Different Types of Software Systems

Types of Software Systems

Software systems can be classified by their purpose and the environment in which they operate. At A2 level, you need to understand four main categories.

Type	Purpose	Examples
Safety-related systems	Systems where failure could endanger life or cause significant damage	Aircraft autopilot, medical devices, nuclear reactor control
Control systems	Systems that manage physical processes using sensors and actuators	Central heating, traffic lights, industrial robots
Expert systems	Systems that emulate human expert decision-making in a specific domain	Medical diagnosis, financial planning, fault diagnosis
Information exchange systems	Systems for sharing and accessing information	The Internet, intranets, search engines, email

---

Safety-Related Systems

Safety-Critical vs Safety-Related

Term	Definition
Safety-critical	Failure could directly cause death, serious injury, or major environmental damage. The system must work correctly.
Safety-related	The system contributes to safety but is not the sole means of protection. Failure increases risk but other safeguards may exist.

Examples of safety-critical systems:

• Aircraft flight control systems
• Medical infusion pumps and ventilators
• Railway signalling systems
• Nuclear reactor shutdown systems
• Automotive braking systems (ABS, autonomous emergency braking)

High Dependability Requirements

Safety-critical systems must meet exceptionally high standards across several properties:

Property	Description
Reliability	The system must operate correctly for extended periods without failure. Often measured as Mean Time Between Failures (MTBF).
Availability	The system must be operational and accessible whenever needed. Expressed as a percentage (e.g. 99.999% uptime = ~5 minutes downtime per year).
Safety	The system must not cause harm to people or the environment, even in the event of failure.
Security	The system must resist deliberate attacks and unauthorised access.
Maintainability	The system must be easy to repair and update without compromising safety.

Techniques for Achieving Dependability

Technique	Description
Redundancy	Duplicate critical components so that if one fails, another takes over. Triple modular redundancy (TMR) uses three systems and a voting mechanism.
Fail-safe design	If the system fails, it defaults to a safe state (e.g. traffic lights go to all-red, not all-green).
Formal methods	Use mathematical models and proofs to verify that the system meets its specification. Eliminates ambiguity.
Diversity	Use different hardware, software, or algorithms for redundant components so a single bug cannot affect all copies.
Watchdog timers	A hardware timer that resets the system if the software stops responding (indicating a crash or hang).
Defensive programming	Extensive input validation, error handling, and assertions in the code.

Development of Safety-Critical Systems

Developing safety-critical systems is a specialised field with requirements beyond standard software engineering:

Aspect	Requirement
Standards	Must comply with domain-specific standards (e.g. DO-178C for aviation, IEC 61508 for industrial, ISO 26262 for automotive)
Certification	Independent third-party assessment and certification before deployment
Documentation	Extensive documentation of requirements, design decisions, test results, and risk assessments
Testing	Exhaustive testing including formal verification, simulation, and real-world testing. 100% code coverage may be required.
Traceability	Every requirement must be traceable through design, implementation, and testing
Review	Multiple levels of peer review, including independent safety assessment
Lifecycle	Strict adherence to development lifecycle models with defined review gates

---

Control Systems

Nature and Scope of Computer Control

A control system uses computers to manage and regulate physical processes. The basic structure involves:

1. Sensors measure the current state of the physical process (temperature, pressure, speed, position)
2. Processor compares the measured value against the desired value (set point)
3. Actuators make physical adjustments based on the processor’s output (open valves, adjust motors, switch heaters)

Open-Loop vs Closed-Loop Control

Feature	Open-Loop	Closed-Loop
Feedback	No feedback — output is not measured	Uses feedback — output is measured and compared to the set point
Accuracy	Less accurate — cannot correct for disturbances	More accurate — continuously adjusts based on measured error
Complexity	Simpler and cheaper	More complex
Example	A toaster with a timer (runs for set time regardless of toast colour)	A thermostat (measures temperature and adjusts heating)

Feedback Loop

The feedback loop in a closed-loop control system works as follows:

1. The sensor reads the current value of the controlled variable
2. The comparator calculates the error = set point − measured value
3. The controller determines the appropriate response based on the error
4. The actuator adjusts the physical process
5. The cycle repeats continuously

