Internet Flashcards
Explain how the internet is a federated system.
Allows interoperability and information sharing between semi-autonomous de-centrally organized lines of business.
How does network relate to internet?
Internet is not a network technology. It ties different networks together.
What are the problems facing the Internet?
- Internet Scale
- Competing Entities must Cooperate
- Enormous Diversity of Devices
Formal definition of Internet.
A federated system of enormous scale with large dynamic range
and diversity that is asynchronous in operation failure prone and
constantly evolving
Describe the design paradigms of the internet. (5)
Decentralized control
A best-effort service model
Route around trouble
Dumb infrastructure with smart endhosts.
The end to end design principle - keep the core network simple and have functionality for reliability and the two endpoints.
Layering
Explain best effort service model.
No guarantee on whether or when data will be delivered.
No notification of outcome.
What do the following stand for?
SDN
NFV
DSL
ISO
OSI
SDN - Software Defined Networks
NFV - Network Functions Virtualisation
DSL - Digital Subscriber Line
ISO - International Organization for Standardization
OSI - Open System Interconnection
Explain layering principle of internet.
A part of a system with well-defined interfaces to other parts. One layer interacts only with the layer above and layer below. Two layers interact only through the interface between them. Each layer should have a single standard protocol.
If layer A is a Session protocol using PDU what is the unit exchanged called?
SPDU
Explain the hierarchical structure of the layers.
Each layer offers certain services to the higher layers without revealing how those services are implemented.
T or F. Communication is always horizontal in the Internet.
False.
T or F. Except in the physical medium, communication is always vertical in the Internet.
True
Explain encapsulation in network.
Encapsulation allows higher-level protocols to utilize lower-level protocols while still allowing protocols at every level of the stack to add their own protocol-specific data to each packet. The N + 1 - PCI is not visible to layer N.
T or F. The packet size will increase as a packet goes from application to link layers when being sent.
True
T or F. The packet size will decrease as a packet goes from link to application layers when being received.
T
Name four main layers and a process utilised on each of them.
Application layers (HTTP)
Transport Layer (TCP)
Internet Layer (IP)
Link Layer (Hardware)
What do the following stand for?
SAP
SDU
PDU
PCI
SAP - Service Access Point
SDU - Service Data Unit
PDU - Protocol Data Unit
PCI - Protocol Control Information
What is SAP?
Point between a layer and the next highest one.
What is SDU?
Is the block of data coming from the higher layer.
What is PCI?
Is the header that is added onto the SDU to make the relevant data block at this layer.
What is PDU?
Is the block of data that is given to the next layer underneath the current layer.
How are packets transported through network by IP?
Datagram packet switching
Explain how TCP/IP Reference Model is a client-server model.
A client machine requests a service and runs only when needed. A server program, runs all the time waiting for incoming requests.
What do the following stand for?
NIC
UDP
Network Interface Card
User Datagram Protocol
What is ‘port’ number?
Address of application that we must send the packet to.
What is UDP?
UDP attaches port numbers without any reliability or congestion control unlike TCP.
Explain what TCP does?
Gives each byte of data its own Sequence Number. Assigns port number, but also has congestion control and reliability.
What is MSS?
Maximum TCP Segment Size. This is the block size for what the receiver can take, usually set at the start and not changed.
What is congestion window?
How much I can send which changes all the time. Grows with each successful transmission.
What is receiver’s advertised window?
This is the maximum window size, congestion window cannot exceed it. Can be changed.
What is Threshold?
This is where the receiver goes from exponential increase to linear increase. It is initially set but changes when there is a timeout.
T or F. Packet loss due to transmission is quite common.
False
What does ACK stand for?
Acknowledgment
Explain TCP Congestion Control from setup.
When connection initialised, sender intialises congestion window to MSS and sends one MSS.
Grows exponentially, add one MSS to congestion window and send 2 MSS ….
If we try to exceed the receivers advertised window then congestion window = receivers advertised window.
When the congestion window reaches or exceeds the threshold, send that number of MSS, then increase linearly by one MSS.
Procedure for timeout.
When a timeout occurs, there is a reset of the connection parameters.
Reset threshold to 1/2 current congestion window.
Reset congestion window to one MSS.
Formula for time to send full window?
= RTT * number of succeful transmission before a full window can be sent.
What did Barbara Liskov win? When?
The Computer Science equivalent of the Nobel Prize in 2008.
T or F. Extracting a Network PCI from a Transport PDU results in a Network SDU.
F
List the layers in a network in order.
Application, Transport, Network, Data Link, Hardware.
T or F. A Transport PDU is the same as a Network SDU.
True.
T or F. A Transport SDU is identical to a Network PDU.
False.
