Routing Protocols (Part II)

Interior Gateway Protocol (IGP):

Border Gateway Protocol (BGP)

Computer Network II (CSE 461)

 BGP Autonomous Systems
• An autonomous system (AS) is a collection of networks under
a single technical administration.
• IGP operates within an autonomous system.
• BGP is used between autonomous systems.
• Exchange of loop-free routing information is guaranteed.

2
 IGP and EGP
• Interior Gateway Protocol (IGP) operates
within the same autonomous system.
– Examples of IGP are: RIP, OSPF, and EIGRP
• Exterior Gateway Protocol (EGP) operates
between multiple autonomous systems.
– Example of EGP is BGP.
– BGP is the only EGP available.

3
 Basic BGP Features
• Open Standard
• Exterior Gateway Protocol
• Designed for inter-autonomous system domain
routing.
• Designed to scale huge internetwork like Internet.
• Classless:
– Supports net classes: Fixed -FLSM, Variable- VLSM, Inter-
domain- CIDR, auto and manual summary (BGP 4).
• Updates are incremental and triggered.
• Path vector protocol

4
 Basic BGP Features (cont.)
• It sends updates to manually defined neighbors
as unicast.
• BGP is an application layer protocol which uses
TCP for reliable transmission.
• It uses port 179 for TCP.
• Metric = Attribute
• Administrative distance
– 20 external updates (eBGP);
• path is coming from different autonomous system.
– 200 internal updates (iBGP);
• path is coming from the same autonomous system.

5
BGP: Autonomous System Path
Advertisement

6
BGP: Loop Prevention Mechanism

7
BGP: Loop Prevention Mechanism
(cont.)

AS 500 sending all information, it

knows, to AS 100.

Autonomous System: Path Loop Detection

8
 Application of BGP
• When to use BGP
– Autonomous system working as transit
autonomous system (example is ISP)
– Autonomous system connected to multiple
autonomous system (when the autonomous
system is multi-homed) data traffic path entering
or leaving autonomous system need to
manipulated.

9
 Application of BGP (cont.)
• When not to use BGP
– When we have single-homed autonomous system
– Lack of resources like memory and less processing
power in routers
– Limited understanding about BGP route filtering
and path selection process

10
 Types of ISP Connections
• Single-homed
• Dual-homed
• Multi-homing
• Dual multi-homed

11
 Single-homed
• A site with a single ISP connection is single-
homed.
• This is fine for a site that does not depend
heavily on Internet or WAN connectivity.
• Routing: Either use static routes or receive a
default route from ISP and advertise the site
routes to the ISP.

12
 Dual-homed
• A dual-homed site has
two connections to the
same ISP either from
one router or two
routers.
• One link might be
primary and the other
backup or the site load
balance over both links.
• Either static or dynamic
routing is used.
13
 Multi-homing
• Multi-homing means
connecting to more than
one ISP at the same time.
• First reason: It is done for
redundancy, for example, if
one ISP fails.
• Second reason: It is done
for backup, for example, if
one ISP provides a better
path to frequently used
networks.
• BGP is typically used with
multi-homed connections.

14
 Dual multi-homed
• Dual-multi-homed
means two connections
to multiple ISPs.
• This gives redundancy.
• BGP is used with the
ISPs and can also be
used internally.

15
 ISP Connections Summary
• Single homed: you are connected to a single ISP
using a single link.
• Dual homed: you are connected to a single ISP
using dual links.
• Single multi-homed: you are connected to two
ISPs using single links.
• Dual multi-homed: you are connected to two ISPs
using dual links.

16
 Connecting to the Internet with BGP
route reception options
• Default routes from the provider(s)
– Easy on resources internal traffic routed to nearest
BGP router
• Some routes + default routes
– Allows for selection of some paths with others
falling back to a default route
• All routes (full table)
– Hard on resources but guarantees the most direct
path is taken

17
 BGP Neighbors
• BGP neighbors are routers forming TCP
connection for exchanging BGP updates.
• BGP neighbors are also called BGP peers or
BGP speakers.
• There are two types of BGP neighbor
relations:
– iBGP (internal BGP)
– eBGP (external BGP)

18
iBGP Neighbors

19
eBGP Neighbors

20
 BGP Databases
• Neighbor Table
– A list of all configured BGP neighbors
– It has to be manually configured using neighbor command
#show ip bgp summary
#show ip bgp neighbors
• BGP Forwarding Table
– A list of networks known by BGP along with their paths
and attributes
#show ip bgp
• Routing Table
– List of best paths to destination networks
#show ip route

21
 BGP Configuration
R1(config)# router bgp 500
R1(config-router)#neighbor 1.1.1.2 remote-as 500
R1(config-router)#neighbor 4.4.4.1 remote-as 500
R1(config-router)#network 10.1.1.0
R1(config-router)#network 1.1.1.0
R1(config-router)#network 4.4.4.0
R1(config-router)#no auto-summary
R1(config-router)#no synchronization
R1(config-router)#exit

22
 BGP Split Horizon Rule
• An update send to one iBGP neighbor should
not be send back to another iBGP neighbor.
• To prevent routing loops within an
autonomous system.
• Hence, BGP split horizon is necessary to
ensure that routing loops are not started
within an AS. Solutions with:
– Full mesh neighborship
– Use route reflectors

23
 Solution: BGP Split Horizon Rule
• Full mesh neighbor-ship
– Every router should be a neighbor of every other
router within the AS.
– iBGP neighbors need not to be directly connected but
they must be reachable to each other.
• Use route reflectors
– It is a special BGP ability to re-advertise routes learned
from an internal peer to other internal peers. So,
rather than requiring all internal peers to be fully
meshed with each other, route reflection requires only
that the route reflector be fully meshed with all
internal peers.

24
 BGP Split Horizon (Review)

Establishing
Full Mesh
Neighborship

25
iBGP full mesh scalability concerns
• Administration
– Configuration management on increasingly large number of
routers.
• Number of TCP sessions
– Total number of sessions = n(n-1)/2
e.g. 5*(5-1)/2 = 10 Sessions (required for peers’ adjacencies)
– Maintaining extreme number of TCP sessions create extra
overhead
• BGP table size
– A higher number of neighbors generally translate to a higher
number of paths for each route
– Memory consumption

26
 Router Reflectors
• Scalable alternative to an iBGP full mesh.
• Allows a router (route reflector) to advertise routes
received from an iBGP peer to another iBGP peer.
• Client updates server.
• Server updates all remaining clients.
• All clients should establish neighbor with only one
server.
• Clients will not establish neighbor with any other
client.
• In case if there are two servers, each server establish
neighborship with other server and clients.

27
Router Reflectors cont.

28
 Configuring BGP Authentication
• BGP supports authentication using Message
Digest 5 (MD5) algorithm.
• When authentication is enabled any TCP
segment, belonging to BGP, exchanged between
the peers is verified and accepted only if
authentication is successful.
• For authentication to be successful both the
peers must be configured with the same
password.
• If authentication fails, BGP neighbor relation
doesn’t establish.
29
Configuring BGP Authentication (cont.)

30
 BGP Peer Groups
• Apply same policies to multiple neighbors.
• Useful when many neighbors have the same
policies.
• Updates are generated once per peer group.
• Configuration is simplified:
BGP neighbors who share the same outbound policies together
can be under what is called a BGP peer group. Instead of
configuring each neighbor with the same policy individually, a
peer group allows you to group the policies which can be
applied to individual peers thus making efficient update
calculation along with simplified configuration.

31