Types of Control Systems

Type	Description	Examples
Embedded systems	Computer built into a device to control its operation	Washing machine, microwave, car engine management
SCADA	Supervisory Control and Data Acquisition — large-scale industrial control with remote monitoring	Power stations, water treatment, oil pipelines
PLCs	Programmable Logic Controllers — rugged industrial computers for real-time control	Manufacturing lines, conveyor systems, robotic arms
Real-time systems	Must respond to inputs within guaranteed time limits	ABS braking, pacemakers, air traffic control

Benefits and Implications of Automation

Benefits:

Benefit	Description
Consistency	Automated processes produce identical results every time, reducing defects
Speed	Computers operate faster than humans for repetitive tasks
24/7 operation	Automated systems can run continuously without breaks
Dangerous environments	Robots can work in hazardous conditions (radiation, extreme heat, toxic chemicals)
Precision	Computer-controlled machines achieve accuracy beyond human capability
Cost reduction	Lower long-term labour costs once the initial investment is recovered

Implications:

Implication	Description
Job displacement	Automation replaces human workers in manufacturing, logistics, and service industries
Deskilling	Remaining workers may need fewer skills as machines handle complex tasks
Dependency	Over-reliance on automated systems creates vulnerability if they fail
Initial cost	High capital investment for equipment, installation, and programming
Maintenance	Requires specialised technical staff to maintain and repair automated systems
Ethical concerns	Questions about accountability when automated systems cause harm
Retraining	Workers displaced by automation need retraining for new roles

---

Expert Systems

What is an Expert System?

An expert system is a computer program that emulates the decision-making ability of a human expert in a specific, narrow domain.

Components of an Expert System

Component	Function
Knowledge base	Stores facts about the domain and IF-THEN rules that represent expert knowledge. For example: IF temperature > 38°C AND cough = persistent THEN suggest = pneumonia (confidence 0.7)
Inference engine	The reasoning mechanism. Applies rules to known facts to derive new conclusions. Uses forward chaining (data-driven: start with facts, apply rules to reach a conclusion) or backward chaining (goal-driven: start with a hypothesis, work backwards to find supporting facts).
User interface	Allows users to input information (answer questions) and receive advice. Should be clear and accessible to non-experts.
Explanation facility	Explains how and why the system reached its conclusion. Essential for user trust and for verifying correctness.
Knowledge acquisition facility	Tools for domain experts to add, modify, and validate rules in the knowledge base.

Expert System Shell

An expert system shell is a generic framework that provides the inference engine, user interface, and explanation facility without any domain-specific knowledge. A knowledge base is then added to create a complete expert system for a particular domain.

Knowledge Acquisition

The process of extracting knowledge from human experts is the knowledge engineering bottleneck:

• Interviews with domain experts
• Observation of experts at work
• Analysis of case studies and documentation
• The knowledge engineer translates expert knowledge into IF-THEN rules
• Challenge: experts often cannot explicitly articulate their reasoning (tacit knowledge)

Effects on Professional Groups

Professional Group	Effects
Doctors	Expert systems can assist with diagnosis, suggest treatments, and flag drug interactions. May improve accuracy but raises concerns about liability and over-reliance. Does not replace clinical judgement.
Lawyers	Legal expert systems can search case law, assess case strength, and suggest relevant precedents. Makes legal advice more accessible but cannot handle nuanced interpretation.
Financial advisers	Can assess risk profiles and recommend investment portfolios. Democratises financial advice but may not account for unusual circumstances.
Engineers	Fault diagnosis systems help identify problems in complex equipment. Speeds up maintenance but requires accurate knowledge base.

