At 12:41 PM 1/30/02, Cebuano wrote: >i would think that regardless if there's a new >router with a higher RID that comes on line, the DR/BDR >should be the default MASTER to initiate the exchange >since he's got all the topology/links info in the area, except >of course for scenarios where there is no DR/BDR.
Both neighbors have information to send. The master/slave business is just a temporary relationship to allow the neighbors to exchange their information in a reliable fashion. There's no real reason for one router instead of the other to become the master. Remember that protocol design is modular. You should keep the database synchronization process separate from the DR/BDR election. The synchronization process is the first step in the adjacency-building process. Each router describes its database by sending a sequence of database description packets to its neighbor. Each database description packet has a sequence number. Database description packets sent by the master (polls) are acknowledged by the slave through echoing of the sequence number. Both polls and their responses contain summaries of link-state data. The master is the only one allowed to retransmit database description packets. The OSPF protocol developers could have chosen some other method to ensure reliability, such as opening a TCP session or inventing a client/server protocol with the DR acting as the server on networks that have a DR. Instead they invented a master/slave protocol. It's just how they decided to implement it. If you think about it, you can see that their method has some advantages. If you were in a computer science protocol development class, you could write an essay on why their method is best. As a CCIE candidate, however, I'm tempted to say, "why ask why?" ;-) Priscilla >Elmer >----- Original Message ----- >From: "Rogell, Dennis" >To: "'Cebuano'" >Sent: Wednesday, January 30, 2002 12:19 PM >Subject: RE: OSPF Startup questions [7:33711] > > > > You can make the Dr the higher rid , and the answer to 3 is it looks at > > sequence numbers not timestamps.If the information it receives is the same > > but the sequence number is greater that will be entered into ls database. > > > > hth > > > > Dennis Rogell CNE, CCNP > > nextiraone > > Formally Milgo Solutions > > Email : [EMAIL PROTECTED] > > Phone: (954) 846-5128 > > > > > -----Original Message----- > > > From: Cebuano [SMTP:[EMAIL PROTECTED]] > > > Sent: Wednesday, January 30, 2002 08:38 > > > To: [EMAIL PROTECTED] > > > Subject: OSPF Startup questions [7:33711] > > > > > > Hi, group. > > > I have a few questions to iron out regarding OSPF startup. > > > > > > 1. EXSTART - master/slave is created between each router and its >adjacent > > > DR/BDR. > > > Q: But this "election" on who the master will be is mute because the > > > router with the higher RID (thus) the DR/BDR acts as the master, right? >If > > > not, shouldn't the DR act as the master anyway since he ( or she) is the > > > central distribution point for the area's topology? > > > > > > 2. LOADING - slave router sends an LS request if the master's DBD has a > > > more > > > up-to-date link-state entry. > > > Q: "up-to-date" meaning timestamps?? But...what if there's no NTP > > > server > > > to synchronize them? > > > > > > Thanks. > > > Elmer ________________________ Priscilla Oppenheimer http://www.priscilla.com Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=33775&t=33711 -------------------------------------------------- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
