COMP247 Data Communications

a)As we have seen, the Internet layer of TCP/IP has two protocols – IPv4 and IPv6. The transport layer provides two main protocols TCP and UDP (along with some special-purpose, minor protocols). But these are for very different reasons. Why are the Internet and Transport layers very different in providing two protocols.
b) Since UDP does very little that is different from IP, why do we even have a transport layer such as UDP
c) Describe how DDoS attacks may be mounted against TCP and UDP services. In what way will being connection oriented be different for TCP and UDP

a) Perform the following boolean operations on the given bit patterns 
i. not 11100010
ii. 11100010 or 00001110
iii. 11100010 and 00001110
b) What are the subnet masks for the following slash address blocks
i. /9
ii. /17
iii. /31
c) Are the following address masks legal? If so, what are their slash forms If not,why are they not legal
i. 255.255.248.0
ii. 255.224.252.0
iii. 255.255.255.224
iv. 255.255.156.0
d) How many computers are in networks with the following address blocks.
i. /13
ii. /26

e) A small company is assigned the class C network 205.67.35.0. You need to divide this network to provide subnets for three departments using fixed-length subnetting. The departments are approximately the same size. Show the network
and broadcast addresses for each subnet. Show the slash notation. How many hosts are on each subnet Are the resulting subnets class C networks.

a) Explain why CSMA/CD scheme used in traditional Ethernet LANs cannot be used in 802.11 wireless LANs.
b) A supermarket uses WiFi to enable staff to check inventory levels and to perform price checks. Staff have complained that the wireless scanners they use to check stock levels do not work reliably in some isles. The supermarket manager has contracted you to diagnose the wireless problems and correct them.

Due to the heterogenous nature of networks in terms of network infrastructure, the transport and the internet layers have to accommodate the various protocols that could be in use in various LANs, WANs, WLANs or Wireless WANs. Separation of tasks is a key factor in this process. At any one time, during a session, if a protocol cannot be implemented in one of the two layers, it will be implemented on the other. This includes flow, error and congestion control. The aim of this is to provide reliability and resilience in network connection and the most critical point is between the transport layer and the internet layer, where the two networks actually connect.

The Transmission Control Protocol (TCP) is a connection-oriented protocol, which first establishes a logical link between two computers before transmitting any data. It provides flow, error and congestion control. TCP is ideal for Local Area Networks where all computers “know” each other and there is no overwhelming traffic.

The User Datagram Protocol ( UDP) is a connectionless protocol, which means it does not need to establish an logical link before transmitting. When UDP transmits IP datagrams they become independent of each other and can take any of n  routes to reach the destination host. It provides no flow, error or congestion control. UDP is useful for WANs, especially the internet, where there is no time for the various reliability controls. This task is relegated to the responsible protocol.

In the internet layer, the dominant Internet Protocol (IP) is a connectionless protocol which has no error, flow or congestion controls. The IP determines the address structure and format of  a data packet to be used through its variants, IPv4  and IPv6.

1. Since UDP does very little that is different from IP, why do we even have a transport layer protocol such as UDP

The User Datagram Protocol is capable of sending multiple requests to a server(s) through multiplexing and to receive multiple data through demultiplexing. This enables the user to run multiple network application at the same time from his device and to receive data simultaneously from the servers he/she is connected to.

The above is achieved by the use of port numbers at the application level of the TCP/IP stack. Port numbers consist of the source address and the destination address, each of which is 16 bits. Hence a port number is made up of a total of 32 bits. When the client computer sends multiple requests, the port number multiplexes the requests by combining the various signals and transmitting them over a single channel. The receiving computer demultiplexes the signals and directs them to their respective applications. The returned data passes through the same process when the server answers.

 The main difference between UDP and IP is that the former’s datapacket/ da has a header with unique identification for a specific device

1. Describe how DDoS attacks may be mounted against TCP and UDP services. In what way will being connection oriented be different for TCP and UDP.

A distributed denial of service attack (DDoS) is an attack executed over a network (mostly WANs) from multiple sources, where the perpetrator seeks to make a machine or network resource unavailable to its legitimate users by temporarily or indefinitely disrupting services of a host connected to the network. Accomplished by flooding the target machine or resource with redundant and bogus requests by using numerous IP addresses in an attempt to overload systems and prevent some or all legitimate requests from being fulfilled.

UDP has no flow or congestion control and services could be almost instantly be unavailable. TCP may be able to handle the flood of incoming packets since it has flow and congestion controls, but if the packets number overwhelms it will succumbe could succumb.

1. Perform the following boolean operations on the given bit patterns 
2. not 11100010 = 00011101
3. 11100010 or 00001110  = 11101110
4. 11100010 and 00001110 = 0000001

11111111.11111111.10000000.00000000 = 255.255.128.0

iii.) /31

11111111.11111111.11111111.11111110 = 255.255.255.254

1. Are the following address masks legal? If so, what are their slash forms? If not, why are they not legal.

i.) 255.255.248.0

     The address is legal

     /29

ii.) 255.224.252.0

The address is not legal. The second and third octets in the address are both partially set, making it difficult to determine the network class it belongs to(B or C) and therefore making subnetting impossible.

iii.) 255.255.255.224

The address is legal

/27

iv.) 255.255.156.0

The address is not legal. No zeros are supposed to come before any ones in the net id bit set.

