Exam Questions Flashcards
What is a private network?
A private network is a network that is not connected to the public internet. By convention a private network is given a subnet address from one of the reserved IP address
Describe the typical main forms of attack that can be made on the security of computer systems.
Eaves dropping, replay, denial of service – others can be described
Describe briefly public key and symmetric key encryption and the main ways in which they differ.
Public key – uses two keys, the private key is kept secret by the owner, and the public key is known openly. Any plaintext can be encoded by either the public or private key, but the opposite is required to decode. Encoding is computation intensive. Therefore used to encode relatively short messages.
Symmetric key - uses same key to encode and decode. It must be kept secret by sender and receiver. Encoding is much less computation intensive than public so used for large messages.
Describe digital certificates.
Digital certificates are used to prove the authenticity of the party presenting the certificate. They are in the form of a statement that is authenticated by a trusted party. The authentication is normally signed by the trusted party using their private to encode the signing and is authenticated by decoding with the public key
TCP, the Transmission Control Protocol is designed to provide a reliable service. Describe its important features to support this functionality.
A reliable transport will provide the following as features:
Error detection
Error correction (retransmission or forward error correction)
Flow control
Guaranteed order of delivery
TCP also provides traffic shaping through congestion algorithm and window protocol to maintain full transmission rate
Multi-media applications sending a real-time video stream would generally send packets at regular intervals. What would be the preferred transport layer protocol to support such an application? Explain your answer.
UDP is used to transport packets in the network. UDP has no retransmission; this is not required in this application as arrival time and order of packets is unpredictable and may arrive too late. RTP would be used to carry the multimedia packets as payload of the UDP packets in order to carry information on order, timing and sequence. Other protocols may be used in addition to support multi-media application such as H323, SIP and RSVP.
Describe the main physical topologies employed to implement networks.
Bus, star, point to point, (wireless)
Describe the CSMA/CD medium access protocol.
Carrier Sense, multiple access, collision detection – used for cable bus network
All nodes monitor the bus for silence before attempting to access the bus (carrier sense)
When silence is detected, all nodes waiting to transmit are allowed to attempt to transmit their data (multiple access).
Transmitting nodes monitor their transmission to detect if it is corrupted because 2 or more nodes are attempting to transmit at the same time (collision detection)
Transmission of data (minimum packet length) must be longer than maximum round trip time of the network to ensure corruption can be detected
When collision detected, sending station exerts a jamming signal, then transmission terminates.
Transmitting stations must alter the period before attempting to resend in order to resolve contention. This is a random selection from a number of slots. If there are further collisions then the number is increased, by a factor of 2. This is termed binary exponential back off.
Describe how network nodes are enabled to deliver an IP packet to the specific physical node with that IP address within the subnet.
The network node needs to resolve the IP address to the physical MAC address in order to deliver to the specific physical node. This is normally achieved through the ARP. The sending node broadcasts an ARP request within the subnet, and the node with that IP address responds. The sending node can then resolve the IP address to MAC address and sends the packet. Normally nodes will cache the address resolution for a period of time.
Describe the characteristics expected of a reliable transport layer, such as the TCP of the TCP/IP protocol suite.
Deliver data in same order as sent
Retransmit lost or corrupted data
Flow control to prevent loss by application
Explain the window protocol of the TCP and how it overcomes the limitation of round trip time to achieve full rate transmission within an extended network.
Multiple packets are transmitted and are in flight before acknowledgement is received for the first transmitted packet. The window is made large enough so that packets are transmitted continually and there is no pause to wait for an acknowledgement, and so transmission is maintained at full speed.
Explain how TCP manages its rate of transmission to ensure that a network does not become congested and bandwidth is shared fairly between users. . Illustrate your answer to show typical throughput.
TCP uses the congestion algorithm with slow start -up to match transmission rate to prevailing network conditions. Slow start up sends 1, 2, 4, 8 etc packets until loss, to ramp up to transmission speed, then uses linear growth. If there is further loss, then transmission rate is halved. Suitable diagram to be added.
Explain when the UDP might be the protocol of choice over TCP.
UDP is used for simple applications that have a single request and response which can each be placed in a single packet. This reduces significantly the number of packets (typically 2 in place of 9).
UDP is used for real time applications, such as multi-media, in order to deliver packets timely as there is no retransmission which could cause freeze in play out if retransmitted packet is not received in time.
Describe what a subnet is in a network system
A subnet is a single address domain, where all network nodes are addressed with the same subnet address and differentiate by their node address. This normally corresponds to a single LAN domain.
Many organisations choose to segment their network using a combination of routers and switches. Describe the reasons for this approach and advantages that it can provide.
Physical extension of network – subnets located in separate locations
Increase capacity – physical limitation of limited number of node addresses on single subnet
Reduce load – subnet can become overloaded so distribute to multiple subnets
Security – isolate secure systems on separate subnet
Resilience – isolate faults (e.g. virus)
