{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,21]],"date-time":"2026-10-21T21:39:11Z","timestamp":1792618751110,"version":"4.5.0"},"reference-count":124,"publisher":"Association for Computing Machinery (ACM)","issue":"4","license":[{"start":{"date-parts":[[2026,10,16]],"date-time":"2026-10-16T00:00:00Z","timestamp":1792108800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/legalcode"}],"content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":["Form. Asp. Comput."],"published-print":{"date-parts":[[2026,12,31]]},"abstract":"<jats:p>Formal methods have long promised enhanced software reliability through rigorous mathematical reasoning, yet concerns over scalability, cost, and accessibility have historically limited their industrial adoption. This article critically evaluates the public record on the use of formal methods in engineering at Amazon Web Services (AWS). Drawing on publicly available case studies, engineering reports, and technical documentation, we examine reported applications of model checking, SMT solving, and lightweight formal methods across domains such as distributed systems, cloud networking, cryptography, and security policy verification. We organise the analysis around 16 targeted questions across five themes: (1) technical impact, (2) organisational adoption, (3) scalability in complex and heterogeneous systems, (4) long-term strategic value, and (5) future directions. By placing these AWS case studies alongside historical precedents at companies such as IBM, Microsoft, and Google, we highlight both continuity and innovation in industrial practice. Across the published case studies, authors report benefits such as early bug detection, performance optimisation, and strengthened security assurance, including integration with CI\/CD workflows in some settings. However, the evidence base is selective and does not support organisation-wide claims about prevalence, cost-effectiveness, or ROI; quantitative baselines are sparse, and survivorship bias is plausible. We discuss the resulting methodological limits, highlight challenges such as maintaining specifications under continuous evolution, and outline a research agenda to improve empirical validation, methodological refinement, and the preservation of industrial verification artefacts.<\/jats:p>","DOI":"10.1145\/3815784","type":"journal-article","created":{"date-parts":[[2026,6,8]],"date-time":"2026-06-08T21:00:47Z","timestamp":1780952447000},"page":"1-73","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Formal Methods in Industry: A Critical Evaluation of Their Use at Amazon Web Services"],"prefix":"10.1145","volume":"38","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-7955-2702","authenticated-orcid":false,"given":"James","family":"Woodcock","sequence":"first","affiliation":[{"name":"College of Software, Southwest University","place":["Chongqing, China"]},{"name":"Department of Electrical and Computer Engineering, Aarhus University","place":["Chongqing, China"]},{"name":"Computer Science, University of York","place":["Chongqing, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4589-1500","authenticated-orcid":false,"given":"Peter Gorm","family":"Larsen","sequence":"additional","affiliation":[{"name":"Electrical and Computer Engineering, Aarhus Universitet Science and Technology","place":["Aarhus, Denmark"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"320","published-online":{"date-parts":[[2026,10,16]]},"reference":[{"key":"e_1_3_4_2_2","doi-asserted-by":"publisher","DOI":"10.1017\/CBO9780511624162"},{"key":"e_1_3_4_3_2","doi-asserted-by":"publisher","DOI":"10.1145\/503272.503274"},{"key":"e_1_3_4_4_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-642-22110-1_14"},{"key":"e_1_3_4_5_2","doi-asserted-by":"publisher","unstructured":"Clark W. Barrett Roberto Sebastiani Sanjit A. Seshia and Cesare Tinelli. 2009. Satisfiability modulo theories. (2009) 825\u2013885. DOI:10.3233\/978-1-58603-929-5-825. This chapter surveys the field of Satisfiability Modulo Theories (SMT) which extends Boolean satisfiability (SAT) by incorporating background theories such as arithmetic arrays bit-vectors and uninterpreted functions. The authors present the foundations of SMT solving including decision procedures DPLL(T)-based frameworks and combination methods for multiple theories. They review major SMT solvers their architecture and optimisations that enable scalability to large industrial problems. Applications range from hardware and software verification to model checking program analysis and test generation. The paper highlights theoretical advances practical implementations and open research challenges establishing SMT as a cornerstone of modern automated