[Ndn-lib] Request for comments about KeyLocator in Data packets in NDN-TLV format

Alex Afanasyev alexander.afanasyev at ucla.edu
Thu Apr 10 16:30:33 PDT 2014


Our NDN-TLV specification (http://named-data.net/doc/ndn-tlv/data.html) currently defines KeyLocator as part of specific signature types, since it may make sense for some signatures and doesn't make sense for other.  

However, yesterday I realized that there is inconsistency in this specification.  In Interest packet we have defined PublisherPublicKeyLocator, which can be present in any Interest packet as part of selectors (http://named-data.net/doc/ndn-tlv/interest.html#selectors).   The "problem" is that the generic Interest defines KeyLocator, while "generic" Data does not and in order to match selector, one needs to check signature type, figure out whether this signature can contain KeyLocator and then try to match it (this is just logical view, implementation, obviously, can do it more efficiently).

What I was thinking is to define optional KeyLocator field as part of generic specification of SignatureInfo field (http://named-data.net/doc/ndn-tlv/signature.html), instead of it be a mandatory field for some signature types.   This does not change the wire format in any ways, just reorganizes logical structure of the specification.

I would appreciate your opinions on the issue.


More information about the Ndn-lib mailing list