[Hipsec-rg] future work on naming objects in HIP

[miika.komu at hiit.fi (Miika Komu)]

Henderson, Thomas R wrote:

Hi,

>> -----Original Message-----
>> From: Gyu Myoung Lee [mailto:gmlee at icu.ac.kr] 
>> Sent: Wednesday, April 08, 2009 10:01 AM
>> To: Henderson, Thomas R; hipsec-rg at listserv.cybertrust.com
>> Cc: 'Samu Varjonen'; Pascal.Urien at enst.fr; 'Robert 
>> Moskowitz'; 'Andrei Gurtov'
>> Subject: RE: future work on naming objects in HIP
>>
>>
>> Dear All,
>>
>> As we have discussed several times on object related work, it 
>> seems that
>> most of experts have some interesting points but we still 
>> have technically
>> alternative solutions. Thus, it will be more productive way 
>> to officially
>> work together on this topic as a RG item  
>>
>> First of all, in order to initiate this work as a RG item, I 
>> believe that we
>> sincerely need to work on a broadly scoped high-level informational
>> document. In addition, we also need to develop documents on detailed
>> solution spaces simultaneously.
>>
>> In this case, hip-object document will be a candidate for high level
>> informational document which contains basic concept, various technical
>> options and use cases, etc. I can take the responsibility for 
>> developing
>> this document and also would like to invite several experts 
>> for valuable
>> comments and inputs.
>>
> 
> Pascal, Robert, Miika,
> 
> You've all expressed some interest in pursuing this topic as a RG work
> item.  Gyu Myoung Lee above suggested a possible concrete path forward.
> Would you like to pursue the above with him, or other activities in
> parallel?  Does anyone else want to express opinions about going forward
> with this work?

I'd support the document to become an RG item. If nothing else, I can at 
least offer feedback. Few comments below:

   * I suggest that authors adopt proposal #2 as suggested in section 
5.3.2 because it is more backwards compatible than #1.
   * What are the implications of sending object id in R1 or I2 instead 
of HOST_ID? It affects the HIT calculation and signatures are not 
needed. Do we need HIPv2 in the header or a new critical flag?
   *  RFID communications probably requires running HIP without IP, or 
does it? If yes, this has some implications on the way the HIP header 
checksum should be calculated and we need discuss about fragmentation 
issues. Also, I am not so familiar with mobility mechanisms in ad-hoc 
networks.
   * Do the authors have plans to implement the draft?