Network Layer: Control Plane Flashcards
Introduction
Network-Layer Functions
Forwarding: move pckts from router input to output, local action, data-plane
Routing: determine route from src-dest, global action, control-plane
Introduction
Two approaches to structuring network control plan
- Per-router control (traditional)
- Logically centralized control (software defined networking)
Introduction
Per-Router Control Plane
- Traditional
- Individual routing algorithm components in each router
- Router makes its own routing table
Introduction
Software-Defined Networking (SDN) Control Plane
- Remote controller computes, installs forwarding tables in routers
Introduction
Routing Protocol Goal
Determine “good” paths from sending hosts to recieving host through network of routers
Introduction
Path
- Sequence of routers
- Packets traverse from given initial src host to final dest host
Introduction
“Good” Path
Least cost, fastest, least congested
Introduction
Routing Alogirithm Classification
- Global
- Static
- Dynamic
- Decentralized
Introduction
Routing Algorithm Classification: Global
- All routers have complete topology, link cost info
- Uses link state algorithms
Introduction
Routing Algorithm Classification: Static
Route changes slowly over time
Introduction
Routing Algorithm Classification: Dynamic
- Routes change more quickly
- Periodic updates or in reponse to link cost changes
Introduction
Routing Algorithm Classification: Decentralized
- Iterative process of computation, exchange of info with neighbors
- Routers intilially only know link cost to attached neighbors
- uses distance vector algorithms
Routing Protocols: Link State
Dijkstra’s Link-State Routing Algorithm
All link costs known by all nodes
Centralized: network topology, link cost known to all nodes, accomplished via “link state broadcast*”
Computes least cost path frfrom one node to all other nodes
Iterative: after k iterations, know least cost path to k destinations
Routing Protocols: Link State
Dijkstra’s Example
Routing Protocols: Link State
Djikstra’s Algorithm Complexity
- n Nodes
- Each of n iterations, need to check all nodes, w, not in N
- n(n+1)/2 comparisons: O(n^2) complexity