Chapter 4 Flashcards
Link-State Routing Protocols and Areas
- Link-state routing protocols like OSPF and IS-IS partition routing domains into areas.
- Areas are groups of routers that exchange link-state information.
- There is a special area known as the backbone area (Area 0), which connects all other areas.
- Example: Routers R1, R2, and R3 can be members of both the backbone area and nonbackbone areas (e.g., Area 1, Area 2).
- A router that is a member of both the backbone area and a nonbackbone area is called an Area Border Router (ABR).
Routing Within and Between Areas
- Within an area, all routers exchange link-state advertisements to develop a complete map of the area.
- Link-state advertisements from non-ABR routers do not leave the area, which improves scalability.
- To route between nonbackbone areas, packets travel through the backbone area (Area 0).
- Area border routers summarize routing information from one area and advertise it into the backbone area.
- Routers in the backbone area then summarize and advertise this information into nonbackbone areas.
Tradeoff Between Scalability and Routing Optimality
- Dividing a domain into areas trades off routing optimality for scalability.
- All packets between nonbackbone areas must travel through the backbone area, even if a shorter path exists.
- This design decision ensures scalability by limiting the number of routers that need to exchange routing information.
- Example: Even if R4 and R5 are directly connected, packets cannot flow between them if they are in different nonbackbone areas.
Virtual Links and Routing Flexibility
- Virtual links allow network administrators to connect routers not directly connected to the backbone to routers in Area 0.
- Example: A virtual link can connect R8 in Area 1 to R1 in Area 0, making R8 part of the backbone.
- Virtual links improve routing optimality by allowing non-ABR routers to participate in backbone routing.
- The cost of the virtual link is determined by routing information exchanged in the respective nonbackbone area.
Autonomous Systems (AS) and Routing
- Autonomous systems (AS) provide hierarchical aggregation of routing information in large networks like the Internet.
- Routing in AS is divided into:
- Intradomain routing (within an AS)
- Interdomain routing (between ASs or routing domains)
- AS model allows each AS to use its own intradomain routing protocols and policies independently.
- Interdomain routing involves sharing reachability information between ASs.
Challenges in Interdomain Routing
- Interdomain routing requires each AS to define its own routing policies.
- Example: Policies might prefer one path (AS X) over another (AS Y), avoid carrying traffic between specific AS pairs, and prioritize certain providers over others.
- Complex policies need to be supported without relying on other ASs, due to competitive and confidential reasons.
- Interdomain routing protocols must handle misconfigurations and malicious behaviors from other ASs.
History and Evolution of Interdomain Routing Protocols
- Exterior Gateway Protocol (EGP) was the first interdomain routing protocol but had limitations with the Internet’s evolving topology.
- Border Gateway Protocol (BGP), specifically BGP-4, replaced EGP and supports a graph model for interconnecting ASs.
- BGP can accommodate non-tree-structured internetworks, such as multiprovider networks.
- Today’s Internet is a complex network of interconnected ASs, mainly operated by private ISPs.
Types of Autonomous Systems (AS)
- Stub AS: Connects to only one other AS and carries local traffic only.
- Multihomed AS: Connects to multiple ASs but refuses to carry transit traffic.
- Transit AS: Connects to multiple ASs and carries both transit and local traffic.
- AS types influence routing decisions and policies within the network.
Goals and Challenges in Interdomain Routing
- Interdomain routing aims to find loop-free paths compliant with AS policies.
- Challenges include scale (handling 700,000+ prefixes), diverse routing policies, and trust between ASs.
- Interdomain routing focuses on reachability rather than optimizing path costs across multiple ASs.
- The autonomous nature of ASs complicates path cost calculations due to varying metrics and policies.
Basics of BGP
- Each AS has one or more border routers responsible for forwarding packets between autonomous systems (AS).
- Border routers may also function as BGP speakers, which communicate routing information with other BGP speakers in different ASs.
- BGP is not a distance-vector or link-state protocol; it advertises complete paths as enumerated AS sequences to reach specific networks.
- This path-vector approach is crucial for making policy decisions and preventing routing loops in complex AS networks.
