IMPERIAL COLLEGE OF SCIENCE, TECHNOLOGY AND MEDICINE EXAMINATIONS 2019 BEng Honours Degree in Electronic and Information Engineering Part II MEng Honours Degree in Electronic and Information Engineering Part II BEng Honours Degree in Mathematics and Computer Science Part III MEng Honours Degree in Mathematics and Computer Science Part III MSc in Computing Science for Internal Students of the Imperial College of Science, Technology and Medicine This paper is also taken for the relevant examinations for the Associateship of the City and Guilds of London Institute PAPER C527 COMPUTER NETWORKS AND DISTRIBUTED SYSTEMS Monday 29th April 2019, 14:00 Duration: 120 minutes Answer THREE questions Paper contains 4 questions Calculators not required Section A (Use a separate answer book for this Section) 1a Explain the function of client and server stubs in Remote Procedure Calls. b Briefly explain how Java RMI avoids the use of server stubs (or skeletons). c A system implemented using Java RMI enables users to share their music and playlists. Songs and playlists are hosted on music servers and can be accessed remotely. A music server allows users to upload new songs and to create new playlists. A playlist has a name and contains several songs. Playing a playlist returns the audio bytes for all the songs in the playlist. A song has a title and stores its music as an array of bytes. You can assume that the implementation for the Song class already exists. Consider that the objects described above conform to the interfaces given below. public interface MusicServerI extends Remote { public SongI uploadSong(String title, byte[] song) throws RemoteException; public PlayListI createPlayList(String name) throws RemoteException; } public interface PlayListI extends Remote { public String getName() throws RemoteException; public void addSong(SongI song) throws RemoteException; public byte[] play() throws RemoteException; } public interface SongI extends Remote { String getTitle() throws RemoteException; byte[] play() throws RemoteException; } i) Briefly discuss whether parameters and return values are passed by value or by reference in the interfaces given above. ii) Give a Java RMI implementation for the MusicServer class. iii) Give a Java RMI implementation for the PlayList class. iv) Give a Java RMI implementation for an application that adds a song that exists on one music server to a new playlist on server rmi://mymusic.com/server. The RMI reference of the song, in its URL format and the name of the new playlist are parameters of the application. Strict Java syntax is not required but your answer must include all the components of the RMI invocations. State explicitly any assumptions that you make. The three parts carry, respectively, 15%, 15%, and 70% of the marks. c Imperial College London 2019 Paper C527 Page 1 of 4 2a i) Explain what a certificate (as used in PGP and X.509) is and list 5 fields that you would expect to find in a certificate. ii) Explain what is meant by a certification authority hierarchy in X.509 and what steps need to be taken to verify a certificate. Use a diagram to illustrate your answer. iii) Briefly explain how trust is handled differently in PGP and X.509. iv) It has been argued that secure group communication i.e., where messages are sent to a group of recipients, is impractical with PGP. Discuss, using examples, why this might be the case and how such group communication could be realised. b A company riss.com needs to synchronise its computers to the Coordinated Universal Time (UTC) using the Network Time Protocol (NTP). To this end it needs to host one or several NTP servers, which synchronise with external NTP servers and with which the company’s computers can synchronise. riss.com has several thousand computers. NTP interactions use the UDP protocol and servers use port 123 to receive queries or to issue queries to other servers. The company also advertises its activities on a web-server available from the Internet and its users are allowed to access external web-servers. i) Assuming riss.com uses a screened subnet firewall architecture discuss how many time servers the company would need and where they should be placed with respect to the internal and external firewalls. ii) Assuming riss.com uses a single time server placed in the DMZ between the external and the internal firewalls give the configuration for the external and the internal firewalls in the format below. Rule No. Protocol Dir. Source Address Src. Port Dest. Address Dest. Port TCP Flags Action The two parts carry equal marks. c Imperial College London 2019 Paper C527 Page 2 of 4 Section B (Use a separate answer book for this Section) 3a i) Explain the general idea behind data framing and discuss its advantages and disadvantages. On which layer of the OSI Reference Model is data framing defined? ii) List 2 framing methods that do not require perfect time-synchronisation between sending and receiving nodes and explain how they work. Discuss any shortcomings these methods might have and how these shortcomings might be overcome. iii) What is the inter-frame gap that is defined in IEEE 802.3 and why is it used? Explain how burst mode works in this context. b Explain what ICMP is and what it is used for. Indicate which layer of the OSI Reference Model it belongs to and explain why. Furthermore, list two different programs that utilise ICMP. Discuss their use cases and how exactly ICMP is utilised to achieve their tasks. The two parts carry, respectively, 65% and 35% of the marks. c Imperial College London 2018 – 2019 Paper C527 Page 3 of 4 4 Consider the computer network given below, which consists of a router, a switch, a hub, and 4 workstations (W1, W2, W3, W4). The workstations are connected to the ports (p1, p2) of the switch or hub, which connect to the network cards of the router (i.e. eth1 and eth2). The router has a third network card, eth3, which is connected to the network of an Internet Service Provider and which provides connectivity to the Internet. The network uses the Internet Protocol (IP) and the IEEE 802.3 standard with an MTU size of 1500 bytes. Workstations W1 and W3 are correctly configured with: W1 W3 MAC Address 8c:dc:d4:57:61:ab 8c:dc:d4:16:b7:5a IP Address 146.169.25.10 146.169.28.5 Network Mask 255.255.252.0 255.255.252.0 Default Gateway 146.169.27.254 146.169.28.1 Answer the following questions and explain all your answers: a All devices are properly configured and all workstations can communicate with each other as well as the Internet. What are the IP addresses and network masks of the network cards eth1, eth2, and eth3 of the router? What would a correct static routing table on the router look like? b Explain what backwards learning is and how it works in the context of the operation of a switch. c Based on the given information above give a detailed explanation of how 2000 bytes of data that is passed to the network layer would be transmitted from workstation W1 to workstation W3 using the given addresses. Describe the operation of all involved protocols and communication devices in chronological order. Your explanation should only relate to the network layer and data link layer of the OSI Reference Model. Medium access control does not have to be considered and you can assume that all channels are free of errors. The three parts carry, respectively, 25%, 15%, and 60% of the marks. c Imperial College London 2018 – 2019 Paper C527 Page 4 of 4