reasoning.","DOI":"10.3233\/978-1-58603-929-5-825"},{"key":"e_1_3_4_6_2","doi-asserted-by":"publisher","DOI":"10.3233\/978-1-58603-929-5-825"},{"key":"e_1_3_4_7_2","doi-asserted-by":"publisher","DOI":"10.1145\/3689374"},{"key":"e_1_3_4_8_2","doi-asserted-by":"publisher","DOI":"10.1007\/3-540-08766-4"},{"key":"e_1_3_4_9_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-76554-4_2"},{"key":"e_1_3_4_10_2","doi-asserted-by":"publisher","DOI":"10.1080\/01621459.1976.10480949"},{"key":"e_1_3_4_11_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-21690-4_33"},{"key":"e_1_3_4_12_2","doi-asserted-by":"publisher","DOI":"10.1145\/3729175"},{"key":"e_1_3_4_13_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-58298-2_8"},{"key":"e_1_3_4_14_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-642-14203-1_12"},{"key":"e_1_3_4_15_2","volume-title":"Model Checking","author":"Clarke Edmund M.","year":"1999","unstructured":"Edmund M. Clarke, Orna Grumberg, and Doron A. Peled. 1999. Model Checking. MIT Press. Retrieved from http:\/\/books.google.de\/books?id=Nmc4wEaLXFEC. This foundational textbook presents model checking as an automated technique for verifying finite-state concurrent systems against temporal logic specifications. It introduces the theoretical basis, including temporal logics (LTL, CTL), automata theory, and fixpoint characterisations, as well as algorithmic approaches to state exploration. The book addresses state explosion through abstraction, symmetry reduction, and symbolic methods using BDDs. Practical applications include hardware and communication protocol verification. Numerous examples and case studies demonstrate how model checking systematically detects design errors early in the development process. The text established model checking as a cornerstone of formal verification and remains an authoritative reference for researchers and practitioners."},{"key":"e_1_3_4_16_2","doi-asserted-by":"publisher","DOI":"10.1145\/242223.242257"},{"key":"e_1_3_4_17_2","volume-title":"The CMMI Institute Announces CMMI Development V2.0","author":"Institute CMMI","year":"2018","unstructured":"CMMI Institute. 2018. The CMMI Institute Announces CMMI Development V2.0. Retrieved from https:\/\/cmmiinstitute.com\/news\/press-releases\/march-2018\/annoucingv2. Press release announcing CMMI Development V2.0."},{"key":"e_1_3_4_18_2","doi-asserted-by":"publisher","DOI":"10.1145\/2491245"},{"key":"e_1_3_4_19_2","doi-asserted-by":"publisher","DOI":"10.3233\/JCS-2009-0340"},{"key":"e_1_3_4_20_2","doi-asserted-by":"publisher","DOI":"10.1145\/2491956.2462184"},{"key":"e_1_3_4_21_2","doi-asserted-by":"publisher","DOI":"10.1145\/2491956.2462184"},{"key":"e_1_3_4_22_2","doi-asserted-by":"publisher","DOI":"10.1145\/3357231"},{"key":"e_1_3_4_23_2","article-title":"Formal methods: Oversold? underused? a survey","volume":"1812","author":"Gleirscher Mario","year":"2018","unstructured":"Mario Gleirscher and Diego Marmsoler. 2018. Formal methods: Oversold? underused? a survey. CoRR abs\/1812.08815 (2018). arXiv:1812.08815. http:\/\/arxiv.org\/abs\/1812.08815. The paper presents a survey on the use of formal methods (FMs) in mission-critical software. It finds that while industry sees FMs as useful, ease of use is a challenge, with scalability, skills, and education as key barriers. The study, based on 192 responses, suggests a growing intention to adopt FMs, highlighting the need for improved education, tools, and empirical evidence to enhance adoption.","journal-title":"CoRR"},{"key":"e_1_3_4_24_2","doi-asserted-by":"publisher","DOI":"10.1007\/S10270-023-01124-2"},{"key":"e_1_3_4_25_2","doi-asserted-by":"publisher","DOI":"10.1145\/3477355.3477367"},{"key":"e_1_3_4_26_2","doi-asserted-by":"publisher","DOI":"10.1007\/3-540-54834-3_34"},{"key":"e_1_3_4_27_2","doi-asserted-by":"publisher","DOI":"10.48550\/ARXIV.2301.02206"},{"key":"e_1_3_4_28_2","unstructured":"Laura Humphrey. 2019. Formal Methods for V&V and Certification. Retrieved from https:\/\/cra.org\/ccc\/wp-content\/uploads\/sites\/2\/2019\/10\/Laura-Humphrey.pdf. This presentation discusses the use of formal methods for verification validation and certification of complex and autonomous systems. It highlights the growing software complexity that renders traditional V&V approaches infeasible and explains formal methods such as theorem proving model checking and abstract interpretation. The talk emphasises scalability usability and integration challenges explores hybrid systems and runtime assurance as alternatives and presents successful industry applications in aerospace cybersecurity and cloud computing. Future directions include automating formal methods enhancing education and improving cybersecurity for autonomous systems."