NOTIFY is reserved basically for error diagnostics. RFC5203 describes registration extensions for HIP. Why don't use those instead of NOTIFY? Lauri Silvennoinen has done a pretty good job in modularizing the registration extensions and it should be easy to add your own extensions:

http://hipl.hiit.fi/hipl/doxygen/registration_8c.html
http://hipl.hiit.fi/hipl/doxygen/hiprelay_8c.html

The registration does not support UPDATE packets yet because it's being rewritten.

I am working on the design of new services over hip however I'm not so sure on how strict must the design be according to the protocol and the RFCs.

I am assuming that you are doing research? Standards compliance affects deployment and feasiblity, hence it is a good bonus.

Other related posts: