BGP Basics and Network Management Quiz

Computer Networks: The Network Layer

• This presentation discusses the network layer and its control plane.
• The outline covers various topics such as introductions, routing algorithms, intra-AS routing (OSPF), inter-AS routing (BGP), SDN control planes, ICMP protocol, and network management protocols.

Routing Among the ISPs: BGP

• BGP (Border Gateway Protocol) is a decentralized, asynchronous protocol used for inter-AS routing.
• BGP routes packets to CIDRized prefixes (representing subnets or collections of subnets), not specific destination addresses.
• A router's forwarding table has entries in the format (x, I), where x is a prefix (e.g., 138.16.68/22) and I is an interface number.

The Role of BGP

• BGP allows routers to obtain prefix reachability information from neighboring autonomous systems (ASs).
• Each subnet advertises its existence to the rest of the internet within each associated autonomous system.
• BGP utilizes a route-selection procedure to determine the "best" routes, considering both reachability and policy.
• BGP routing tables can contain over half a million routes.

Advertising BGP Route Information

• Each router in an autonomous system (AS) is either a gateway router or an internal router.
• A gateway router connects to one or more routers in other ASs and directly interacts with other Autonomous Systems.
• An internal router connects only with hosts and routers within its own AS.
• ASes exchange routing information via routers.

Advertising BGP Route Information Example

• In the displayed network example, AS3 advertises subnet x to AS2.
• AS2 advertises the path AS2-AS3-x to AS1.
• Each AS simultaneously learns about the existence of a path and the existence of the path itself to destination x.
• Routers, not autonomous systems, are the agents acting upon routing information.

Advertising BGP Route Information (TCP Connections)

• BGP routing information exchange happens over semi-permanent TCP connections with port 179.
• External BGP (eBGP) connections span across different autonomous systems.
• Internal BGP (iBGP) connections exist between routers within the same autonomous system.
• Typically, there is one eBGP connection per direct link between gateway routers in different ASes.
• There can be multiple iBGP connections between routers within an AS that don't always correspond to physical links.

Determining the Best Routes

• Routers choose the best route among multiple available paths based on attributes like AS-PATH and NEXT-HOP. The AS-PATH attribute records the sequence of ASes in the route, and the NEXT-HOP attribute identifies the IP address of the next router along the chosen path.
• This prioritizes paths with shorter AS-PATHs, but also considers policy and local preference.

Hot Potato Routing

• The hot-potato routing algorithm prioritizes sending network traffic out of an AS as quickly as possible, without considering costs or intermediate network segments outside the specific AS.
• Two routers in the same AS can use different paths to reach the same destination prefix.

IP-Anycast

• IP-anycast allows replicating content across multiple servers in different countries.
• This allows users to access the nearest server efficiently.
• BGP is used to create paths to multiple copies of the server.

Routing Policy

• Routing policies are used to determine routing within and between autonomous systems.
• Policies favor traffic going to a direct customer through the shortest path, minimizing cost and routing overhead.
• The policy decides which route from multiple possible ones is to be used for traffic with a common destination prefix.

Access ISP Policy

• All the traffic entering or leaving an access network, must be destined to or from the access network where the traffic originated.
  • Selective route advertisement policy is used to only advertise the paths for given access network, without unnecessary routing complexity.

Backbone/Provider ISP Policy

• ISPs commonly use the rule that traffic routed across their backbone network must have a destination or source within a customer network of that ISP.
• Peerings between ISPs are negotiated privately and have confidential agreements.

Why Different Routing Protocols?

• Policy dictates routing decisions between Autonomous Systems (ASs).
• Scaling capability to manage large networks is a key aspect when selecting routing protocols.
• Performance of the chosen routes is a lesser concern between ASs, but it is important inside an AS where routing prioritizes high throughput and minimized end-to-end latency.

Obtaining Internet Presence

• Obtaining internet connectivity requires contracting with and connecting to an ISP.
• The ISP provides an IP address range.
• The customer (e.g., an organization) assigns these IP addresses within the range to devices (e.g., Web Servers, mail servers, DNS servers, and gateway router).
• The customer then contracts with an internet registrar to obtain a relevant domain name which in turn is registered to be associated with the provided IP addresses of its services.

ICMP: The Internet Control Message Protocol

• ICMP is used by systems (hosts or routers) to exchange network-layer information, with a most frequent use-case as an error reporting tool.
• ICMP messages are transported within IP datagrams.
• Format of an ICMP message, includes headers fields for Type and Code that describe the type and specific nuance of the message.

The Internet Control Message Protocol Example Messages

• Ping: the program sends ICMP message type 8, code 0.
• Source quench: ICMP message to force a host that overloaded a router to reduce its transmission rate. This is seldom (if ever) used.
• Traceroute: the program sends a series of UDP packets with incrementing Time-To-Live (TTL) values to determine paths, or routes, needed to reach a target.

Network Management

• Managing network components, including people and software tools in addition to hardware components such as routers, and switches.
• Objectives of network management include deploying, integrating, and coordinating network elements, as well as monitoring, evaluation, analyzing the performance of the operational network, identifying and responding to faults, and implementing changes to meet performance expectations defined by Service Level Agreements (SLAs).

Components of Network Management

• A managing server located in a central network operations center (NOC), to oversee, and execute, all operations pertaining to networked devices within the purview of that NOC.
• Managed devices include all network-connected components: hosts, routers, switches, and other hardware components.
• Network management agents are software processes running on the managed devices to communicate with the managing server.

Ways to Manage the Network

• Command line interface (CLI): a method to manage networks which is prone to errors, and difficult to automate or scale to larger networks.
• SNMP/MIB: a technique using Simple Network Management Protocol (SNMP) that can query and set values of network data via Management Information Base (MIB) objects.
• NETCONF/YANG: data modeling and communication language and protocol used to configure and manage devices remotely and uniformly.

Test your knowledge on Border Gateway Protocol (BGP) and its role in internet routing. This quiz covers key concepts, components of BGP, and aspects of network management. Challenge yourself with questions about routing tables, ICMP messages, and traceroute operations.