},{"key":"e_1_3_4_29_2","doi-asserted-by":"publisher","DOI":"10.1109\/52.896251"},{"key":"e_1_3_4_30_2","doi-asserted-by":"publisher","DOI":"10.1109\/MC.1996.10038"},{"key":"e_1_3_4_31_2","doi-asserted-by":"publisher","DOI":"10.1109\/MC.2006.145"},{"key":"e_1_3_4_32_2","doi-asserted-by":"publisher","DOI":"10.1109\/54.936245"},{"key":"e_1_3_4_33_2","doi-asserted-by":"publisher","DOI":"10.1145\/1743546.1743574"},{"key":"e_1_3_4_34_2","doi-asserted-by":"publisher","DOI":"10.1145\/3522582"},{"key":"e_1_3_4_35_2","doi-asserted-by":"publisher","DOI":"10.4204\/EPTCS.310.6"},{"key":"e_1_3_4_36_2","volume-title":"Specifying Systems, The TLA+ Language and Tools for Hardware and Software Engineers","author":"Lamport Leslie","year":"2002","unstructured":"Leslie Lamport. 2002. Specifying Systems, The TLA+ Language and Tools for Hardware and Software Engineers. Addison-Wesley. Retrieved fromhttp:\/\/research.microsoft.com\/users\/lamport\/tla\/book.html. This book introduces TLA+, Leslie Lamport\u2019s formal specification language for describing and reasoning about concurrent and distributed systems. TLA+ unifies temporal logic with set theory to provide a precise framework for specifying system behaviour, safety, and liveness properties. The text emphasises practical engineering, showing how to write clear specifications, model complex algorithms, and reason about correctness using the TLC model checker. Numerous examples from hardware and software illustrate the approach, including cache coherence protocols and distributed algorithms. The book established TLA+ as a powerful tool for engineers to design reliable systems with mathematical rigour and tool support."},{"key":"e_1_3_4_37_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-642-03466-4_2"},{"key":"e_1_3_4_38_2","volume-title":"Formal Methods in Cloud Security","author":"Le Michael","year":"2021","unstructured":"Michael Le. 2021. Formal Methods in Cloud Security. Retrieved from https:\/\/www.ibm.com\/blogs\/research\/2021\/05\/formal-methods-in-cloud-security\/. A practitioner-oriented IBM Research blog post on the role of formal methods in cloud security. It is useful as non-AWS industrial evidence that major cloud providers view formal verification as relevant to high-assurance cloud infrastructure, security policy, isolation, and system correctness."},{"key":"e_1_3_4_39_2","doi-asserted-by":"publisher","DOI":"10.1007\/S10817-009-9155-4"},{"key":"e_1_3_4_40_2","unstructured":"Belle Lin. 2025. Why Amazon Is Betting on \u2018Automated Reasoning\u2019 to Reduce AI\u2019s Hallucinations. Retrieved from https:\/\/www.wsj.com\/articles\/why-amazon-is-betting-on-automated-reasoning-to-reduce-ais-hallucinations-b838849e. This Wall Street Journal article provides press coverage of AWS\u2019s use of automated reasoning to reduce hallucinations in generative-AI applications. It is useful as an external non-AWS source that shows how automated reasoning has entered broader business and technology discourse especially through Amazon Bedrock Guardrails and Automated Reasoning Checks. The article is particularly relevant because it presents both the commercial motivation making enterprise AI safer for regulated and high-value use cases and the limitations: automated reasoning can mitigate hallucinations where policies and rules can be precisely specified but it is not a general solution to all LLM errors."},{"key":"e_1_3_4_41_2","unstructured":"Rafael S\u00e1 Menezes Edoardo Manino Fedor Shmarov Mohannad Aldughaim Rosiane de Freitas and Lucas C. Cordeiro. 2024. Interval analysis in industrial-scale BMC software verifiers: A case study. arXiv:2406.15281. Retrieved from https:\/\/arxiv.org\/abs\/2406.15281. Evaluates the effectiveness of interval analysis in improving Bounded Model Checking (BMC) for software verification. The authors implement interval analysis in ESBMC and measure its impact on benchmarks that include Intel Core Power Management firmware and SV-COMP\u201923 ReachSafety programs. Results show that interval analysis reduces verification complexity enabling more benchmarks to be solved though at an increased computational cost. The study explores precision trade-offs by comparing integer and wrapped-interval domains and demonstrates that lightweight interval analysis enhances BMC performance while mitigating state-space explosion."},{"key":"e_1_3_4_42_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-540-78800-3_24"},{"key":"e_1_3_4_43_2","doi-asserted-by":"publisher","DOI":"10.1016\/J.CL.2017.04.001"},{"key":"e_1_3_4_44_2","doi-asserted-by":"publisher","DOI":"10.1145\/3209108.3209109"},{"key":"e_1_3_4_45_2","doi-asserted-by":"publisher","DOI":"10.1145\/3623652.3623668"},{"key":"e_1_3_4_46_2","doi-asserted-by":"publisher","DOI":"10.21236\/ADA263403"},{"key":"e_1_3_4_47_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICSE.2013.6606572"},{"key":"e_1_3_4_48_2","unstructured":"REF2021. 