N11. IP Routing Flashcards
What is CIDR?
Classless Inter-Domain Routing
- more efficient address assignment
- super-netting possible (routing info aggregation)
- variable length sub-net mask
What is the difference between fixed topology and dynamic topology in routing?
Fixed topology - what is supposed to be connected to what
Dynamic topology - what is actually connected to what
What are the possible metrics for distance to destination?
- hop count
- link throughput
- link delays
- link error rates
- link financial costs
What routing policies and constraints are there?
Administrative
- commercial agreements to carry traffic from certain sources
- priority routes for some traffic
Security
- avoid untrusted networks
Quality of Service (QoS)
- route certain traffic types via suitable links
Real cost (financial)
- only use high-cost links when all else fails
How are routing protocols used on the internet?
Distributed routing
- no centralised operation and control of protocol
- routing policy may be managed centrally
Two main elements:
- messages: routing updates
- algorithm: find routes
(combination of these provides behaviour of protocol)
Describe the original ARPANET routing algorithm
Advanced Research Projects Agency Network
- Distance-vector (DV) (Bellman-Ford)
- Use queue length as metric (diverts traffic away from congestion)
- high capacity links not specially favoured
- queue lengths are not stable
- oscillations: route flapping
Describe link-state routing algorithms
Each node - addresses cost of local links - distributes info to all nodes - finds lowest cost path to all other nodes Dijkstra's SP algorithm - shortest path tree to all other nodes
(note bellman-ford less efficient than Dijkstra but can handle negative weights on edges)
Features of hierarchical routing
Area - network cloud - single routing protocol within area Connectivity between areas - hierarchy of routers Routing at levels - allows routing info to be aggregated
What are Autonomous Systems?
Internet connectivity is partitioned along administrative boundaries (Autonomous System boundaries)
AS identified by AS numbers
Between ASs - need a common routing protocol (e.g. BGP), policy-based routing
Describe the large scale structure of the internet
Traffic types: - local: intra -AS - transit: inter-AS AS: - stub AS e.g. site network - multi-homed AS e.g. ISP - transit AS e.g. backbone provider Internet: - collection of interconnected stub, multi-homed and transit ASs
What is BGP?
Border Gateway Protocol - inter-AS protocol - between border routers - reachability info - no routing metric - path vector (full AS path) - policy-based routing possible Updates less frequent than intra-AS protocol. Info exchanged with neighbours.
Give some examples of network exploration tools
- ping: check if a remote host is up
- traceroute: find the network path to the remote host
- whois services: IP address registration records
Give the equation for measurement of links in terms of transmission delay.
Tx = b/r
Tx = transmission delay (time taken to put bits on wire) b = number of bits r = data rate (bits per second, b/s)
Give the equation for measurement of links in terms of propagation delay.
Tp = d/Ss
Tp = propagation delay of signal (time taken for a signal to traverse link) d = distance of link (metres) Ss = speed of signal (m/s)
Give the equation for measurement of links in terms of one-way delay.
Td = Tx + Tp
Td = one-way delay on a link (path) (time taken for a signal to traverse a link)
List measurements of paths
- approximation delay for end-to-end path (rather than individual link)
- use ping to measure one-way delay of whole end-to-end path
- use traceroute to see the individual links of the whole end-to-end path
Describe routing subversion (security)
C communicates with S
X wants to capture traffic sent by S to C
X pretends to be a router (offers better route)
Now routes via X and C&S are unaware
Route hijacking
Give examples of routing subversion
Man-in-the-Middle
- X can inspect traffic or perturb, disrupt or modify session
Black hole
- X can simply drop packets (DoS attack)
Replay
- X can keep copies of packets and replay a session later
Similar attack vectors possible at application layer
What is the purpose of intrusion detection systems (IDS)
Detect security breaches based on observations of systems behaviour.
Based on network and systems monitoring.
Establish baseline and detect anomalous behaviour
Legal issues e.g. GDPR
