It is important for you all to attempt every question. There are 6 questions, each question is worth 20 marks giving a total of 120 marks.

It is important for you all to attempt every question. There are 6 questions, each question is worth 20 marks giving a total of 120 marks. The first 70 marks go into bronze, then next 30 into silver and the last 20 marks go into Gold. Q7 is a bonus question worth 2% on top of your course mark.

The techniques of critical thinking are needed for answering the questions below and are what you need to understand and answer the exam questions. We have dealt with these topics both in the lectures, the tutorials and to some extent in the quizzes. Please note that in the exam the most you should spend on any one question is 18 minutes.

Q1 (20 marks). The Virus on a Network model in the NetLogo library is useful for understanding the Netlogo assignments. It can be used to help in modelling the spread of trust or mistrust. How would you evaluate the model for its usefulness to your work in the Netlogo models?

Q2 (20 marks). How does a distributed file system balance the requirement for scalability and fault tolerance?

Q3 (20 marks). What is a complex system and how do we detect emergence? In the NetLogo flocking model that we discussed in class and that you used in the first assignments what, in your opinion, is the emergent behaviour and why?

Q4 (20 marks). The Distributed Hash Table and the Chord lookup is used extensively in overlay networks. What strategies are used to ensure fault tolerance? What happens if the storage capacity of a node is exceeded?

Q5 (20 marks). In considering synchronisation issues of distributed systems we discussed the necessity of capturing causality. How do vector clocks capture causality? Why is this important in considering fault tolerance in invocation processes in object-based distributed systems? 

Q6(20 marks). The reference book makes a distinction between remote objects and distributed objects. What is the difference? Why is it useful to define the interfaces of an object in the Interface Definition Language. What is a client stub? 

Q7(2% course bonus). Outline a simple implementation of a fault tolerant Python space. To answer this question, you have to think in terms of fault-tolerant architectures.