Other代写:COMP255 Requirements Elicitation and Specification Part2



THIS SPEC MAY BE FURTHER DEVELOPED this week. Any changes will be highlighted here…

You have been approached by a client who has an idea for a system that they intend to offer to sell to Air Services Australia for use at Sydney Airport.

The problem as they see it

Currently a consortium including the Defence Science and Technology Group (DSTG) is working on a new air traffic control system for Australian airspace including both civil and defence operations. Such a system is a remarkably difficult and safety critical system development job. It has a long timeline and is behind schedule.

Meanwhile Sydney Airport is operating at maximum legislated capacity for most of the day and minor difficulties caused by for example weather or aviation emergencies can have significant knock-on implications and result in extensive delays or even flight cancellations. Just weeks ago while students were studying hard during the mid-semester break from lectures there was a school holiday period and the air traffic control system at Sydney airport became stuck in curfew operations mode severely limiting the number of flights that could be processed during the important morning peak period. Some details about that software difficulty are reported in the reported difficulty the system was stuck in Operating Mode 1 which is described in publicly available Air Services Australia documents as “not practical except during the required Curfew period.” The Curfew period is between 11pm and 6am when very few aircraft movements are permitted and the only other use of the airport is for emergencies.

The media described the software problem in September as “a glitch”. But Software Engineers and software engineering students know that what the public calls “a glitch” implying a strange event that is not expected to reoccur is in fact evidence of a software fault that could result unpredictably in another software failure at any time.

The idea

Sydney airport operates under special legislative arrangements set out in the Sydney Airport Long Term Operating Plan LTOP. This means that simply importing standard air traffic control and planning systems will not work.

Sydney airport has 10 principal modes of operation numbered 1 to 14A – some proposed modes have been given numbers but are not used. In some sense there are 10 different air traffic control systems required.

The client specialises in distributed systems often using cloud based services to provide scalable software solutions for a wide range of businesses including logistics businesses business which have some needs that are similar to aircraft movement planning an important part of air traffic control. They propose to set up not in the cloud but on Air Services Australia property at Sydney Airport at least ten virtual machines with their standard switching technology. Each virtual machine can host a copy of the airport’s current air traffic control software locked into one of the ten currently used modes and the well-tested switching software can be used to connect controllers’ consoles and radar and transponder inputs to the system running the desired mode at any particular point in time.

This means that the widely tested currently effective software can be used but the known fault with unknown location in that very substantial software system can be by-passed because no single instantiation of the software will ever need to change mode.

The first priority is to get the system of virtual machines and their inputs outputs and interconnections properly specified and tested. As an extension of the work the client will consider also hosting the system in its own cloud-based systems where high speed integration with the long term operations of Brisbane Centre and Melbourne Centre the two bases for long distance air traffic control in Australian Airspace could be arranged and Air Services Australia’s IT costs substantially reduced. This second phase is more speculative but the client believes that if it could demonstrate its feasibility it would open up many new opportunities to use scalable cloud-based systems in air traffic control and the advantages if demonstrated could lead to the displacement of the currently under development DSTG system by higher quality services provided by the client.

Your responsibility

The client has sent a representative to you to ask you to plan the software requirements for the new system.