BGP Path Advertisement
- BGP speakers advertise reachability information for networks assigned to their customers.
- Example: AS 2 advertises networks 128.96, 192.4.153, 192.4.32, and 192.4.3 as reachable directly from AS 2.
- Backbone networks then advertise paths to these networks, indicating the sequence of ASs to reach them, such as (AS 1, AS 2) and (AS 1, AS 3).
- BGP’s path enumeration prevents routing loops by detecting and avoiding paths that lead back to the originating AS.
AS Numbers and Loop Prevention
- AS numbers in BGP must be unique to prevent routing loops.
- AS numbers are 32 bits long and centrally assigned to ensure uniqueness.
- Unique AS numbers are critical for BGP speakers to correctly identify and avoid routing loops in the AS path advertisements.
BGP Route Selection and Advertisement
- A BGP speaker selects the best route to a destination based on its local policies.
- BGP speakers are not obliged to advertise all routes; they can choose not to advertise routes to certain prefixes, implementing policies such as not providing transit.
- Route cancellations in BGP are achieved through withdrawn route messages, a form of negative advertisement.
BGP Communication and Reliability
- BGP runs over TCP for reliable communication.
- TCP ensures that once information is sent from one BGP speaker to another, it does not need to be retransmitted unless changes occur.
- BGP speakers exchange keepalive messages to confirm connectivity and the validity of routes; absence of keepalives indicates route invalidity.
Common AS Relationships and Policies
- Autonomous Systems (ASs) have different relationships reflecting common connectivity needs and business models.
- Three primary relationships are:
- Provider-Customer
- Customer-Provider
- Peer
- These policies ensure that traffic is routed efficiently and economically, aligning with the business interests of each AS.
Provider-Customer Relationship
- Description: Providers connect their customers to the Internet.
-
Policy:
- Advertise all routes to its customer.
- Advertise routes learned from its customer to everyone else.
- Function: Ensures connectivity for customers and helps in routing traffic efficiently.
Customer-Provider Relationship
- Description: Customers receive traffic from and send traffic to the Internet through their provider.
-
Policy:
- Advertise its own prefixes and routes learned from its customers to its provider.
- Advertise routes learned from its provider to its customers.
- Do not advertise routes between providers.
- Function: Provides customers with access to the Internet and ensures efficient routing of traffic.
Peer Relationship
- Description: Symmetrical peering between autonomous systems that view themselves as equals.
-
Policy:
- Advertise routes learned from its customers to its peer.
- Advertise routes learned from its peer to its customers.
- Do not advertise routes from its peer to any provider or vice versa.
- Function: Minimizes costs by exchanging traffic between peers and optimizing routing.
Summary of AS Relationships
-
Provider-Customer Relationship:
- Providers connect customers to the Internet.
- Policy: Advertise all routes to customers; advertise customer routes to everyone else.
-
Customer-Provider Relationship:
- Customers receive and send traffic through their provider.
- Policy: Advertise own prefixes and customer routes to provider; advertise provider routes to customers; do not advertise routes between providers.
-
Peer Relationship:
- Peers are equals and exchange traffic without payment.
- Policy: Advertise customer routes to peers; advertise peer routes to customers; do not advertise routes to or from other providers.
Hierarchical Structure and Tier-1 Providers
- Hierarchical structure:
- Stub networks at the bottom, customers of providers.
- Providers with other providers as customers higher up.
- Tier-1 providers at the top, having customers and peers but not being customers of other providers.
- Tier-1 providers are critical for global Internet connectivity, as they provide extensive reach without the need for transit from other providers.
Business Policies in AS Relationships
- AS relationships are based on business needs and traffic exchange efficiency.
- Provider-Customer relationship focuses on connectivity for customers.
- Customer-Provider relationship ensures customers have access to and from the Internet.
- Peering relationships are symmetrical and minimize costs by exchanging traffic between peers.
Default Route Injection in Stub AS
- Description: In a stub AS that connects to other ASes at a single point, the border router is the only exit point for routes outside the AS.
-
Process:
- Border router injects a default route into the intradomain routing protocol.
- This default route matches any destination that is not explicitly advertised within the AS.
- Function: Provides a simple and effective way for a stub AS to reach networks outside its domain.
