4: The Network Layer

Question 1

What is a network service model?

Answer 1

A list of criteria and desired properties of network functionality

Question 2

What is switching via a bus?

Answer 2

Packets directly transferred from input to output ports over a shared bus, which allows 1 packet through at a time. High traffic can cause buffering in the bus.

Question 3

What is priority scheduling?

Answer 3

Assigning priorities to each packet, and sending the packet with the highest priority first.

Question 4

What is generalised forwarding?

Answer 4

A variant of decentralised switching in which packets can be forwarded according to any combination of their header values.

Question 5

Whar is the control plane?

Answer 5

The section of a router’s architecture that provides the logic determining the topology of the whole network.

Question 6

What is a subnet?

Answer 6

A subset of a network divided by an interface on a router. The subnet part of an IP address is the high order bits; the low order bits distinguish hosts within a subnet.

Question 7

What is software-defined networking?

Answer 7

Remote servers contain the control plane logic for routers.

Question 8

What is the ICMP protocol?

Answer 8

The Internet Control Message Protocol (ICMP) is a network-level protocol that seeks to assist network functionality be reporting errors and operational information, including to routers.

Question 9

What is decentrailised switching?

Answer 9

Using header values in inbound segments to forward them to an output port as per the forwarding table in the memory of the port they came in on.

Question 10

What are switching fabrics?

Answer 10

The topology of a network using switches that defines how the switches are placed and connected, and how data are switched in the network.

Question 11

What is a forwarding table?

Answer 11

A mapping of input to output interfaces used in forwarding in a router.

Question 12

What is longest prefix matching?

Answer 12

Using the longest address prefix that matches destination address when looking for forwarding table entry for a given destination address, to minimise the number of hops to the destination after the next one.

Question 13

What is FIFO (first in first out) scheduling?

Answer 13

Sending out packets in a buffer/queue in the same order they arrive.

Question 14

What is a network bus?

Answer 14

A common linear half-duplex link to which all nodes are connected.

Question 15

What is switching via memory?

Answer 15

When a traditional computer acts as a switch, taking a packet into memory, processing it to find the associated output port, then sending it there.

Question 16

What is an interconnection network?

Answer 16

High-speed computer networks with ndoes specialised to memory and processing.

Question 17

What is switching via interconnection network?

Answer 17

When you use the distributed processors in interconnection networks to process the packets being switched in parallel to optimise performance and overcome bus bandwidth limitations.

Question 18

What is the concept of IP fragmentation and reassembly?

Answer 18

Networks have MTUs - the largest possible link-level frames - so if these are exceeded then IP datagrams are fragmented into smaller ones and reassembled at the final destination.

Question 19

What is Head-of-the-Line (HOL) blocking?

Answer 19

When a slow datagram at the front of the queue at an input port in a switch prevents others in the queue from moving forward.

Question 20

What is buffering?

Answer 20

When high traffic means packets are entered into a queue and processed one at a time, in the order they arrived.

Question 21

What is a scheduling discipline?

Answer 21

Algorithms used for distributing resources among parties which simultaneously and asynchronously request them.

Question 22

What is the NAT protocol?

Answer 22

Network address translation is a method of using one IP address to represent a private subnet to relieve IPv4 address exhaustion.

Question 23

What is output port queueing?

Answer 23

When you buffer multiple packets contending for access to the same output port on a switch, and the arrival rate of packets exceeds the output line speed of the port.

Question 24

What is Round Robin (RR) scheduling?

Answer 24

Sort packets into groups called classes and cyclically output a packet from each class, internally ordered chronologically.

Question 25

What is IP?

Answer 25

Internet Protocol is a network-layer protocol that handles the routing and addressing of packets between nodes on a computer network.

Question 26

What is IPv6?

Answer 26

The sixth and most recent version of the Internet Protocol.

Question 27

What is a link’s MTU?

Answer 27

max transfer size = the largest possible amount of data in a single link-level frame on a network

Question 28

How are fragmented IP packets reassembled?

Answer 28

With header flags that identify the original datagram, length and offset from the start of the original datagram.

Question 29

What is an IP address?

Answer 29

A 32-bit identifier for a host on the Internet that acts an interface for routers.

Question 30

What is an interface?

Answer 30

A connection between a host/router and a physical link.

Question 31

How do you identify a subnet?

Answer 31

Separate all LANs connected to different interfaces on routers.

Question 32

What is CIDR?

Answer 32

Classless InterDomain Routing is a method for identifying subents and hosts in IP addresses; the first portion of the IP address distinguishes subnets and the second hosts. A shorter subnet part leaves more bits to identify hosts, so subnet can be larger. But since less for subnet identification can have fewer of these larger networks.

Question 33

What is DHCP?

Answer 33

Dynamic Host Configuration Protocol is a network management protocol in which a server dynamically assigns IP addresses to hosts on the Internet.

Question 34

What is ICANN?

Answer 34

Internet Corporation for Assigned Names and Numbers is the body that allocates Internet domains and addresses and manages DNS.

Question 35

Why is NAT used?

Answer 35

To relieve IP address exhaustion (in v4) and so that Internet access in a network is abstracted to one access point: nodes on the subnet in question aren’t concerned about it.

Question 36

How is NAT implemented?

Answer 36

A NAT translation table is used to map incoming datagrams to devices by port number, and outgoing datagrams replace their source IP address from that of the device to that of the router operating NAT.

Question 37

How does IPv6 differ from IPv4?

Answer 37

It increases addresses to 128 bits to fix exhaustion. Further changes to the header to speed up processing and forwarding and allow QoS measures to reduce jitter, latency, and loss. Also options flags, no checksum, and a new version of ICMP (v6).

Question 38

What is the format of the IPV6 datagram?

Answer 38

The header contains priority, flow label, next header, TTL, payload length, source address, and dest address. Then payload.

Question 39

What are IPv6 options?

Answer 39

Additional non-prescriptive flags that alter how IPv6 datagrams are processed at routers or their destination host.

Question 40

What is ICMPv6?

Answer 40

The new version of ICMP in IPv6 which includes new message types and multicast group management functions.

Question 41

What is tunnelling in IPv6?

Answer 41

When you split up an IPv6 datagram and send it as the payload of an IPv4 datagram when going through IPv4 routers.

Question 42

Why aren’t people all taking up IPv6?

Answer 42

Overhead to upgrade, systems been using IPv4 for entire Intenet history, ignorance of issues, timely and costly to upgrade all low-level Internet infrastructure.