Data Communications And Networks Part 2

Front 1

Application layer

Back 1

Hold network applications and many related protocols (HTTP for requesting and/or transferring files, SMTP for email, FTP for transferring files between two end systems, SSL, etc.) Another key application-layer protocol, DNS (Domain Name System), works to give Internet end systems more user-friendly names.

Front 2

Transport layer

Back 2

Data transfer between processes. Deliver messages from the application layer to the other endpoint(s). Uses either the TCP (connection-focused) or UDP (bandwidth-focused) protocols.

Front 3

Network layer

Back 3

Relays datagrams, or network-layer packets, between hosts. Follows the lone IP protocol to ensure that end systems and routers work to receive datagrams. However, also follows many available routing protocols to decide transport paths between sources. Provides logical communication between hosts.

Front 4

Link layer

Back 4

Helps the network layer transport a packet from one node to the next. These nodes can include hosts or routers, which are where the link layer takes over from the network layer. These different links may follow different protocols (Ethernet, Wifi, etc.)

Front 5

Physical layer

Back 5

Within the link layer, the physical layer works to deliver each bit of data across the link. Depending on the link, different protocols might come into play.

Front 6

What are the two key network-layer functions?
(most important in datagram network)

Back 6

* Routing: determine route taken by packets from source to destination (like planning a trip from the source to the destination)

* Forwarding: move packets from router’s input to appropriate router output (like the process of getting through a single interchange)

Front 7

Some network architectures have a 3rd important function. What is it?
(this and the other two = the three most important in a virtual-circuit network)

Back 7

* Connection setup: before datagrams flow, two end hosts AND intervening routers establish a virtual connection

In network layer: between two hosts (may also include intervening routers)

In transport layer: between two processes

Front 8

Network service model

Back 8

service model for a "channel" transporting datagrams from sender to receiver

Front 9

Example services for individual datagrams

Example services for a flow of datagrams

Back 9

individual datagrams: guaranteed delivery, guaranteed delivery with less than 40 msec delay

flow:
In-order datagram delivery, guaranteed minimum bandwidth to flow, restrictions on changes in interpacket spacing

Front 10

Describe the forwarding tables for both datagram networks and virtual-circuit networks.

Back 10

Datagram forwarding tables use a routing algorithm and the packet’s destination address to decide the output link.

VC forwarding tables have four sets of data: incoming interface, incoming VC#, outgoing interface, and outgoing VC#. The first two determine the last two.

Front 11

Compare and contrast link-state and distance-vector routing algorithms.

Back 11

Link-state routing uses Djikstra’s algorithm to look at all paths and calculate the least-cost paths from a source to the destination. Distance-vector routing uses dynamic programming in the form of the Bellman-Ford Equation: each node periodically sends a distance vector estimate to each of its neighbors. The equation looks at these distances to determine the least-cost path. A malfunctioning router has a different effect in both algorithms. In link-state routing, an error (particularly a node that gives an incorrect link cost) remains with the node and doesn’t propagate. With distance-vector routing, however, the incorrect information would be passed to other nodes.

Front 12

Inter-autonomous-system routing
(Inter-AS protocol)

Back 12

The (multiple) administrators want to control how exactly traffic is routed and who routes through the network. Policy can even take precedence over performance.

Front 13

Intra-autonomous-system routing
(Intra-AS protocol)

Back 13

Only one administrator, more focused on performance

Front 14

When a host joins a multicast group, must it change its IP address to that of the multicast group it is joining?

Back 14

No, the host maintains its own IP address.

Front 15

Suppose that routers were subject to conditions that might cause them to fail fairly often. Would this argue in favor of a VC or a datagram architecture? Why?

Back 15

Datagram, because it has “elastic” service and superior flexibility that would be more accommodating to router failure. Datagram architectures provide a connectionless service.

Front 16

Suppose that a source node and a destination require that a fixed amount of capacity always be available at all routers on the path between the source and destination node, for the exclusive use of traffic flowing between this source and destination node. Would this argue in favor of a VC or a datagram architecture? Why?

Back 16

VC, because it allows stricter requirements than datagram architecture. VC architectures provide a connection-focused service that allows the fixed capacity described above.

Front 17

Suppose that the links and routers in the network never fail and that routing paths used between all source/destination pairs remains constant. In this scenario, does a VC or datagram architecture have more control traffic overhead? Why?

Back 17

Datagram architecture will have more traffic overhead because packets between the same source and destination may take different paths. VC architecture will be more consistent: its connection-based architecture will produce no overhead in this kind of scenario.

Front 18

How big is the MAC address space? The IPv4 address space? The IPv6 address space?

Back 18

MAC: 6 bytes / 48 bits
IPv4: 32 bits
IPv6: 128 bits

Front 19

Virtual circuits

Back 19

“source-to-dest path behaves much like
telephone circuit”
 performance-wise
 network actions along source-to-dest path

 call setup, teardown for each call before data can flow
 each packet carries VC identifier (not destination host
address)
 every router on source-dest path maintains “state” for
each passing connection
 link, router resources (bandwidth, buffers) may be
allocated to VC (dedicated resources = predictable
service)

Front 20

Virtual circuit implementation
a VC consists of 3 things:

Back 20

1. path from source to destination
2. VC numbers, one number for each link
along path
3. entries in forwarding tables in routers
along path

Front 21

Datagram vs VC end systems

Back 21

VC: “dumb” end systems
Datagram:  “smart” end systems
(computers)
 can adapt, perform
control, error recovery
 simple inside network,
complexity at “edge”
 many link types
 different characteristics
 uniform service difficult

Front 22

Two key router functions

Back 22

1) run routing algorithms/protocol (RIP, OSPF, BGP)
2) forward datagrams from incoming to outgoing
link

Front 23

switching rate

Back 23

rate at which packets can transfer from inputs to outputs

often measured as multiple of input/output line rate
 N inputs: switching rate N times line rate desirable

Front 24

Three types of switching fabrics

Back 24

Switching Via Memory - first generation routers. traditional computers with switching under
direct control of CPU

Switching Via a Bus - datagram from input port
memory

Switching Via An Interconnection
Network - overcome bus bandwidth
limitations

Front 25

Head-of-the-Line (HOL) blocking

Back 25

queued
datagram at front of queue prevents others in
queue from moving forward

Front 26

An IP address is a 32-bit identifier for...

Back 26

a host and the router interface

subnet part = high order bits
host part = low order bits

Front 27

Subnet

Back 27

device interfaces with same subnet part of IP address

can physically reach
each other without
intervening router

Front 28

CIDR

Back 28

CIDR: Classless InterDomain Routing

 subnet portion of address of arbitrary length
 address format: a.b.c.d/x, where x is # bits in
subnet portion of address

Front 29

DHCP

Back 29

Dynamic Host Configuration Protocol
dynamically get address from a server
 “plug-and-play”

can return more than just address

Front 30

address format: a.b.c.d/x, where x is

Back 30

# of bits in
subnet portion of address