Binder 1 PDF

Preview

Summary

mkkjikjik...

Description

Page | 3

Problem overview Dragon Boat Racing is a growing sport in Australia. In recent world championships held in Korea Australian teams performed well. Both the women’s and men’s teams finishing in the top five. In case you are unfamiliar with the sport the photo shows a dragon boat race. The boats have a crew of up to 20 paddlers, a sweep and a coxswain who shouts encouragement (among other things) and controls the pace of paddling by beating on the drum.

Dragon boat races are short by boat race standards. Mostly they are 200 metres. Since the boats are large they aren’t the kind of thing that every team owns and takes around the country to regattas. Instead the two sets of boats are usually supplied. One set will be racing, the other will be changing crews in preparation for the next race. The boats are usually assigned designated race lanes so teams are allocated to lanes, using whatever boat is assigned to that lane. Dragon boat regattas have all the usual problems of sporting tournaments. More teams arrive than were expected and teams withdraw during the day. The regatta is run in heats. Heat winners and, usually, the fastest loser progress to the final. The rest compete in repechage races for remaining places in the final. A repechage is “a heat (as in rowing or fencing) in which the best competitors who have lost in a previous round compete for a place or places yet left in the next round” As an indication of the size of the task of organising races, a recent regatta involved seven divisions, between 10 and 24 crews per division. The smaller divisions require two heats with winners and fastest loser going to the grand final, the rest competing in repechage heats. Winners of the repechage heats also going to the final. The larger divisions had heats (and repechages), semi-finals and finals. A crew might gain entry into a semi-final only to withdraw due to one or more of their crew members becoming unavailable. When that happens, usually the fastest of the remaining losers will be invited to compete in the semi-final. Business processes Plan regatta categories. These are combinations of; Gender {Women, open, mixed} Age {Grand master, master, premier, youth, junior} Crew Size {20, 10} Accept entries. Page | 4

Schedule races (heats, repechages, semi-finals, finals) Allocate crews to heats, repechages, semi-finals, finals. Notify crews of event, time, lane. Withdraw crew. Promote crew to race in an event (after a crew withdraws). Record race results. Usage narratives

1. Regatta organiser I want to manage the regatta entries much more easily. Preferably the clubs should be able to enter their crews directly into the events through some web page. Surely we are past the age of sending a piece of paper. That usually led to late entries, changed entries and all manner of problems. If the clubs can manage their own entries then all those problems can be sorted out by the clubs right up until race day. It becomes a different matter on race day. If a club wants to withdraw a crew from an event then we need to reshuffle the race entries, depending where we are in the day. If it’s a heat or a repechage then they simply withdraw. But if a crew is withdrawing from a semi-final or final then we like to invite the next fastest crew to compete. It makes for better racing if we have all lanes full in the race. Right now, crews are allocated to lanes for a race. But that can leave a gap in the middle if a crew withdraws. Depending how much notice we get, we could shift crews across lanes so that the crews are closer to the crowd and each other. We could also allocate faster crews to the centre lanes as they do in swimming, or we could allocate them to lanes closer to the crowd. Sometimes we would like to reschedule entire races so that crews are not racing too close together. Races might be short but they are tiring. Crews need a certain amount of time to recover so if it happens that one crew hasn’t enough time to recover it would be nice to be able to shift the entire race without causing chaos. We could do that if we were able to send out notices to all the crew captains about their changed schedules.

2. Crew captain I need to know if we could be called on to race again. If we are finished for the day, we all like to relax with a beer or go home. But is isn’t fun trying to put a crew back together after they have all stopped for the day. Yes, I know we should all wait until the day is over before clocking off but some of these events drag on and on. The weather isn’t always fun. When the sun is out no-one minds waiting around but when it is cold, wet and raining we don’t want to be standing around anymore than we have to. Better we wait in the tent. So we want to know when we are racing, which lane and at what time so we leave the comfort of shelter as late as possible and spend as little time as possible in the weather. If someone pulls out I want to be able to nominate a replacement, if I can find one. We try to keep within the rules but can sometimes be caught out putting someone in a crew who shouldn’t be there. We would rather know before the race so we can withdraw or find someone else, than after the race when someone protests and we get disqualified. Getting notices on my phone would be good. That way I don’t have to wait around for postings on noticeboards. The whole thing could respond faster to changing circumstances.

3. Dock marshal It can get confusing when crews pull out. Suddenly we have a crew rushing onto the dock, not quite sure which lane they should be in, tripping over people who are waiting for their event. Often the crew and even lane allocation can change more than once, leading to Page | 5

Page | 7

Part B, Question 3 Sequence diagram for Dragon Boat Race

Page | 8

Part B, Question 3. Modify the system to add security.

