1. COMP/ELEC 429/556 Introduction to Computer Networks
Inter-domain routing
Some slides used with permissions from Edward W. Knightly, T. S. Eugene Ng, Ion Stoica, Hui Zhang
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

2. Previous focus: Intra-Domain Routing
Interior router
Border router
AS-1
AS-3
AS-2
Intra-domain routing protocol aka Interior Gateway Protocol (IGP)
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

3. Today’s focus: Inter-Domain Routing
Interior router
Border router
AS-1
AS-3
AS-2
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

4. Inter-Domain Routing Considerations
Global connectivity is at stake
Inevitably leads to one single protocol that everyone must speak
Unlike many choices in intra-domain routing
What are the requirements?
Scalability
Flexibility in choosing routes
If you were to choose, link state based or distance vector based?
Border Gateway Protocol (BGP)
A hybrid between link state and distance vector
“Path vector”
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

5. Border Gateway Protocol Part I: E-BGP
AS2
AS1
E-BGP
border router
interior router
Two types of routers
Border router, Interior router
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

6. Border Gateway Protocol Part II: I-BGP
AS2
AS1
E-BGP
I-BGP
border router
interior router
Two types of routers
Border router, Interior router
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

7. BGP Operations (Simplified)
Establish session on
TCP port 179
AS1
BGP session
Exchange all
active routes
AS2
While connection
is ALIVE exchange
route UPDATE messages
Exchange incremental
updates
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

8. Destination IP address prefix + attributes values
BGP Update Messages
Update : Announcing new routes or withdrawing previously announced routes.
Update =
Destination IP address prefix + attributes values
(e.g. a routing path)
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

9. Part I: E-BGP, Share connectivity information across ASs
you can reach net
prefix A via addr and
the path is “AS2”
AS2
AS1
E-BGP
addr A R1
BGP table at R1:
dest network prefix AS path next hop
A AS addr
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

10. Part II: I-BGP, Carrying Info within an AS
E-BGP update
I-BGP used to disseminate learned routes to all routers in AS
I-BGP updates
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

11. Part II: I-BGP, Carrying Info within an AS
you can reach net
B via addr1 and the
path is “AS3”
you can reach net
B via addr2 and
the path is “AS3 AS2”
E-BGP
E-BGP
you can reach net
B via addr1 and the
path is “AS3”
addr2 A
AS1
AS2
addr1
AS3 B
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

12. Attributes are Used to Select Best Routes
/24
pick me!
/24
pick me!
/24
pick me!
Given multiple
routes to the same
prefix, a BGP speaker
must pick at most
one best route
/24
pick me!
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

13. Example: Multiple AS Paths
128.2/16
AS701
AS7018
128.2/16
9 701
128.2/16
AS1239
AS73
Default choice: Pick shortest path
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

14. Shorter Doesn’t Always Mean better
Is path 4 1 better
than path 3 2 1?
AS 4
AS 3
AS 2
AS 1
AS can use custom policies other than shortest path
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

15. Benefits of BGP Design Path Vector style routing Advantages:
Distance vector algorithm with extra information
For each route, store the complete path (ASs)
Advantages:
can make policy choices (choose among many possible learned paths) based on set of ASs in path
can easily avoid loops
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

16. Announcing and Choosing Routes
BGP may learn many different paths for a destination network
Learns only reachability information, no performance metrics
Not about optimizing anything
All about policy (business and politics)
What a BGP speaker announces or not announces to a neighbor determines what routes may get used by that neighbor
Router chooses among paths based on policy
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

17. Nontransit vs. Transit ASes
ISP 2
Internet Service
providers (often)
are transit
networks
ISP 1
NET A
Nontransit AS
might be a corporate
or campus network.
Traffic NEVER
flows from ISP 1
through NET A to ISP 2
(At least not intentionally!)
IP traffic
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

18. Selective Transit NET A NET B NET C NET A provides transit
between NET B and NET C
and between NET D
and NET C
NET A DOES NOT
provide transit
Between NET D
and NET B
NET A
IP traffic
NET D
Most transit networks transit in a selective manner…
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

19. Customers and Providers
IP traffic
Customer pays provider for access to the Internet
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

20. The Peering Relationship
customer
provider
Peers provide transit between
their respective customers
Peers do not provide transit
between peers
Peers (often) do not exchange $$$
traffic
allowed
traffic NOT
allowed
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

21. Peering Provides Shortcuts
customer
provider
Peering also allows connectivity between
the customers of “Tier 1” providers.
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

22. Import Routes provider route customer route peer route ISP route From
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

23. Export Routes provider route peer route customer route ISP route To
From
provider To
peer To
peer To
customer To
customer
filters
block
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

24. How can routes be marked as “provider”, “peer”, “customer”, “isp” ?
Use “Community Attribute” in route announcement
Used for signaling
within and between
ASs
A community attribute is 32 bits
By convention,
first 16 bits is
ASN indicating
who is giving it
an interpretation
community
number
Very flexible
BECAUSE it
has no predefined
meaning
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

25. BGP Issues BGP designed for policy not performance
Susceptible to router misconfiguration
Blackholes: announce a route you cannot reach
Slow convergence time
Rate limiting and route flap dampening
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

26. Combining IGP and BGP BGP Process IGP Process OS kernel
BGP Routing tables
IGP Process
IGP Routing tables
OS kernel
Forwarding Table Manager
Forwarding Table
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

27. Combine BGP and IGP Tables to Create Forwarding Table
I-BGP: /16
Next Hop =
/16
E-BGP
IGP
/30
destination
next hop
AS 1
AS 2
/30
Forwarding Table
/16
destination
next hop
/30
BGP
/16
destination
next hop +
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

28. Life Cycle of a Packet in the Internet
Address Resolution Protocol (ARP)
On the same subnet, need to map IP address to MAC (e.g. Ethernet) address
Host and router have ARP cache to store the IP-MAC pairs
In case of no match in ARP cache, broadcast an ARP request with the IP address in question and the device with the IP address will reply with its MAC address
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University

29. Life Cycle of a Packet in the Internet
For each hop in the network, do the following steps:
Decapsulate the Ethernet frame to get the IP header (except no need to do this at the source)
Check routing table by the destination IP address, get the next-hop IP address and the network interface
Learn the MAC address of the next hop (look up in ARP cache or broadcast an ARP request)
Encapsulate the IP packet into an Ethernet frame with the destination MAC address
Send the Ethernet frame out from the next-hop network interface
T. S. Eugene Ng eugeneng at cs.rice.edu Rice University