2021. Impact case study database: Automated Reasoning for Amazon Cloud Security. Retrieved from https:\/\/results2021.ref.ac.uk\/impact\/submissions\/cc7adea4-1b27-4b89-b822-bebb19650fab\/impact. This is an impact case study for the UK Research Evaluation Exercise 2021. It is from University College London (UCL) and highlights the Department of Computer Science\u2019s significant contributions to various sectors through interdisciplinary research and innovation. Key impacts include advancements in artificial intelligence human-computer interaction and cybersecurity which are driving improvements in healthcare finance and public policy. UCL\u2019s collaborative approach and strategic partnerships have facilitated the translation of research into practical applications benefiting both industry and society."},{"key":"e_1_3_4_49_2","volume-title":"The Exponential Cost of Fixing Bugs","year":"2019","unstructured":"Sanket. 2019. The Exponential Cost of Fixing Bugs. Retrieved from https:\/\/deepsource.io\/blog\/exponential-cost-of-fixing-bugs\/. The blog highlights how the cost of fixing software bugs grows exponentially as they progress through the development lifecycle. A defect introduced during requirements or design may cost little to correct at that stage, but if discovered during testing or in production, it can be orders of magnitude more expensive due to rework, patching, and customer impact. Sanket illustrates this with industry data and emphasises practices like code reviews, static analysis, continuous integration, and automated testing to catch errors early. The article emphasises the importance of investing in early defect detection to minimise long-term costs and enhance software quality."},{"key":"e_1_3_4_50_2","volume-title":"Safe and Reliable Software and the Formal Verification of Prim\u2019s Algorithm in SPARK","author":"Wheelhouse Brian S.","year":"2023","unstructured":"Brian S. Wheelhouse. 2023. Safe and Reliable Software and the Formal Verification of Prim\u2019s Algorithm in SPARK. Master\u2019s thesis. Air Force Institute of Technology, Wright-Patterson Air Force Base, Ohio. Retrieved from https:\/\/scholar.afit.edu\/etd\/6945. AFIT-ENG-MS-23-M-070. The author formally verifies examples in SPARK, including implementations of Prim\u2019s algorithm, demonstrating how modern verification tools can support reliable graph-algorithm implementations. The thesis also discusses SPARK and Rust as languages that support safer software development, and contrasts basic verification with more comprehensive functional-correctness proof obligations."},{"key":"e_1_3_4_51_2","doi-asserted-by":"publisher","DOI":"10.1109\/MC.2006.340"},{"key":"e_1_3_4_52_2","doi-asserted-by":"publisher","DOI":"10.1145\/3477355.3477363"},{"key":"e_1_3_4_53_2","doi-asserted-by":"publisher","DOI":"10.5555\/235337"},{"key":"e_1_3_4_54_2","doi-asserted-by":"publisher","DOI":"10.1145\/1592434.1592436"},{"key":"e_1_3_4_55_2","doi-asserted-by":"publisher","DOI":"10.1007\/S00165-007-0060-5"},{"key":"e_1_3_4_56_2","doi-asserted-by":"publisher","DOI":"10.1145\/2854065.2854081"},{"key":"e_1_3_4_57_2","doi-asserted-by":"publisher","DOI":"10.1007\/3-540-48153-2_6"},{"key":"e_1_3_4_58_2","unstructured":"Amazon Science. 2023. Computing on Private Data. Retrieved from https:\/\/www.amazon.science\/blog\/computing-on-private-data. A clear Amazon Science overview distinguishing secure multiparty computation from differential privacy as two approaches to computing on sensitive data. The article is a useful background for AWS\u2019s privacy-preserving computation programme explaining when exact cryptographic computation is appropriate and when noisy statistical protection through differential privacy is more suitable. Accessed: 2025-08-28"},{"key":"e_1_3_4_59_2","article-title":"Formal verification makes RSA Faster \u2014 and Faster to Deploy","author":"Science Amazon","year":"2024","unstructured":"Amazon Science. 2024. Formal verification makes RSA Faster \u2014 and Faster to Deploy. Blog post, Retrieved from https:\/\/www.amazon.science\/blog\/formal-verification-makes-rsa-faster-and-faster-to-deploy. Describes verified and optimised big-number arithmetic used ins2n-bignum, including a constant-time coding discipline to mitigate timing-based side channels.","journal-title":"Blog post"},{"key":"e_1_3_4_60_2","volume-title":"Amazon Science Formal Methods Tag Page","author":"Science Amazon","year":"2024","unstructured":"Amazon Science. 2024. Amazon Science Formal Methods Tag Page. Retrieved from https:\/\/www.amazon.science\/tag\/formal-methods. A curated Amazon Science tag page collecting AWS publications, articles, blog posts, and code resources on formal methods, formal verification, model checking, automated reasoning, and security. It provides useful evidence of the institutional breadth and public visibility of AWS\u2019s formal-methods activity, but should be treated as an index resource rather than as a primary technical publication."