Computing Dissertation Topics 2026 — 80+ Ideas Across Every Technical Area

Computing dissertation topics fail most commonly at the execution stage topics that are technically interesting in principle but impossible to implement, evaluate, or empirically study within the word count and timeline available. "A neural network for detecting cybersecurity threats in real time" sounds compelling until the scope question arrives: trained on what data, evaluated against what baseline, implemented in what infrastructure, and completed within how many weeks? The gap between an interesting technical idea and a researchable dissertation topic is almost always the specificity and feasibility of the execution plan.

Three-part structure: problem, method, and evaluation

The most consistently researchable computing dissertation topics are built around three distinct elements. The problem is the specific technical question being investigated not "machine learning in healthcare" but "whether convolutional neural networks outperform support vector machines in classifying skin lesion images from the ISIC 2019 dataset." The method is the specific technical approach being used implementation, literature based evaluation, comparative benchmarking, or empirical study of existing systems. The evaluation is how you'll know whether the approach worked accuracy metrics, performance benchmarks, usability scores, or statistical comparison against a baseline. A topic that has all three specified before the literature review begins is considerably more likely to produce a focused, defensible dissertation than one that leaves evaluation criteria implicit.

Implementation versus evaluation: the most important early decision

Computing dissertations divide into two broad types: those that involve building or implementing something new (a system, application, algorithm, or prototype) and those that evaluate, compare, or empirically study existing systems. Both are academically legitimate and UK examiners assess both types on rigour of execution rather than technical ambition. An evaluation study that rigorously benchmarks three existing deep learning architectures on a standardised dataset, analyses the results systematically, and discusses their implications for a specific application domain is often stronger than an implementation study that builds something novel but evaluates it superficially. Knowing which type your programme typically expects and which type fits your technical skills and available resources is worth confirming with your supervisor before committing to a topic.

Cloud computing dissertation topics are among the most data rich in computing research, since the major cloud providers (AWS, Microsoft Azure, Google Cloud Platform) publish benchmark data, pricing information, and technical architecture documentation that provides accessible secondary data. Vendor lock in risk research connects to the NIST Cloud Computing definition and the academic literature on cloud portability examining how dependency on cloud provider specific APIs, data formats, and services creates switching costs that limit competition. Serverless computing research connects to the Function as a Service (FaaS) literature and to the published benchmark studies (SPEC Cloud benchmarks, the CNCF serverless whitepaper) that provide standardised performance comparison data. Multi tenant cloud security research has a specific and developed academic literature following high profile cloud data breaches the Spectre and Meltdown vulnerabilities (2018) demonstrated that hardware level security assumptions underlying multi tenant isolation required fundamental reassessment, and the subsequent academic literature on hardware security and hypervisor vulnerabilities provides a well developed theoretical base for cloud security dissertations.

• The reliability of cloud infrastructure for traffic management in rural banking

• Cost efficiency trade offs between public, private, and hybrid cloud deployment

• Security risks in multi tenant cloud architectures

• Serverless computing's impact on application scalability and cost

• Cloud migration strategies for legacy enterprise systems

• Load balancing techniques in distributed cloud environments

• Cloud computing's role in supporting remote and hybrid work infrastructure

• Evaluating vendor lock in risk across major cloud service providers

Quantum computing dissertation topics are unusual in that the gap between theoretical possibility and practical implementation is still very large which creates a specific challenge for computing students who want to write about quantum computing but can't actually run experiments on quantum hardware. The most productive dissertation approaches for quantum computing topics are therefore either literature based comparative evaluation (comparing the theoretical properties of different quantum algorithms or error correction schemes using the academic literature), simulation based (using IBM Qiskit, Google Cirq, Microsoft Q#, or Amazon Braket simulator environments to implement and test quantum circuits without physical hardware access), or industry and policy analysis (examining the practical barriers to quantum computing adoption in specific sectors using Gartner Hype Cycle data, National Quantum Initiative programme reports, or the UK National Quantum Strategy published in 2023). Quantum cryptography research connects directly to the current real world concern about post quantum cryptography NIST's Post Quantum Cryptography standardisation project selected its first quantum resistant algorithms in 2022 (CRYSTALS Kyber for key exchange, CRYSTALS Dilithium for digital signatures), making the transition timeline and implementation challenges an immediately relevant research topic.

