Hi all,
I'm not sure who has the editorial pen right now, but I thought I'd suggest
changes to the draft which implement the RFC5444 message changes we discussed
this week. As soon as we've added them to the draft, I'd like to send our
changes to the MANET list, too (or even before?).
Since we have this nifty generic AODVv2 message format, all we have to change
is the translation to RFC5444:
(side note: Imo, “The following sections show the TLVs that apply to each
address.” is a bit misleading, could we change this to “The following sections
discuss show the TLVs that apply to the specific addresses of the Address
Block.” or something in that direction to make clear that “all” is meant
per-address here?)
First, we'll have to remove the “(optional)” for TargSeqNum and specify that
the TLV shouldn't contain any value (as advised by chris here:
http://www.ietf.org/mail-archive/web/manet/current/msg17665.html) if the last
TargSeqNum is unknown.
8.1.4.2. Address Block TLVs for TargAddr
+----------------+--------------+------------+----------------------+
| Data Element | TLV Type | Extension | Value |
| | | Type | |
+----------------+--------------+------------+----------------------+
| TargSeqNum | TARG_SEQ_NUM | 0 | The last known |
| | | | TargSeqNum for |
| | | | TargAddr, if |
| | | | available, |
| | | | None otherwise |
+----------------+--------------+------------+----------------------+
Similarly, 8.2.4.1. Address Block TLVs for OrigAddr needs to be changed from
No Address Block TLVs apply to OrigAddr in a RREP.
to:
+--------------+---------------+------------+-----------------------+
| Data Element | TLV Type | Extension | Value |
| | | Type | |
+--------------+---------------+------------+-----------------------+
| OrigSeqNum | ORIG_SEQ_NUM | 0 | None |
+--------------+---------------+------------+-----------------------+
Then, we need to either introduce a new TLV type for all Addresses of a RERR
which are not PktSource or recycle one of the optional ones (i.e. make them
non-optional, but empty if no value for them exist). For the sake of
economicalness, I opted for the latter and made SeqNum mandatory:
8.4.4.2. Address Block TLVs for Unreachable Addresses
+----------------+-------------+------------+-----------------------+
| Data Element | TLV Type | Extension | Value |
| | | Type | |
+----------------+-------------+------------+-----------------------+
| SeqNumList | SEQ_NUM | 0 | Sequence Number |
| | | | associated with |
| | | | invalid route to the |
| | | | unreachable address |
| | | | if available, |
| | | | None otherwise. |
| MetricTypeList | PATH_METRIC | MetricType | None. Extension Type |
| (optional) | | | set to MetricType of |
| | | | the route to the |
| | | | unreachable address. |
+----------------+-------------+------------+-----------------------+
(Also, I'm not quite sure why Value for MetricTypeList says “None.”?)
And that should be about it!
Regards,
Lotte
