Suvani,

  From the slave's perspective, a packet with SeqNum X+1 from
  the master is an ack that the master received slave's packet X,
  so, the slave must remove the LSA descriptions it sent in
  packet X from the DB summary list and send the new portion.

  HTH.

--
Alex Zinin
Wednesday, May 16, 2001, 9:44:15 AM, Suvani Kaura wrote:
> Hi Alex,
> I think last time I missed something while phrasing my question: The slave sends the DBD packet
> with seq. num. "x+1" in response to a DBD from the master with seq. num. "x+1" (this pkt from the
> master contains the summary of LSAs in its database). So, in my original question I meant to say
> that after receiving the DBD with the next higher seq number from the master, the slave is sending
> the exact same lsa headers (as it sent in DBD with seq num. "x" ) as part of the DBD packet with
> seq Number "x+1". The table below illustrates this:
> Sender Seq#  Flags  DBD Contents
> R1     x     IMM    Empty
> R2     z     IMM    Empty
> R2(S)  x      MS    DB Summ. with "n" LSA Hdrs
> R1(M)  x+1     M    DB Summ. with "m" LSA Hdrs
> R2(S)  x+1    MS    DB Summ. with the same "n" LSA
>                     Hdrs as prev DBD sent by R2.
> ...
> ...
> Is this behaviour of the slave correct?
> The RFC says on page 103-104:
> " Each new Database
> Description Packet copies its DD sequence number from the
> neighbor data structure and then describes the current top of
> the Database summary list.  Items are removed from the Database
> summary list when the previous packet is acknowledged."
>>From my understanding, the master considers a DBD pkt from the slave as an acknowledgement when
> the slave's DBD has the same sequence number as the last DBD pkt that the master had sent. But the
> RFC does not talk about what an acknowledgement is for the Slave. In the above example, should the
> slave have removed the items that it sent in the DBD with seq. number "x" upon receiving the DBD
> from the master with seq. number "x+1"?
> Thanks in advance,
> Suvani.
> --- Alex Zinin <azininNEXSI.COM> wrote:
>> Suvani,
>>
>>  This is the master who increments the SeqNum in the DBD
>>  packets. Even if the slave have to retransmit its DBD
>>  packet (which is probably what is happening), it should
>>  not increment the SeqNum.
>>
>> --
>> Alex Zinin
>>
>> Tuesday, May 15, 2001, 1:11:30 PM, Suvani Kaura wrote:
>>
>> > Hi All,
>>
>> > Is it OK for the slave in the Database Exchange
>> > process to first send some LSA Headers in response to
>> > a packet from the master (with sequence number "x"),
>> > and in its next DBD packet (with seq number "x+1") to
>> > resend the same LSA headers? In between sending "x"
>> > and "x+1", the slave didn't receive any requests or
>> > DBD packets from the Master with those LSAs in them.
>>
>> > RFC 2328 isnt explicit about this scenario.
>>
>> > Thanks in advance,
>> > Suvani.
> __________________________________________________
> Do You Yahoo!?
> Yahoo! Auctions - buy the things you want at great prices
> http://auctions.yahoo.com/

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]