Page | 9

Part B, Question 4

Part C: Long-answer question (10 marks) Select one question from a choice of three, listed below. Part C is worth ten marks. Write your answer in the second 5-page booklet provided. You are expected to write between 2-4 pages for your answer. Only one question will be marked. If you attempt more than one question, only the first one attempted will be marked, unless clearly crossed out. 1. What are the main features and constraints of a web based systems and what architectural techniques are used to respond to those features and constraints. 2. Web systems can face demands of rapidly changing workload. Describe how a web system architecture can respond to variable workload. Discuss the benefits and costs of your proposed architecture. 3. Your company will develop a mobile application to extend its existing system to customers on the move. Describe the usability concerns of such an application and how you would respond to them. Page | 10

Part A: Short-answer questions (10 marks) Write your answers in the 5-page booklet. There are ten questions, worth 1 mark each. 1. Name one stakeholder that would not normally review the architecture? 2. Which architectural quality opposes ‘performance’ and is it opposed? 3. What is the main purpose of the conceptual architecture? 4. What concerns remain unresolved after the execution architecture has been designed? 5. Which architecture quality is achieved by the implementation architecture? 6. Briefly describe one of the major architectural patterns. 7. There are six main software architecture qualities. Name three of them. 8. Why is it important to document the architecture rationale? 9. Choose one software architecture quality and describe the architecture tactics used to achieve it. 10. What are the main concerns that need to be resolved in product line architectures?

Select one question from a choice of three, listed below. Part C is worth ten marks. Write your answer in the second 5-page booklet provided. You are expected to write between 2-4 pages for your answer. Only one question will be marked. If you attempt more than one question, only the first one attempted will be marked, unless clearly crossed out. 1. What are the main concerns of a product line system. Describe how these are normally dealt with in a product line architecture. 2. In web system architectures there are three main authorization strategies; role based, resource based and claims based authorization. Describe one of them then describe the circumstances in which it is suitable and circumstances in which it would not be suitable. 3. When a failover system is configured with an active and a dormant system, the dormant system might not be up to date when the active system fails. Usually this is because data is not necessarily provided to both active and dormant and because the active system may change the status of the system without telling the dormant system. Describe how you would design an active - dormant failover system so there was minimal loss of data when the dormant system takes control

Page | 2

OVER

Part B: Design questions (20 marks) This section of the exam asks you to exercise your architectural design skills. There are four questions in this section. Write your answers to Part B in the 20-page booklet, unless instructed to write your answers in this paper. 1. Identify the main quality attributes and their relative importance to this system. Group components to show how you would design an execution architecture that achieves each of them. (5 marks) 2. Draw an implementation architecture by putting the components in packages that would be deployed to separate servers. (5 marks) 3. Choose and describe an architectural pattern that suits this problem. One of the diagrams show how this pattern would be implemented. (5 marks) 4. If this system was extended by adding a tablet or smartphone app (android or IOS app), how would you extend the proposed future architecture to support this. (5 marks). Smart Water Management Most water meters in Australia require someone to be physically present to read the meter. Then the reading must be entered into a system. Once the information is in the system invoices for water usage can be generated and various analyses of water usage run. The problem is that the information is always historic, lagging by as much as two months behind the present. This makes it hard to know patterns of water usage or to detect leaks from changes in water consumption. Western Australia Water Corporation in Perth, Yarra Valley Water in Victoria, Toowoomba City Council and Gold Coast City Council conducted investigations on smart metering in their jurisdictions. Currently South-East Queensland (Urban Water Security Research Alliance) conducts end-use study incorporating the Sunshine Coast, Brisbane, Ipswich and the Gold Coast in Queensland. Sydney Water and various consultants have been using smart metering technology since 1996 to conduct water efficiency audits for their business customers. They have also conducted research projects on residential use and houses with rainwater tanks. To support a rollout of smart meters and smart metering to all consumers, any of these water suppliers will need to modify their systems to accept data directly from the meter and to provide real time information to consumers. The new smart meters simply add a SIM Data logger to the existing water meter that enables readings to be taken at fixed intervals, stored on the SIM card then communicated to the water supplier using GPRS (General Packet Radio Service). Essentially the SIM data logger acts like a mobile phone to send data to the water supplier. With these new SIM data loggers, water suppliers will be able to provide their customers with real time information about their water consumption as well as alert them about any sudden changes in water usage patterns. This information would be available to consumers through a web page for their account. Some corporate accounts are likely to have multiple water meters, some smart meters and some older meters. Corporate account customers may want to see information for a one or more selected meters. For example, a school district may have several meters at each school, some for swimming pool water, some for sports ground water, some for toilet water and some for general consumption. A school district may want a report or graph showing a comparison of swimming pool water use across a number of schools or may want to see a report of all water use by the one school over several months. Part of the change to introduce smart meters will include making the system secure against unauthorised access. Page | 3

OVER

Original system architecture The original system architecture for water management was a simple database centric architecture in which data was entered from the data collector in batches. That is, a meter reader would take their hand held recorder to read a number of meters located at customer’s houses. When they returned to the office, meter readings would be downloaded from the hand held device. The conceptual architecture is shown in Figure 1.