Specific Route Injection by Border Routers
- Description: Border routers inject specific routes learned from external ASes into the intradomain routing protocol.
-
Process:
- Border router learns specific network prefixes (e.g., 192.4.54/24) from BGP.
- Injects these specific routes into the intradomain protocol with associated costs.
- Function: Allows routers within the AS to learn how to reach specific external network prefixes efficiently.
Interior BGP (iBGP) in Backbone Networks
- Description: In backbone networks, where a large number of prefixes are learned from BGP, iBGP is used to distribute this information internally.
-
Process:
- iBGP redistributes BGP-learned routes to all routers within the AS.
- Maintains a mesh of iBGP sessions among all routers in the AS.
- Function: Enables all routers in the AS to determine the best exit (border) router for reaching any external prefix.
Integrating Interdomain and Intradomain Routing
- Description: Combining BGP (interdomain) with an intradomain protocol (IGP) to provide complete routing information within an AS.
-
Process:
- Border routers exchange BGP information (eBGP) with external ASes.
- iBGP is used to distribute this information to all routers within the AS.
- Each router also maintains an IGP to learn internal routes within the AS.
- Routers use both BGP-learned information and IGP-learned routes to determine the best path to each external prefix.
- Function: Provides comprehensive routing capabilities, ensuring efficient routing of packets both within and outside the AS.
Introduction to IPv6
- Purpose: Address limitations of IPv4, such as address exhaustion.
- Design Goals: Simplify address assignment, improve routing efficiency, enhance security, and support new services.
Address Format
- Address Length: 128 bits (compared to 32 bits in IPv4).
- Representation: Eight groups of four hexadecimal digits separated by colons (e.g., 2001:0db8:85a3:0000:0000:8a2e:0370:7334).
- Zero Compression: Omitting leading zeros and replacing consecutive zeros with “::” (e.g., 2001:db8::8a2e:370:7334).
Address Types
-
Unicast: Identifies a single interface. Types include:
- Global Unicast: Globally unique addresses routable on the internet.
- Link-Local: Used for communication within a single link (e.g., FE80::/10).
- Unique Local: Used within a site or organization (e.g., FC00::/7).
- Multicast: Identifies multiple interfaces, typically in the same group.
- Anycast: Identifies multiple interfaces, but packets are delivered to the nearest one.
Header Format
- Simplified Header: Compared to IPv4, the IPv6 header is streamlined to improve processing efficiency.
-
Fields:
- Version: Indicates the protocol version (6 for IPv6).
- Traffic Class: Used for traffic prioritization.
- Flow Label: Identifies flows of packets for special handling.
- Payload Length: Length of the payload following the header.
- Next Header: Indicates the type of the next header (similar to the Protocol field in IPv4).
- Hop Limit: Replaces the Time to Live (TTL) field in IPv4.
- Source and Destination Addresses: 128-bit addresses of the sender and receiver.
Extension Headers
- Purpose: Provide additional functionalities and options.
-
Types:
- Hop-by-Hop Options: Processed by every node along the path.
- Destination Options: Processed by the destination node.
- Routing Header: Specifies a list of intermediate nodes to be visited.
- Fragment Header: Supports packet fragmentation.
- Authentication Header (AH): Provides packet integrity and authentication.
- Encapsulating Security Payload (ESP): Provides confidentiality, data integrity, and authentication.
IPv6 Address Autoconfiguration
-
Stateless Address Autoconfiguration (SLAAC):
- Process: Hosts generate their own addresses using a combination of locally available information and router advertisements.
- Router Advertisements: Routers periodically send advertisements to announce their presence and provide network information.
-
Stateful Configuration (DHCPv6):
- Purpose: Provides additional configuration options such as DNS server addresses.
- Process: Similar to DHCP in IPv4, where a server assigns addresses and other network information to hosts.
Transition Mechanisms
- Dual Stack: Allows IPv4 and IPv6 to coexist on the same network infrastructure.
- Tunneling: Encapsulates IPv6 packets within IPv4 packets to traverse IPv4 networks.
- Translation: Converts IPv6 packets to IPv4 packets and vice versa.