[dtn] rfc5050(bis) proposed revisions

["Templin, Fred L" <[e-mail]>]

Hello,

Below is a list of (proposed) revisions for rfc5050(bis) as found in
Appendix A of 'draft-burleigh-bpv7'. Please post any comments or
suggestions to the list.

Thanks - Fred
[e-mail]

---

Appendix A.                  Summary of Revisions

   This specification differs from RFC-5050 in a number of ways.  The
   revisions that seem to the author to be most significant are listed
   below:

     . Amplify the discussion of custody transfer.  Move current
        custodian to an extension block, of which there can be multiple
        occurrences (providing possible support for the MITRE idea of
        multiple concurrent custodians, from several years ago); define
        that block in this spec.
     . Add the notion of "embargoes", i.e., what do you do when a
        route unexpectedly goes bad for a while?  This entails adding
        another extension block (Forwarding Anomaly) and another
        administrative record (Reopen Signal).
     . Incorporate the Compressed Bundle Header Encoding [RFC6260]
        concepts into the base specification: nodes are explicitly
        identified by node numbers, and operations that pertain to
        nodes are described in terms of node numbers rather than
        endpoint IDs.
     . Add basic ("imc") multicast to the BP spec.  This entails
        adding another administrative record, Multicast Petition.
     . Add Destination EID extension block for destinations that can't
        be expressed in "ipn"-scheme and "imc"-scheme URIs.  Define it
        in this spec.
     . Incorporate the "Extended Class of Service" features into the
        base specification.
     . Restructure the primary block, making it immutable.  Add CRC.
        Remove the dictionary.
     . Add optional Payload CRC extension block, defined in this spec.
     . Add block ID number to canonical block format (to support
        streamlined Bundle Security Protocol).
     . Add bundle age extension block, defined in this spec.
     . Define two other extension blocks in this spec: previous node
        number, hop count.
     . Clean up a conflict between fragmentation and custody transfer
        that Ed Birrane pointed out.
     . Remove "DTN time" values from administrative records.
        Nanosecond precision will not be meaningful among nodes whose
        clocks are not closely synchronized, and absent that feature
        the administrative record's bundle creation time suffices to
        indicate the time of occurrence of the reported event.
     . Note that CL protocols are supposed to discard data that they
        find to have been corrupted.