T or F. The MSS in TCP can not be changed once set.
True
T or F. Extracting a Data Link PCI from a Data Link PDU results in a Network PDU.
True.
T or F. The congestion window in TCP changes a lot of the time.
True
T or F and explain. In UDP an IP Host and Port uniquely defines a host’s application process.
False.
If two hosts are communicating, the combination of source IP, source port, destination IP, and destination port defines the complete connection context.
T or F and explain. TCP preserves message (packet) boundaries.
Packets are treated as a stream.
False.
T or F. The actual communication in the OSI layered model is vertical.
True
T or F. The Internet only gives a best effort service with no guarantees.
True
T or F. A layered architecture means that each layer can look inside any other layer.
False.
T or F. The communication latency in the Internet varies by billions.
False.
T or F. Encapsulation means the data is secure.
False.
T or F. The Internet is designed around being a Synchronous system.
False.
T or F. The bandwidth in the Internet varies by billions.
True.
T or F and explain. There are four layers in the OSI Model.
False.
They are seven.
T or F. The bandwidth in the Internet varies by millions.
False.
T or F. Layer two in the OSI model is the Data Link Layer.
True.
T or F. The number of layers is counted from the bottom in the OSI model.
True.
T or F. The Internet is designed to give guaranteed communication services.
False.
T or F. The Transport Layer is the first end to end layer in the OSI model.
True.
T or F. The internet has centralised control.
F
T or F. The actual communication in the OSI layered model is horizontal.
False.
T or F. The communication latency in the Internet varies by millions.
True.
T or F. The MSS in TCP usually changes a lot during the connection.
False.
T or F. A hostname can have many IP addresses associated with it.
True
T or F. The receiver’s advertised window (RAW) can be changed at any time in TCP
True.
T or F. A MAC address can support many IP addresses.
False.
T or F. The Internet is a federated system.
True.
T or F. An IPv4 address has 48 bits.
False.
T or F. An N+1 layer PDU is the same as a N layer SDU.
True
T or F and explain. UDP sets up a connection for data transfer at the Transport Layer.
False.
UDP is connectionless
T or F. A domain name can be resolved to multiple MAC addresses.
True.
T or F. There are more Class B IP addresses than there are Class C addresses.
False.
T or F. There are over 4 billion IPv4 addresses.
True.
T or F. RIPE assigns all the IP addresses.
False.
T or F. Class A addresses start with numbers between 0 and 127 in the first octet.
False.
T or F. ICANN supports and controls the assignment of all IP addresses.
True.
T or F. An IP address can have multiple MAC addresses.
False.
T or F. An IP address can support multiple Domain names.
True.
T or F. The congestion window can be greater than the Receivers Advertised Window.
False.
T or F. The Threshold can be greater than the Receivers Advertised Window.
True.
T or F. RIPE assigns the IP addresses in the Asia Pacific Region.
False.
What do the following stand for?
ARP
MAC
Address Resolution Protocol
Medium Access Control
What is MAC address?
Physical address of computer. Invariant in time, built into hardware.
Is IP always the same?
No, internet address can change.
Explain the relationship between domains, IP addresses and MAC addresses.
IP to MAC is 1:1
Domain to IP is n:k
So Domain to MAC is n:k
How many bits in IPv4 address space?
32
Distinguish class A, B, C subnets.
Class A: 1 bit locked to 0. 2^7 - 1 networks. 2^24 - 2 hosts per.
Class B: 1st bit locked to 1, 2nd to 0. 2^14 networks. 2^16 - 2 hosts per.
Class C: 1st bit locked to 1, 2nd to 1, 3rd to 0. 2^21 networks. 2^8 - 2 hosts per.
How can A, B and C classes be easil distinguished?
First part of IP:
A:
1 - 126
B:
128 - 191
C:
192 - 223
List the different regions and their IP providers.
RIPE - Europe
ARIN - America
APNIC - Asia
What does CIDR stand for? Explain.
Classless Inter-Domain Routing.
Assigns IP addresses by building up prefixes assigned and using netmasks.
What does the netmask /28 mean?
Means the network id of the IP is 28 bits and 4 bits are left for the host id.
What is network function virtualisation?
Allows network to prioritise tasks with more urgency by building intelligence within the endhosts of a network, e.g. autonomous cars.
What is edge computing?
Involves reducing reliance on big data centres and the cloud. Means setting up preliminary computers on premises to analyse data briefly. Leads to faster reaction times in case of emergency but is more vulnerable to security risks for valuable data.
Distinguish between net id and host id.
Net id is network number, host id is your place in the network.
T or F. The threshold can be bigger than the Receiver’s advertised window.
True
T or F. A domain name can be associated with only one IPv4 address.
