Cisco Inferno Helping dropbears route packets

OSPF Part II – Building Neighbor relationships

Welcome back for part two. Neighbor relationships. This is an important part before we dive into discussing the details of exchanging routers with LSA's in part III. Before we start that we need to know how neighbors for and what happens area wide amongst routers.

Before we dig down deeper I am going to put up a list of acronyms to help the various components of OSPF piece together.

OSPF TERMS

• LinkState Data Base (LSDB) - database held by router - stores the topology table.
• Shortest Path First (SPF)  - Algorithm used by OSPF to analyse the LSDB. Determines best,       lowest cost router for each prefix
• LinkState Update (LSU) - OSPF message that holds detailed topology information(LSA)
• LinkState Advertisement (LSA)  - OSPF data structures that hold detailed topology information.      LSA's are held in the memory inside LSDB. They are communicated inside LSU's
• Area - Contiguous collection of routers. Actively learn all topology      information within an area.
• Area Border Router  - Has interfaces connected to two or more areas. One interface must be connected to the backbone. Holds topology data for each area it  is connected too. Calculates and advertises routes between areas
• Backbone Router (BB)   - Any router with an interface in area 0.
• Internal Router (INT) - Only has interfaces in one area
• Designated Router (DR) - Elected by OSPF to generate area wide LSA's
• Backup Designated Router  (BDR) - Is a backup DR if DR fails.

Let us start with the Database Exchange Process.

THE DATABASE EXCHANGE PROCESS

• Hello - Discovers neighbours. Has values that allow neighbour-ship, brings to a 2-way state and monitors link status
• Database Descriptor - Exchange brief versions of each LinkState Advertisement during initial topology. Allows router to know list of neighbour's known LinkState Advertisments.
• LinkState Request - Lists LinkState ID of Link State Advertisement sender of the LinkState Request would the like recipient of LinkState Request to supply during exchange.
• LinkState Update - Contains fully detailed LinkState Advertisement. Response to LinkState Request
• LinkState ACKnowledgement - Confirms receipt of LinkState Update

OSPF NEIGHBOUR STATES

When starting OSPF on a router and assigning interfaces to participate in the process the following occurs.

• Down - No hellos before dead interval expires
• Attempt - Neighbour defined with neighbour command.After hello TX before RX
• Init - Hello Received. No local RID or does not pass verification check.
• 2Way - Hello Received. RID and local neighbour checks pass.
• ExStart - Negotiating DBD sequence numbers. Uses Master/Slave logic.
• Exchange - Finished negotiating. Exchanging DBD packets
• Loading - DBD exchanged. Sending LSR, LSU and LSAck. Exchange full LSAs
• Full - Neighbours are fully adjacent. LSDBs for that area are identical.

What is a Designated router?

A designated router is the router that is unique to an area. OSPF uses a DR in an area for two reasons. One is to create and flood Type 2 network LSA's for that subnet. The second is to aid in the process of a database exchange in the subnet. An election for a router to become a DR takes place based on the information in the OSPF hello messages. Hello's list the priority and the RID. The election rules are as follows

• Choose the router with the highest priority
• If tied, highest RID
• The BDR is the next highest

The DR adopts the pseduonode ip address of 225.0.0.6. All routers in the area send their Type 1 LSA's and their ROUTER LSA's to this node. The Type 2 network LSA which the DR sends out is composed of the ROUTER LSA's that are connected to it. The DR creates that T2 LSA for the subnet. The DR assigns the LinkState ID of the DR's interface IP in that subnet. It also lists the DR's RID as the router advertising the LSA.

R1         R2

x

R3         R4

x = pseudonode - emits Type 2 LSA

When a DR and a BDR exist, no elections are made until one failts. If a DR fails then the BDR is elected the DR. This is the case even if a better DR joins the network. A new BDR is elected to fill the place. If a BDR fails the DR remains the same then only a new BDR is elected.

BREATHER!

Phwoar. A lot to start with. As we go through feel free to cross reference posts. I do appreciate the feedback as well as my peers calling me out if there are mistakes. I am no expert but this is a way for me to reinforce what I know. I will do a whole post on LSA's and the exchanging of routes next. Then we can configure some routers and check some output. Theory before the lab makes the lab gel better!

OSPF Part I – Basic OSPF

Welcome to my new series focussing on technologies and expanding on it. Working from basic to thorough we will establish the skills and knowledge required to understand and configure each technology.

Open Shortest Path First is what is known as a link state protocol. Every router in an area has the exact same routing information. By using a cost based metric it is possible to determine the shortest/fastest path a destination prefix. Think of it as having two exits to a maze. One exit is 200 meters ahead but covered in long grass that would slow you down. The other exit is a paved brick path that 250 meters long. OSPF knows the paved brick path is the fastest path out of the maze. OSPF does the same with routes. It knows where a given route is, how to take the quickest of multiple paths and deliver the packet efficiently.

Let's start with the basics.

OSPF - Open Shortest Path First

• Transport: IP
• Port Number: 89
• Metric: Cumulative costs of all outgoing interfaces in a router
• Update: 224.0.0.5 / 224.0.0.6 for Designated Routers
• Updates: Full table for new neighbours, partial otherwise
• Authentication: null, cleartext, and MD5
• VLSM: Includes mask with each route
• Tags: Supports route tagging
• Summarization: Supported at Area Border Routers
• Next-Hop: supports advertising routes with a different next-hop

OSPF uses a link state logic to calculate its routers. During the process of running OSPF between two neighbours the following occurs over three steps.

Step 1 - Neighbour Discovery involves finding the neighbouring routers running OSPF and exchanges enough information so routers know if they should become neighbours and exchange topology data. OSPF keeps a list of neighbours in the neighbour table. OSPF sends hellos to the multicast address of 224.0.0.5.

Step 2 - Topology exchange requires all OSPF routers in the area to send messages to each other. Information about the topology table is stored in the topology database known as the Link State Database. Stored in the LSDB is the router-id for each routers, each routers interface, IP address, mask, subnet, and finally the list of routers reachable by each router on each interface.

Step 3 - Router computation consists of each router analyzing the top data to choose the best from its OWN perspective. The link-state algorithm used is SPF. SPF choose the shortest route for each reachable subnet/next-hop/exit interface.

Enabling OSPF can be performed under the global router process or on an interface level. Remember that on a global level that any interface that has an IP address that is covered by a network range statement in OSPF will be included in the routing process. OSPF will check for additional information when forming adjacencies with neighbours.

The following information is sent in a hello. Hello packets are a critical part of the process of establishing and maintaining OSPF neighbours.

OSPF RID, Area ID,
Stub Area flag, Router Priority,
Dead Interval, Hello Interval,
DR IP, BDR IP, Subnet Mask,
Authentication Digest,
List of Neighbours.

Alright - The first part is done for now. Tune into part II shortly for more information regarding Designated and Backup Designated Routers, Router ID's and an introduction to LSA types.