},{"key":"e_1_3_4_61_2","volume-title":"Calculating the Cost of Downtime","unstructured":"Atlassian. [n. d.]. Calculating the Cost of Downtime. Retrieved from https:\/\/www.atlassian.com\/incident-management\/kpis\/cost-of-downtime. Amazon Science\u2019s 2024 blog series on Formal Methods and Security highlights AWS\u2019s ongoing efforts to integrate rigorous verification into cloud infrastructure. The posts showcase applications of TLA+, automated reasoning, and model checking across various domains, including distributed storage, cryptography, and access control policies. They describe how these techniques prevent subtle design flaws, improve security guarantees, and support compliance at scale. The series also emphasises cultural aspects, training, knowledge sharing, and tooling that enable widespread adoption across engineering teams. By publishing these accounts, AWS aims to share lessons with industry and academia while demonstrating leadership in trustworthy cloud computing."},{"key":"e_1_3_4_62_2","volume-title":"Incident Management for High-velocity Teams: Calculating the Cost of Downtime","unstructured":"Atlassian. [n. d.]. Incident Management for High-velocity Teams: Calculating the Cost of Downtime. Retrieved from https:\/\/www.atlassian.com\/incident-management\/kpis\/cost-of-downtime"},{"key":"e_1_3_4_63_2","volume-title":"Amazon Verified Permissions is now generally available","author":"Services Amazon Web","year":"2023","unstructured":"Amazon Web Services. 2023. Amazon Verified Permissions is now generally available. Retrieved from https:\/\/aws.amazon.com\/about-aws\/whats-new\/2023\/06\/amazon-verified-permissions-generally-available\/. Announces the general availability of Amazon Verified Permissions, a managed authorisation service using the Cedar policy language."},{"key":"e_1_3_4_64_2","unstructured":"Amazon Web Services. 2023. Cryptographic computing on AWS. Retrieved August 28 2025 from https:\/\/aws.amazon.com\/security\/cryptographic-computing\/"},{"key":"e_1_3_4_65_2","unstructured":"Amazon Web Services and Leidos. 2023. Enable fully homomorphic encryption with Amazon SageMaker endpoints for secure real-time inferencing. Retrieved from https:\/\/aws.amazon.com\/blogs\/machine-learning\/enable-fully-homomorphic-encryption-with-amazon-sagemaker-endpoints-for-secure-real-time-inferencing\/. A joint AWS\u2013Leidos technical blog post showing how fully homomorphic encryption can be combined with Amazon SageMaker endpoints for secure real-time ML inference over encrypted data. It is useful evidence of AWS\u2019s practical interest in privacy-preserving AI and confidential cloud analytics especially for sensitive sectors such as defence intelligence healthcare and regulated enterprise computing. Accessed: 2025-08-28."},{"key":"e_1_3_4_66_2","volume-title":"Introducing Cedar, an Open-source Language for Access Control","author":"Services Amazon Web","year":"2023","unstructured":"Amazon Web Services. 2023. Introducing Cedar, an Open-source Language for Access Control. Retrieved from https:\/\/aws.amazon.com\/about-aws\/whats-new\/2023\/05\/cedar-open-source-language-access-control\/. AWS open-sources Cedar and describes verification-guided development, combining automated reasoning and differential testing."},{"key":"e_1_3_4_67_2","unstructured":"Amazon Web Services. 2024. Minimize AI Hallucinations and Deliver up to 99% Verification Accuracy with Automated Reasoning Checks Now Available. Retrieved August 28 2025 from https:\/\/aws.amazon.com\/blogs\/aws\/minimize-ai-hallucinations-and-deliver-up-to-99-verification-accuracy-with-automated-reasoning-checks-now-available\/"},{"key":"e_1_3_4_68_2","unstructured":"Amazon Web Services and PwC. 2024. PwC and AWS Build Responsible AI with Automated Reasoning on Amazon Bedrock. Retrieved from https:\/\/aws.amazon.com\/blogs\/machine-learning\/pwc-and-aws-build-responsible-ai-with-automated-reasoning-on-amazon-bedrock\/. A PwC\u2013AWS case-study blog post on using Automated Reasoning checks in Amazon Bedrock Guardrails as a responsible-AI governance layer. The post is useful evidence that formal methods are being positioned as a practical compliance infrastructure for generative AI especially in regulated domains such as EU AI Act risk classification life sciences content review and utility operations. Accessed: 2025-08-28."},{"key":"e_1_3_4_69_2","article-title":"AWS Clean Rooms Differential Privacy","author":"Services Amazon Web","year":"2026","unstructured":"Amazon Web Services. 