1. d) How many computers are in networks with the following address blocks? (2 marks)

i.) /13

2¹⁹ -2 = 524286 computers

ii.) /26

2⁶-2 = 62 computers

205. e) A small company is assigned the class C network 205.67.35.0. You need to divide this network to provide subnets for three departments using fixed-length subnetting. The departments are approximately the same size. Show the network and broadcast addresses for each subnet. Show the slash notation. How many hosts are on each subnet? Are the resulting subnets class C network.

• Set the first two bits to one on the subnet mask’s last octet to accommodate three subnets, as follows:-

Subnet mask =      11111111.11111111.11111111.11000000 = 255.255.255.192

Number of subnets = 2² = 4 subnets

Hosts per subnet = 2⁶-2 = 62 hosts

Subnet addresses broadcast address Host addresses

Subnet 1:  205.67.35.0/26 205.67.35.63 1 - 62

Subnet 2:  205.67.35.64/26 205.67.35.127 65-126

Subnet 3 : 205.67.35.128/26 205.67.35.191 129-190

Subnet 4:  205.67.35.192/26 205.67.35.255 193-254 

802. Explain why CSMA/CD scheme used in traditional Ethernet LANs cannot be used in 802.11 wireless LANs. 

Carrier Sense Multiple Access with Collision Detection (CSMA/CD) schemes operate in full duplex mode. They can both send and receive signals at the same time. That is why devices implementing them are able to transmit while at the same time receiving signals to detect collisions.

In contrast, a majority of wireless LAN devices are half duplex. They cannot transmit signals and receive them at the same time, making it impossible to implement the CSMA/CD scheme. CSMA/CD also detects collisions by measuring power levels on the medium being used, which in the case of wireless LAN radio frequencies, is low to detect a packet collision.

There is also the performance issue because CSMA/CD and even Collision Avoidance (which is halfplex) schemes use acknowledgements for error checking in its data transmissions, hence lowering throughput for wireless devices accessing a central point.

1. A supermarket uses WiFi to enable staff to check inventory levels and to perform price checks. Staff  have complained that the wireless scanners they use to check stock levels do not work reliably in some isles. The supermarket manager has contracted you to diagnose the wireless problems and correct them. The supermarket is 60m by 40m. Assume that the useful range of WiFi within the warehouse will be 15m radius. The access points are organised into a grid with the following channel allocations:

i.)You have determined that the channel allocations are not optimal, how would you improve upon the current channel allocations

The WiFi grid in the supermarket could be experiencing possible co-channel and adjacent channel interference problems with only channel 1 and 6 on the grid. These are the most common problems in a 2.4 Ghz spectrum WiFi grid.

Co-channel interference (CCI) occurs when two access points are transmitting at the same frequency, in the same vicinity. The two signals tend to combine to form a bigger virtual access point, where they both will contend for a receiving device. WiFi utilizes the CSMA/CA scheme when CCI occurs in a grid to check for traffic, until the medium is clear. In this situation their may be delays in connections.

Adjacent channel interference (ACI) occurs when a transmission occurs on overlapping channels. Either of the channels may introduce noise and interference to the other channel, causing the receiving device to suspend connection indefinitely.

As ACI and CCI are the most likely problems with only two channels available, the grid could be optimized by introducing a third channel (channel 11) in the plan. A visualization software can be used to determine the effectiveness of the new plan. The new plan would look something like this:-

Supermarket stock may also be interfering with the signals since wireless connections may be affected by some materials. In this situation, the solution could be to adjust antennas or slightly move the device’s location. For antennas mounted on the ceiling, tilting downwards and sideways could do the trick.

ii.) As you are reconfiguring the access points you notice that none of them currently have any security enabled. Is this a problem, and if so, what type of security would you recommend

Disabled security is a problem that also needs to be addressed. The security needs to be enabled to prevent unauthorized access and possible damage to the supermarket servers.

The recommended type of security is the Wi-Fi Protected Access (WPA) wireless security, which is more secure than its predecessor the Wired Equivalent Privacy (WEP) wireless security.

WPA encrypts the network with a 256 bit key, which is longer than the one WEP used. This makes it stronger and the network not easy to break into.

The most current standard of WAP is WAP2. A wireless intrusion system can also be used to set wireless security policies.

References

Comer, D.E, (2009). Computer networks and internet. Pierce Prentice Hall. ISBN 10: 0-13-606127-3

Bononi, L., Conti, M., Gregori, E., (2012). Run-Time Optimization of IEEE 802.11 Wireless

LANs performance. National Research Council (CNR) IIT Institute

Davis J.S.: Channel Assignment Schemes by (n.d). Retrieve from Four Layers of TCP/IP model,

Comparison and Difference between TCP/IP and OSI models (2017)-http://www.omnisecu.com/tcpip/tcpip-model.php [17th May, 2017]

IP Addressing and Subnetting for New Users (August 10, 2016). Retrieved from 

http://www.cisco.com/c/en/us/support/docs/ip/routing-information-protocol-rip/13788-3.html [17th May, 2017]

Ketonen, V., (2014). Wi-Fi / WLAN Performance Management and Optimization. 7 Signal, The

WiFi performance Company.