questions on nhrp-04.txt and nhrp-mib-00.txt for Stockholm ROLC

[David Horton <[e-mail]>]

re: nhrp-04.txt

Should the request message contain a "Station is a router" bit, analagous
to the Q bit in the request message?

=========================================================

re: nhrp-mib-00.txt

On page [9] Is AddressType the set of numbers in rfc1700 (Assigned Numbers)
Page 91 "Address family numbers"? (Which is 16 bit).

On page [13/14] Should the type of nhrpIfNetworkID be IpAddress?

On page [17] Declaration of nhrpServerDestNet duplicated

On pages [17/18] Last attribute inconsistently named 
(a) nrhpServerEntryStatus and (b) nhrpServerRowStatus, shouldn't it just be
nrhpServerStatus

On page [20] I presume nhrpAddrDestLinkAddr corresponds to 
{NBMA length + NBMA address} fields of the NHRP packets. Isn't there
a need for an address type field to correspond to the {Address Type}
field of the NHRP packet?

[On page 20] does nhrpAddrEntryType need an additional type for NHS servers
receiving a Request packet containing their source NHRP address. (Or is this
treated as a register or one of the other types?)

On pages [19/21] Last attribute inconsistently named 
(a) nhrpAddrEntryStatus and (b) nhrpAddrRowStatus, shouldn't it just be
nhrpAddrStatus?

Does an ifIndex have to be defined for NHS operating in server mode? 
MUST the NHS be part of the LIS that it is serving? Could one NHS serve
multiple LIS from a single central address?

On page [16] NextHop is first index.
I would have expected the order to be { DestNetAddr, DestNetMask, NextHop },
is there a reason the other ordering was chosen?

On page [19] Is a NextHop needed too?
I don't understand the role of the net mask fully. 
In particular with caching for a server mode configuration.

1.1.1.1    1.1.1.*       1.1.2.*     1.1.2.1        1.3.*.*
    H1 ---{ LIS 1 }-----{ LIS 2 }-- Router B - - - -
             |             |
  (1.1.1.2) NHS 1         NHS 2 (1.1.2.2)

Say H1 sends in a request for an address in 1.3.*.* somewhere say 1.3.1.1, 

NHS1 will forward the request to NHS 2 as there is an entry in the nhrpServer
table {NextHop=1.1.2.2(NHS 2), DestNet=1.3.0.0, DestMask=255.255.0.0}

NHS 2 responds with a NHRP Reply {Dest=1.3.1.1, NextHopIP=1.1.2.1, NBMA=RouterB}

NHS 1 sends this back to H1. What can NHS 1 cache from this in the nhrpAddr?
{DestNet=1.1.2.1, DestMask=255.255.255.255, LinkAddr=RouterB's NBMA} ?
-- this seems correct but not much use
{DestNet=1.3.1.1,DestMask=255.255.0.0,LinkAddr=RouterB's NBMA}?
-- this seems wrong
If there were a NextHop field, we could avoid subsequent requests having to
be forwarded to NHS 2, and be able to cache :-
{DestNet=1.3.1.1,DestMask=255.255.0.0,LinkAddr=RouterB's NBMA,NextHopIP=1.1.2.1}

In server mode, my understanding is that Router B doesn't have to have 
an NHRP server resident, just know where one can be found. In fabric
mode I can see how the model works.

on page [8] an 'experimental' branch?
Can we get an experimental branch for this MIB?

and now some picky ones :-

on page [16] nhrpIfRowStatus should be ::= { nhrpIfEntry 11 }

on page [16] syntax for nhrpServerEntry should be NhrpServEntry
Also shouldn't the ACCESS for the Entry be not-accessible?

on page [17] declaration for nhrpServerDestNet is repeated

on page [20-21] there are two { nhrpAddrEntry 4 } objects

on page [21] nhrpAddrRowStatus should be ::= { nhrpAddrEntry 7 }

There are some other little syntax errors, like extra or missing ",".
I have a version that I pass to our SNMPv1 parsing tools, if anyone wants it?

regards,
David

 =============================================================================
 David Horton
 Centre for Information Technology Research
 University of Queensland, 4072        Australia
 Email: [e-mail]         Phone [phone]   Fax [phone]
 =============================================================================