2026. AWS Clean Rooms Differential Privacy. Documentation, Retrieved from https:\/\/docs.aws.amazon.com\/clean-rooms\/latest\/userguide\/differential-privacy.html. AWS documentation describing Clean Rooms Differential Privacy, a managed privacy-preserving analytics capability that adds calibrated noise to aggregate query results. The page is notable because it explicitly states that the service uses AWS\u2019s SampCert sampler, providing a concrete example of formally verified differential privacy technology in an operational cloud service. Accessed 17 Jan 2026.","journal-title":"Documentation"},{"key":"e_1_3_4_70_2","article-title":"Open Source Cryptography","author":"Services Amazon Web","year":"2026","unstructured":"Amazon Web Services. 2026. Open Source Cryptography. Web page, Retrieved from https:\/\/aws.amazon.com\/security\/opensource\/cryptography\/. Overview of AWS open-source cryptographic and transport libraries, includings2n-tls and AWS-LC. Accessed 17 Jan 2026.","journal-title":"Web page"},{"key":"e_1_3_4_71_2","volume-title":"Shuttle: A Library for Testing Concurrent Rust Code","author":"Labs AWS","year":"2025","unstructured":"AWS Labs. 2025. Shuttle: A Library for Testing Concurrent Rust Code. Retrieved from https:\/\/github.com\/awslabs\/shuttle. Open-source library that controls thread scheduling to enable reproducible, randomized concurrency testing for Rust."},{"key":"e_1_3_4_72_2","volume-title":"TraceForge: Systematic Concurrency Exploration for Distributed Systems in Rust","author":"Labs AWS","year":"2025","unstructured":"AWS Labs. 2025. TraceForge: Systematic Concurrency Exploration for Distributed Systems in Rust. Retrieved from https:\/\/github.com\/awslabs\/TraceForge. Open-source library for systematically exploring message interleavings for distributed-systems tests and Rust-encoded specifications."},{"key":"e_1_3_4_73_2","article-title":"aws-lc-verification","author":"Labs AWS","year":"2026","unstructured":"AWS Labs. 2026. aws-lc-verification. GitHub repository, Retrieved from https:\/\/github.com\/awslabs\/aws-lc-verification. Verification artefacts and documentation for selected AWS-LC configurations; highlights configuration and dependency assumptions (e.g., interaction withs2n-bignum). Accessed 17 Jan 2026.","journal-title":"GitHub repository"},{"key":"e_1_3_4_74_2","doi-asserted-by":"publisher","DOI":"10.23919\/FMCAD.2018.8602994"},{"key":"e_1_3_4_75_2","doi-asserted-by":"publisher","DOI":"10.1109\/MS.2019.2930609"},{"key":"e_1_3_4_76_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-540-24756-2_1"},{"key":"e_1_3_4_77_2","unstructured":"Vishnu Bharathi. 2023. Paper notes: Use of Formal Methods at Amazon Web Services. Available online at https:\/\/vishnubharathi.codes\/blog\/paper-notes-use-of-formal-methods-at-amazon-web-services\/. Vishnu Bharathi\u2019s blog post reviews Newcombe\u2019s influential paper on the use of TLA+ at Amazon Web Services. It summarises how AWS engineers applied formal specification and model checking to critical services such as S3 DynamoDB and EBS. Bharathi highlights key themes: the practicality of TLA+ for uncovering subtle bugs the cultural shift in treating formal methods as \u201cdebuggable design documents\u201d and the importance of management support for adoption. The notes also emphasise lessons on training scalability and the integration of verification into everyday workflows. The post offers an accessible entry point for practitioners interested in AWS\u2019s pioneering industrial use of formal methods."},{"key":"e_1_3_4_78_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-10082-1_3"},{"key":"e_1_3_4_79_2","doi-asserted-by":"publisher","DOI":"10.1145\/3477132.3483540"},{"key":"e_1_3_4_80_2","unstructured":"James Bornholt and Andy Warfield. 2021. How Automated Reasoning Helps Amazon S3 Innovate at Scale. Retrieved from https:\/\/aws.amazon.com\/blogs\/storage\/how-automated-reasoning-helps-us-innovate-at-s3-scale\/?utm_source=chatgpt.com. AWS S3 consists of 300+ microservices and handles trillions of objects making correctness a major challenge. AWS uses automated reasoning tools to check system correctness and security at scale. Zelkova an automated reasoning tool ensures strong security guarantees for AWS IAM policies. AWS is working to develop provably correct systems but challenges remain in scalability and specification complexity."},{"key":"e_1_3_4_81_2","unstructured":"Marc Brooker. 2015. How Amazon Web Services Uses Formal Methods. Available online at https:\/\/brooker.co.za\/blog\/2015\/03\/29\/formal.html. This blog post highlights AWS\u2019s application of formal methods to enhance system reliability and security. He references a Communications of the ACM article detailing AWS\u2019s integration of these techniques into its development processes. Brooker also cites Leslie Lamport\u2019s article \u201cWho Builds a House without Drawing Blueprints?