• Quantum simulation approaches to molecular modelling

• Comparing quantum and classical algorithms for optimisation problems

• The current state of quantum error correction techniques

• Practical barriers to quantum computing adoption in industry

• Quantum cryptography and its implications for existing security standards

• Evaluating near term quantum hardware limitations for real world applications

Cybersecurity dissertation topics have particularly strong empirical data availability the CICIDS2017 (Canadian Institute for Cybersecurity Intrusion Detection System) dataset, the NSL KDD dataset, and the UNSW NB15 dataset are all publicly available labelled datasets for network intrusion detection research, making machine learning based phishing and intrusion detection evaluation studies well suited to undergraduate and Masters dissertations without requiring access to production security environments. Zero trust architecture research connects to NIST SP 800 207 (the foundational zero trust architecture standard published in 2020) and to the growing enterprise adoption literature following the White House's 2021 executive order requiring federal agencies to adopt zero trust principles. GDPR compliance research for UK organisations connects to the ICO's enforcement notices (which provide documented real world compliance failure case studies), the NCSC's Cyber Security Breaches Survey (published annually, providing sector level data on attack types and organisational response), and the UK GDPR's post Brexit operational context. Ransomware research has a particularly current data environment: the NCSC, Verizon's annual Data Breach Investigations Report, and IBM Security's Cost of a Data Breach report all provide annual quantitative data on attack patterns, sector distribution, and financial impact.

• Evaluating phishing detection techniques using machine learning

• The effectiveness of multi factor authentication against modern attack vectors

• GDPR compliance challenges for organisations handling cross border data

• Ransomware attack patterns and organisational response strategies

• Zero trust security architecture adoption in enterprise networks

• Vulnerability assessment methodologies for IoT device ecosystems

• The human factor in cybersecurity: social engineering resistance training

• Privacy preserving techniques in large scale data collection

Information systems and e learning dissertation topics sit at the intersection of computing and organisational theory drawing on both technical literature (information systems architecture, database design, learning management systems) and social science frameworks (knowledge management theory, technology adoption models, organisational learning). Knowledge management research draws on Nonaka and Takeuchi's SECI model (socialisation, externalisation, combination, internalisation) as the foundational knowledge creation framework, and on the more recent literature on enterprise social software and communities of practice. Learning analytics research connects to the ACM Learning @ Scale conference proceedings and to the Journal of Learning Analytics, examining how data generated within learning management systems (click data, time on task, forum participation, quiz performance) can be used to personalise learning recommendations and identify at risk students. E recruitment platform research connects to the academic HCI literature on algorithmic hiring, the FCA's guidance on AI in employment decisions, and the UK Equality Act 2010's implications for automated screening tools that may perpetuate historical hiring biases.

• Using IT to improve inter organisational knowledge management

• Building an effective information system for e learning delivery

• Collaborative social network tools for information gathering and classification

• E recruitment platform standards: current challenges and future directions

• The role of scientific innovation within the broader information society

• Internet based services and their impact on e marketing and digital service delivery

• Comparing e marketing service models built on network infrastructure

• Learning analytics and their effectiveness in personalising e learning outcomes

Operating systems and systems architecture dissertation topics connect the foundational computer science theory of OS design to current deployment contexts real time operating systems in safety critical embedded applications, containerisation and orchestration in cloud native deployments, and memory management in resource constrained IoT and edge computing devices. Real time OS performance in safety critical applications connects to the IEC 61508 functional safety standard and the academic literature on worst case execution time (WCET) analysis both of which are expected reference points in any dissertation examining RTOS for automotive, avionics, or medical device applications. Container orchestration platform comparison (Kubernetes versus Docker Swarm versus Apache Mesos) connects to the Cloud Native Computing Foundation (CNCF) published survey data on orchestration platform adoption and to the academic benchmarking literature in ACM/IEEE conference proceedings. Operating system security hardening for critical infrastructure connects to the NCSC's Cyber Assessment Framework (CAF) and the US CISA's critical infrastructure security guidance both providing authoritative secondary sources for policy oriented OS security dissertations.

• Operating system architecture advantages for long term medical settings use

• Real time operating system performance in safety critical applications

• Virtualisation technology's impact on system resource efficiency

• Comparing container orchestration platforms for enterprise deployment

• Memory management strategies in resource constrained embedded systems

• Operating system security hardening for critical infrastructure

Most UK computer science programmes expect a consistent chapter structure, and skipping or merging sections tends to cost marks for clarity even when the underlying technical work is strong. The chapter structure below is the standard expected by UK computing examiners, with the specific content requirements for each section.