False
T or F. Encapsulation means that I do not have to understand the headers from higher layers.
True
T or F. When the subnetting mask is changed within my network, I must inform the other routers outside of my network.
False
T or F. A Network SDU is identical to a Transport PDU.
True
T or F. The IP address 153.182.201.187/30 has a broadcast address of 153.182.201.191.
False
T or F. 171.77.205.0 is a valid Host IP address with a mask of 255.255.128.0.
True
T or F. A MAC address is resolved by DNS to give a single IPv4 address.
False
T or F. A MAC address is resolved by ARP to give a single IPv4 address.
True
T or F. In TCP the congestion window can not be greater than the threshold.
False
T or F. Extracting a Network PCI from a Transport PDU results in a Network SDU.
False
T or F. Even though IP is connectionless and unreliable, it is still possible to have a reliable connection oriented service that uses IP.
True
T or F. In UCP an IP Host and Port uniquely defines a host’s application process.
False
T or F. A class C network has 256 possible hosts.
False.
T or F. A Domain name can not be resolved to many MAC addresses (using DNS and ARP)
False
T or F. Adding a Network PCI to a Transport PDU results in a Network PDU.
True
T or F. 194.17.41.7 is a valid host IP addresses with a Mask 255.255.255.248
False
T or F. An IP address with /26 will have a broadcast IP address ending in 255,191, 63 or 127
True
T or F. In a class C network the broadcast IP address must end in 255.255
False.
T or F. Broadcast address is 137.53.255.255 with a mask of 255.255.224.0 for an IP address 137.53.184.24?
False.
T or F. An IP address with /23 will have a broadcast IP address ending in 255.
True.
T or F. The decimal notation of the final byte of an IP address with /28 will always be even.
False.
Why does UDP exist? Can we let users send raw IP packets through network?
UDP exists to extend IP’s host-to-host delivery service into a process-to-process communication service.
No you can not let them send raw IP. Without any additional addressing information the destination network layer process responsible for delivering data to application layer would not know what to do with an arriving IP datagram.
Is it possible to have more than one TCP connection between 2 ports at the same time?
There can be multiple from port 1 or port 2 but not multiple linking port 1 and port 2 as TCP connection is identified uniquely by (port 1, port 2) pair.
If IP handles reassembly, does TCP not have to worry about data arriving in wrong order?
No. IP datagrams are sorted but the datagrams themselves can be out of order.
Why is the transmission of information between two peer
processes in either the TCP/IP Model or the OSI Model called
virtual transmission?
The processes are concerned only with the information content being exchanged, and not the way in which this information is actually transmitted.
List two ways in which the OSI Reference Model and the
TCP/IP Reference Model are the same, and two ways in which
they differ.
S:
- Both layered models.
- End-users don’t have to be concerned about the physical communication network.
D:
- OSI model has 7 layers, TCP/IP model has
4 layers.
- OSI connectionless or connection-oriented service whereas TCP/IP model only provides a connectionless.
Distinguish between where the congestion window and the transmission size sent occur.
At start congestion window is set to 1 MSS
Then in 1st transmission, 1MSS is sent and congestion window is set to 2 MSS.
Then in 2nd transmission, 2MM is sent, and congestion window is set to 4 MSS.
Distinguish between ARP and Reverse ARP.
ARP - Gets MAC for given IP address
Reverse ARP - Host broadcasts MAC address to get its own IP address.
Using subnetting what is the net id, subnet id and host id roles.
The Netid defines the site
The Subnetid defines the physical network
The Hostid defines the actual machine
T or F. Bits arrive on wire in physical layer.
True.
T or F. Packets delivered across links and local networks in network layer.
False.
Datalink Layer
T or F. Packets delivered between networks for global delivery in network layer.
True
What does OTN stand for? What is it used for?
Optical Transport Network
Transfer between Autonomous systems.
When is fragmentation of a datagram needed? How is it implemented?
Is needed when an intermediate network has a max frame size too small to carry the datagram.
All fragments of a given datagram have the same identification value.
How does IPv6 differ from IPv4?
128 bit address vs 32 bit address
Why is IPv6 not implemented?
Not compatible with IPv4, so hard to roll out universally.
Host with IP address 128.96.34.15 with subnet mask 255.255.255.128
How do you find the subnet the host is attached to?
Perform bitwise AND operation between values of ip address and subnet mask.
T or F. The broadcast address on a subnet always ends with 255.
False.
T or F. Subnetting and CIDR are really the same.
False.
T or F. The MSS in TCP is the maximum size of the segment that a device can send of the receiver can receive before the TCP and IP header is added.
True.
T or F. The broadcast address for the IP 192.168.153.33 /26 is 192.168.153.63
True