Figure 1: Original system architecture for water management

Part A: Short-answer questions (10 marks) Write your answers in the 5-page booklet. There are ten questions, worth 1 mark each. 1. What is the purpose of stakeholder analysis? 2. What are the two main approaches to identifying the primary stakeholders? Briefly describe how each helps to identify potential stakeholders. 3. There are four main runtime quality attributes that affect a system’s architecture. Name 3 of them and give a short description of each 4. There are five major and two minor non-functional quality attributes that can affect a system’s architecture. Name three of them and briefly describe typical architectural tactics for achieving them. 5. Describe what a view is and why we need different views in architecture. You may refer to the three architectures we used in this subject. 6. Use case maps are a technique for exploring and evaluating the behaviour of the architecture. Describe how you apply use case maps to the conceptual architecture and what you can learn from doing so. Use a diagram to facilitate your explanation. 7. Name three triggers for identifying concurrency needs in the execution architecture view. Illustrate each trigger with a concrete example. 8. What are the major constraints on the architecture of embedded systems? 9. Describe how the implementation architecture is related to the conceptual architecture. 10. Give two reasons why interfaces are important. Phrase your answer in architectural terms.

1

Over/...

Part B: Design questions (20 marks) This section of the exam asks you to exercise your architectural design skills. There are four questions in this section. Write your answers to Part B in the 20-page booklet, unless instructed to write your answers in this paper. Refer to the text and diagrams provided on the following pages. 1. Explore the behaviour of the conceptual architecture, by applying use case maps. Do this for the 4 scenarios listed on page 5. For each scenario, draw the trace of the use case map on the supplied working copies of the conceptual architecture diagram. Clearly reference each diagram in your answer. Start the trace with a named event. As you cross each component, identify the responsibility exercised by it. Decide which use case maps to draw on a single diagram – some issues may be exposed only through multiple traces crossing the same component. Produce a list of component responsibilities. Then reason about any potential weaknesses in the current architecture exposed by the use case maps, which will need to be rectified in the next iteration. [6 marks] 2. Identify the main quality attributes and their relevance importance to this system. Describe how the proposed architecture achieves each of them. Describe how the execution architecture should be modelled to better achieve the relative importance of the quality attributes. Draw a process view to illustrate your decisions. [5 marks] 3. Improve the architecture by applying the following two architectural patterns – pipeline and client-server. As a result, redraw the conceptual or execution architecture (whichever is most appropriate), clearly highlighting the application of these patterns. Provide justification for your decisions, taking into account the quality attributes most significant for this system. [5 marks] 4. Develop an implementation architecture showing both infrastructure and application components. Include justification for your choices of software technologies, network protocols and API calls. You have to work with the following constraints: a. The build-time structure must suit a distributed development team, where some employees are in Asia and Europe, as well as Australia. b. The team of programmers in Asia is highly skilled in MySQL and PHP for online database management/access. c. The hardware on the satellites was installed ten years ago. The hardware consists of a real-time embedded processor - a reconfigurable FPGA (Field Programmable Gate Array). The software is written in C++. The hardware platform cannot be replaced, but new versions of software, compatible with the platform, can be uploaded. [4 marks]

2

Over/...

Problem overview

Howard Hugs, a rich Australian entrepreneur-turned-conservationist, is outraged by a foreign government's plan to kill humpback whales in Australian waters and in the Southern Ocean whale sanctuary. Rather than pursue diplomatic means of resolving the problem, he has decided to build and deploy a warning system that governments and other concerned organisations can use to try and prevent the slaughter of humpbacks. This system, the Whale Early Warning System (WEWS), will track whales and whaling vessels, and try and predict encounters between them. Howard knows that the technology to do this is only just "there", but he is determined to try anyway. He is even prepared to pay for it by selling his garage full of Rolls-Royces and driving a Volkswagen instead. Howard has already had some initial work done on the architecture of the system. You ask him about it. "Well, we've got two sources of data for tracking the whales and the ships. We call the ships the 'enemy'. [laughs] First off, some of the whale pods have individuals tagged with radio transmitters. CSIRO has agreed to let us have a real-time feed from their tracking system, as long as we don't let anyone else have it ... for obvious reasons ... "The second source of data is satellite imaging. I've negotiated a deal where we can get access to some limited real-time bandwidth - but we do pay a lot for it. So we have to maximize use of that bandwidth in two ways. First, by being careful about how we direct the cameras for those times when we have use of them. And second, we can upload program modules onto the satellite itself. Those of course have to be checked very carefully, both for security and real-time performance, but this means we can do some recognition and compression before trying to send the data. "So, it's not perfect, but it's something. Then we have to analyse it and figure out what's a whale, what's a boat, and so on. That's pretty heavy-duty stuff, a lot of those servers will be doing nothing but crunching on that recognition stuff. Then we run the predictive algorithms and ... well, we will have to see how well it goes and then spend a lot of time fine-tuning it, I guess. In-house, we'll have scientists here "helping" the automatic algorithms along and training them. Anyway, we will provide access to the data to any organisation that we approve to have it ... And some of it will be made public on a website as well, we want the public to know what's really going on. "So, we're nearly up and running on the hardware side ... we've got servers into the data centre already, but we need to get the software side of things going pretty fast. That's where you come in. Interested?" (In subsequent conversation, Howard Hugs reveals that he spent so much on WEWS that he can't afford to pay you yet, but you can be chauffeured around in one of the Rolls Royces until the ...