ELEC3506

RoutingLecture 6 PDF15 min

Autonomous systems and routing protocol types

What an autonomous system is, the difference between a routed protocol and a routing protocol, and the three families of interior routing algorithm this unit builds on.

By the end of this page you should be able to

  • Define an autonomous system and say what a gateway router does
  • Distinguish a routed protocol from a routing protocol
  • Explain the difference between an IGP and an EGP
  • Name the three interior gateway protocol types and one example of each
  • List the routing metrics the lecture names and what each one measures

The idea

Not every router in the world runs the same routing software or answers to the same administrator. A university’s network and its ISP’s network are both collections of routers, but they are managed separately and may run completely different routing algorithms internally. Routing has to work both within one such collection and between separate ones, and those two situations turn out to need different tools.

How it works

Autonomous systems

Routers belonging to one organisation are said to belong to one autonomous system (AS). Routers within the same AS run the same routing algorithm — a university’s AS and an ISP’s AS do not have to agree with each other on how routing decisions are made internally.

The routing algorithm run within one AS is called an intra autonomous system routing protocol. Routers that connect different ASes together are called gateway routers, and the routing algorithm they use to talk across that boundary is called an inter-autonomous system routing protocol.

How it works

Routed protocols vs routing protocols

These two terms sound alike and are frequently mixed up.

  • Routed protocols are protocols that get routed over an internetwork — the payload being moved. The lecture’s examples: Internet Protocol (IP), AppleTalk, Novell NetWare.
  • Routing protocols are protocols that implement routing algorithms — the mechanism deciding how the routed protocol gets forwarded.
Routed protocolRouting protocol
What it isThe thing being carried across the internetworkThe mechanism that decides how to forward it
Examples (lecture)IP, AppleTalk, Novell NetWareRIP, OSPF, BGP
A routing protocol decides how a routed protocol travels. IP does not decide its own path; RIP, OSPF and BGP do.

How it works

IGP vs EGP

Routing protocols split into two families based on which side of an AS boundary they operate on:

  • Interior Gateway Protocol (IGP) — an intra-autonomous-system routing protocol. The lecture’s examples: RIP (Routing Information Protocol) and OSPF (Open Shortest Path First).
  • Exterior Gateway Protocol (EGP) — an inter-autonomous-system routing protocol. The lecture’s example: BGP (Border Gateway Protocol).

The lecture is explicit that RIP and OSPF are what this course covers “which are the basic of these protocols” — they are the two IGPs worked through in detail across the rest of this module, with BGP covered separately as the EGP.

How it works

The three interior gateway protocol types

There are three types of interior gateway protocol:

  1. Distance vector based. Uses number of hops to decide the routing path — the distance vector algorithm. Example: RIP.
  2. Link state based. Recreates the exact topology of the entire network: each router, its neighbours, its attached links, and so on. Uses Dijkstra’s algorithm. Example: OSPF.
  3. Balanced hybrid. Exchanges more information than a distance vector protocol, but does not require the full topology and is not as computationally intensive as link state. Example: Enhanced Interior Gateway Routing Protocol (EIGRP).

How it works

Commonly used routing metrics

A metric is what a routing protocol uses to compare candidate paths. The lecture names eight:

  • Hop count — the number of routers a packet passes through.
  • Ticks — delay on a data link, measured in IBM PC clock ticks (approximately 55 ms each).
  • Cost — an arbitrary value, usually based on bandwidth/throughput, dollar expense, or another measurement, assigned by a network administrator.
  • Bandwidth — the data capacity of a link.
  • Delay — the length of time required to move a packet from source to destination.
  • Load — the amount of activity on a network resource, such as a router or a link.
  • Reliability — usually refers to the bit-error-rate of each network link.
  • MTU — Maximum Transmission Unit, the maximum frame length in octets acceptable to every link on the path.

Check yourself

  • An AS is a group of routers under one organisation, running one routing algorithm; gateway routers connect one AS to another.
  • Routed protocols are carried (IP); routing protocols decide how (RIP, OSPF, BGP).
  • IGP = intra-AS (RIP, OSPF). EGP = inter-AS (BGP).
  • The three interior protocol families: distance vector (hop count, RIP), link state (full topology, Dijkstra, OSPF), balanced hybrid (EIGRP).
  • Eight named metrics: hop count, ticks, cost, bandwidth, delay, load, reliability, MTU.

In the exam

  • Do not swap routed and routing protocol. IP is routed; RIP, OSPF and BGP are routing protocols. This pairing is a favourite short-answer or matching question.
  • Match each IGP type to its example and its defining mechanism: distance vector → hop count → RIP; link state → full topology + Dijkstra → OSPF; balanced hybrid → EIGRP.
  • IGP vs EGP is about which side of an AS boundary the protocol runs on, not about which protocol is “better.”
  • Know all eight metrics by name. A question can ask you to name the metric or to say what it measures, in either direction.
  • Ticks has a specific value attached — approximately 55 ms per tick — which makes it an easy fact to be tested on directly.

Check yourself

  1. Two routers belong to the same organisation and run the same routing algorithm as each other. What does the lecture call the group they belong to?
  2. Which pair correctly matches a routed protocol with a routing protocol?
  3. This unit covers RIP and OSPF as the basic examples of interior gateway protocols. What is BGP an example of?
  4. Which routing metric does the lecture describe as measured in units of approximately 55 ms, based on IBM PC clock ticks?
  5. A distance vector protocol and a link state protocol both solve intra-AS routing. What is the key difference the lecture gives?