Effects on Wider Community

Effect	Description
Accessibility	Makes expert-level advice available in areas with few human experts (rural healthcare, developing countries)
Consistency	Always applies rules the same way — no bad days, bias, or fatigue
Cost reduction	Cheaper than consulting a human expert for routine decisions
Limitations	Cannot handle situations outside its knowledge base, lacks common sense, cannot learn from experience (traditional systems)
Trust	People may over-trust or under-trust automated advice

---

Internet and Intranet

Internet vs Intranet

Feature	Internet	Intranet
Access	Public — anyone can connect	Private — restricted to an organisation’s members
Scope	Global	Local (within one organisation)
Security	Varies widely	Protected by firewalls and access controls
Content	General public information	Organisation-specific resources (policies, internal tools, documents)
Protocols	TCP/IP, HTTP/HTTPS	Same protocols but within a private network

How Search Engines Work

Search engines use three main processes to index and retrieve web content:

1. Web Crawling (Spidering)

• Automated programs called web crawlers (or spiders/bots) systematically browse the web
• They follow hyperlinks from page to page, discovering new and updated content
• Crawlers respect robots.txt files that tell them which pages not to crawl
• They periodically re-crawl pages to detect changes

2. Indexing

• The crawler sends discovered pages to the indexer
• The indexer analyses the content: words, their positions, headings, links, images, metadata
• An inverted index is built — mapping each word to the list of pages containing it
• This allows fast lookup: given a search term, instantly find all pages containing it

3. Ranking

• When a user searches, the engine retrieves matching pages from the index
• Pages are ranked by relevance using algorithms that consider hundreds of factors

Key ranking factors:

Factor	Description
PageRank	Measures page importance based on the number and quality of links pointing to it. A page linked to by many high-quality pages ranks higher.
Content relevance	How well the page content matches the search query (keyword frequency, placement in headings)
Freshness	Recently updated content may rank higher for time-sensitive queries
User signals	Click-through rates, time spent on page, bounce rate
Mobile-friendliness	Pages optimised for mobile devices rank higher in mobile searches
Page speed	Faster-loading pages are preferred
HTTPS	Secure pages may receive a small ranking boost

---

Common Contemporary Applications

Application	Description	Key Technologies
E-commerce	Online buying and selling of goods and services	HTTPS, payment gateways, databases, recommendation engines
Social media	Platforms for user-generated content and social interaction	Real-time messaging, content delivery networks, AI moderation
Cloud computing	On-demand access to computing resources over the Internet	Virtualisation, distributed storage, APIs, pay-as-you-go pricing
Internet of Things (IoT)	Network of physical devices with embedded sensors and connectivity	Wireless protocols, edge computing, sensor networks
Streaming services	Real-time delivery of audio/video content	Adaptive bitrate streaming, CDNs, compression codecs
Online banking	Financial services delivered via the Internet	Encryption, two-factor authentication, secure APIs
Telehealth	Remote medical consultations and monitoring	Video conferencing, wearable sensors, electronic health records
Remote working	Working from any location using Internet-connected tools	VPNs, collaboration platforms, cloud storage

---

Effects on Professional Groups and the Wider Community

Effects on Professional Groups

Group	Positive Effects	Negative Effects
Healthcare	Telemedicine reaches remote patients, AI assists diagnosis, electronic records improve coordination	Data security concerns, digital literacy requirements, risk of misdiagnosis via remote consultation
Education	Online learning widens access, digital resources enhance teaching, personalised learning	Digital divide excludes some students, screen fatigue, reduced social interaction
Legal	Online legal databases speed up research, video conferencing enables remote hearings	Cybercrime creates new legal challenges, jurisdiction issues across borders
Journalism	Instant news distribution, citizen journalism, data journalism	Fake news, declining revenue from print, information overload
Retail	E-commerce opens global markets, data analytics improves targeting	High street decline, returns and fraud costs, competitive pressure

Effects on Wider Community

Area	Effects
Communication	Instant global communication via email, messaging, and video calls. But risk of isolation from excessive screen time.
Information access	Vast information freely available. But reliability varies — misinformation spreads easily.
Employment	New job types created (content creators, app developers). But traditional roles disrupted (retail workers, bank tellers).
Privacy	Online activity generates vast data trails. Companies and governments can track behaviour. Data protection laws attempt to balance this.
Democracy	Social media enables political engagement and activism. But also enables manipulation through targeted advertising and echo chambers.
Culture	Global sharing of cultural content. But risk of cultural homogenisation and dominance of English-language content.
Health	Health information and fitness tracking empower individuals. But screen addiction, cyberbullying, and mental health impacts are growing concerns.