• 📄TitleSpecific and clear, reflecting your actual technical aim and research area. "A Study of Machine Learning" is not a title. "Comparative Evaluation of BERT and Logistic Regression for Clinical Text Classification on the MIMIC-III Dataset" is one. The title should communicate the method, the application domain, and the data or system if applicable.

• 📝AbstractA concise summary (200–300 words) of the whole dissertation: research question, method, implementation or data, key findings, and main conclusions. Written last, after the rest of the dissertation is complete, so it accurately describes what was done rather than what was planned.

• 🔍IntroductionStates the research aim and objectives, provides a brief technical context explaining why the problem is important and under addressed, and outlines what the dissertation will cover chapter by chapter. Should establish both the technical significance (what gap does the research address?) and the practical relevance (who benefits from the answer and why?).

• 📚Literature ReviewCritical engagement with existing peer reviewed work identifying where the field has established consensus, where open questions remain, and where the research gap your dissertation addresses sits. Not a catalogue of what papers say, but an analysis of what the field knows, what it disagrees about, and what it hasn't yet adequately investigated. IEEE Xplore and ACM Digital Library are the primary databases; arXiv for very recent work in fast-moving areas like ML. Cite specific conference papers and journal articles, not Wikipedia or general web sources.

• 🔬MethodologyYour research method, data collection or implementation process, and evaluation approach. For implementation dissertations: system design, architecture decisions, and development tools. For evaluation dissertations: dataset selection and justification, baseline selection and justification, evaluation metrics and their appropriateness for the research question, and any statistical analysis approach. Every choice needs justifying against the research question methodology is not a description of what you did but an argument for why this was the right approach.

• 📊ResultsFindings presented clearly performance metrics, benchmark results, usability scores, statistical outputs without interpretation. Tables and figures correctly formatted. The results chapter presents what was found; the discussion chapter explains what it means. Keeping these separate is a structural convention that computing examiners check, and merging them costs marks.

• 💬DiscussionConnects results to the literature review explicitly comparing your findings to what prior work found and explaining agreements, disagreements, and surprising outcomes. Addresses limitations of the study with technical specificity (what does the evaluation metric not capture? what would a larger dataset change?). Identifies directions for future work grounded specifically in what this dissertation found or couldn't determine.

• 🔗ReferencesA complete, consistently formatted reference list. IEEE referencing style is standard for engineering and computing dissertations at most UK programmes; Harvard is used at some institutions. Every in-text citation has a corresponding reference entry; no uncited entries appear. ACM and IEEE papers have specific citation formats that differ from standard journal article formats.

• 📎AppendixSupporting material that would disrupt the main text's flow full algorithm pseudocode, raw data tables, system configuration details, ethics documentation, complete statistical output, or implementation code (if the programme allows code in appendices). Each appendix is referenced explicitly in the main text and numbered sequentially.

The two databases that matter most for UK computing dissertation literature reviews are the ACM Digital Library and IEEE Xplore both are the primary peer reviewed literature sources that UK computing examiners expect to see cited, rather than general web sources alone. A computing literature review that doesn't cite ACM or IEEE papers will be noted as having superficial source engagement, regardless of how accurate the content is.

ACM Digital Library covers computer science research published through ACM including proceedings of major conferences like CHI (Human-Computer Interaction), CCS (Computer and Communications Security), SIGGRAPH (computer graphics), PLDI (programming languages), and KDD (knowledge discovery and data mining), alongside journals including the ACM Transactions series. IEEE Xplore covers engineering and computing research through IEEE including IEEE Transactions journals (on Software Engineering, on Neural Networks and Learning Systems, on Information Forensics and Security, on Knowledge and Data Engineering) and conference proceedings from ICSE (software engineering), CVPR (computer vision), SP (security and privacy), and INFOCOM (networking).

For very recent AI and machine learning work, arXiv (particularly the cs.AI, cs.LG, cs.CL, and cs.CV categories) publishes preprints that often appear months before the peer reviewed version is accepted and indexed in ACM or IEEE. arXiv papers are acceptable dissertation sources and are widely cited in ML research but should be balanced with peer reviewed publications, since preprints haven't been through formal review. When an arXiv paper has a corresponding published version in a conference or journal, cite the published version rather than the preprint.