\u201d emphasising the importance of clear written specifications in understanding and preventing mistakes in system design."},{"key":"e_1_3_4_82_2","volume-title":"Getting into Formal Specification, and Getting My Team into It Too","author":"Brooker Marc","year":"2022","unstructured":"Marc Brooker. 2022. Getting into Formal Specification, and Getting My Team into It Too. Retrieved from https:\/\/brooker.co.za\/blog\/2022\/07\/29\/getting-into-tla.html. Blog post discussing formal specification at AWS and noting teams exploring tools including Dafny and the Rust concurrency-testing library Shuttle."},{"key":"e_1_3_4_83_2","doi-asserted-by":"publisher","DOI":"10.1145\/3712057"},{"key":"e_1_3_4_84_2","unstructured":"Aleks Chakarov Jaco Geldenhuys Matthew Heck Mike Hicks Sam Huang Georges Axel Jaloyan Anjali Joshi Rustan Leino Mikael Mayer Sean McLaughlin et\u00a0al. 2025. Formally verified cloud-scale authorization. (2025). Retrieved from https:\/\/www.amazon.science\/publications\/formally-verified-cloud-scale-authorization. The paper reports on Amazon\u2019s four-year effort to rebuild its AWS authorisation engine a critical component invoked over a billion times per second. The original Java-based engine (AuthV1) was difficult to evolve safely at scale. Instead of verifying it directly the team designed a new engine (AuthV2) in Dafny a verification-aware language and proved it correct with respect to a formal specification derived from AuthV1. To ensure deployability they developed a custom compiler that generated readable idiomatic Java code while preserving performance and stakeholder trust. Extensive differential and shadow testing covering a quadrillion production samples ensured AuthV2 precisely matched AuthV1\u2019s behaviour. AuthV2 was deployed in 2024 without incident offering threefold performance gains and enabling confident future changes. Key lessons include the need to rewrite legacy systems for verifiability integrate verification with practical engineering constraints and complement proofs with large-scale testing. This work represents a landmark in the application of formal verification to cloud-scale security-critical infrastructure."},{"key":"e_1_3_4_85_2","unstructured":"Rod Chapman Adam Petcher Torben Hansen Yan Peng Tancr\u00e8de Lepoint Cameron Bytheway and Panos Kampanakis. 2024. Formal Verification of Cryptographic Software at AWS \u2014 Current Practices and Future Trends. Retrieved from https:\/\/assets.amazon.science\/7a\/bd\/2b2ba3934a97a4a2c83781adc48f\/formal-verification-of-cryptographic-software-at-aws-current-practices-and-future-trends.pdf. Discusses s2n-bignum an open-source library for verified big-number arithmetic used in cryptography. The discussion includes three key aspects: (1) Formal verification of arithmetic operations in cryptographic algorithms. (2) Ensuring security through constant-time implementation to mitigate side-channel attacks. (3) The relevance of machine-verified arithmetic for cryptographic safety."},{"key":"e_1_3_4_86_2","doi-asserted-by":"publisher","DOI":"10.1002\/SPE.2949"},{"key":"e_1_3_4_87_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-96142-2_26"},{"key":"e_1_3_4_88_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-96142-2_26"},{"key":"e_1_3_4_89_2","unstructured":"Byron Cook. 2018. Formal Reasoning About the Security of Amazon Web Services. Retrieved from https:\/\/www.floc2018.org\/speaker\/byron-cook\/. See https:\/\/www.youtube.com\/watch?v=JfjLKBO27nw for a video of the FLoC\u201918 plenary lecture."},{"key":"e_1_3_4_90_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-96145-3_3"},{"key":"e_1_3_4_91_2","unstructured":"Byron Cook. 2024. An unexpected discovery: Automated reasoning often makes systems more efficient and easier to maintain. Retrieved from https:\/\/aws.amazon.com\/blogs\/security\/an-unexpected-discovery-automated-reasoning-often-makes-systems-more-efficient-and-easier-to-maintain\/. This AWS blog highlights how automated reasoning enhances security and improves efficiency and maintainability of systems. Formal verification helps fix bugs optimise performance and reduce debugging efforts making systems easier to update and operate while increasing developer confidence."},{"key":"e_1_3_4_92_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-96142-2_28"},{"key":"e_1_3_4_93_2","doi-asserted-by":"publisher","DOI":"10.1007\/S10703-020-00344-2"},{"key":"e_1_3_4_94_2","doi-asserted-by":"publisher","DOI":"10.1145\/3649835"},{"key":"e_1_3_4_95_2","doi-asserted-by":"publisher","DOI":"10.1145\/3689738"},{"key":"e_1_3_4_96_2","unstructured":"Ankush Desai. 2022. Strange Loop: Formal Modeling and Analysis of Distributed Systems (Finding Critical Bugs Early!). Retrieved from https:\/\/www.thestrangeloop.com\/2022\/formal-modeling-and-analysis-of-distributed-systems-finding-critical-bugs-early.html. YouTube video."},{"key":"e_1_3_4_97_2","doi-asserted-by":"publisher","DOI":"10.1145\/3663529.3663854"},{"key":"e_1_3_4_98_2","unstructured":"Mike Dodds. 2023. The Impact of Provable Security: AWS and Supranational. Retrieved from https:\/\/galois.com\/blog\/2023\/09\/the-impact-of-provable-security-aws-and-supranational\/. Focuses on AWS\u2019s application of provable security techniques for cryptographic security. Discusses collaborations with Supranational and Galois to formally verify cryptographic implementations used in AWS services. Emphasises the importance of provable security in large-scale cloud infrastructures."},{"key":"e_1_3_4_99_2","doi-asserted-by":"publisher","DOI":"10.1109\/SECDEV.2018.00027"},{"key":"e_1_3_4_100_2","volume-title":"Introducing Cedar Analysis: Open Source Tools for Verifying Authorization Policies","author":"Erickson Spencer","year":"2025","unstructured":"Spencer Erickson and Liana Hadarean. 2025. Introducing Cedar Analysis: Open Source Tools for Verifying Authorization Policies. Retrieved from https:\/\/aws.amazon.com\/blogs\/opensource\/introducing-cedar-analysis-open-source-tools-for-verifying-authorization-policies\/. Announces Cedar Analysis (symbolic compiler + CLI) and explains its Lean implementation with mechanically checked soundness and completeness for the SMT translation."},{"key":"e_1_3_4_101_2","unstructured":"Galois Inc.2019. Proving Amazon\u2019s s2n correct. Retrieved from https:\/\/galois.com\/project\/amazon-s2n\/. Details the formal verification of Amazon\u2019s s2n TLS implementation using Galois\u2019s Software Analysis Workbench (SAW) and Cryptol. The authors describe how SAW verifies key components including HMAC and DRBG and how they integrate the verification process into Amazon\u2019s CI\/CD workflows."},{"key":"e_1_3_4_102_2","unstructured":"Wojciech Gawro\u0144ski. 2020. (AWS Maniac) How formal methods helped AWS to design amazing services. Available online at https:\/\/awsmaniac.com\/how-formal-methods-helped-aws-to-design-amazing-services\/. Discusses AWS\u2019s application of TLA+ in designing robust cloud services. AWS uses formal verification to ensure correctness particularly in distributed systems and security policies. Provable security is a major goal leading to the creation of tools like Zelkova and Tiros. PlusCal (a procedural extension of TLA+) simplifies the modeling of concurrent algorithms."},{"key":"e_1_3_4_103_2","doi-asserted-by":"publisher","DOI":"10.1023\/A:1022973506233"},{"key":"e_1_3_4_104_2","doi-asserted-by":"publisher","DOI":"10.1007\/11757283_8"},{"key":"e_1_3_4_105_2","unstructured":"John Harrison. 2023. Verified Computer Arithmetic for Cryptography and Elsewhere. Retrieved from https:\/\/arith2023.arithsymposium.org\/slides\/KN1_JohnHarrison_arith2023_harrison.pdf. This paper discusses s2n-bignum an open-source library for verified big-number arithmetic used in cryptography. Key aspects include: (1) Formal verification of arithmetic operations in cryptographic algorithms. (2) Ensuring security through constant-time implementation to mitigate side-channel attacks. (3) The relevance of machine-verified arithmetic for cryptographic safety."},{"key":"e_1_3_4_106_2","volume-title":"How we built Cedar with automated reasoning and differential testing","author":"Hicks Michael","year":"2023","unstructured":"Michael Hicks. 2023. How we built Cedar with automated reasoning and differential testing. Retrieved from https:\/\/www.amazon.science\/blog\/how-we-built-cedar-with-automated-reasoning-and-differential-testing. Describes modelling Cedar\u2019s authoriser and validator in Dafny, proving properties with automated reasoning, and using differential random testing against the production Rust implementation."},{"key":"e_1_3_4_107_2","unstructured":"Neo Kim. 2024. System Design Newsletter: How Amazon S3 Works. Retrieved from https:\/\/newsletter.systemdesign.one\/p\/s3-architecture. Explains Amazon S3\u2019s architecture detailing its object storage model high availability and 11-nines durability. It covers data redundancy erasure coding and replication across facilities as well as various storage classes to optimise costs performance and scalability."},{"key":"e_1_3_4_108_2","unstructured":"Martin Klepsch. 2015. Formal Methods at Amazon. Retrieved from https:\/\/martinklepsch.org\/posts\/formal-methods-at-amazon.html. Discusses Amazon\u2019s adoption of TLA+ (Temporal Logic of Actions) for specifying and verifying complex concurrent fault-tolerant systems. By using TLA+ Amazon engineers identified bugs and ensured system reliability. The article emphasises the benefits and challenges of integrating formal specifications into engineering workflows."},{"key":"e_1_3_4_109_2","unstructured":"Leslie Lamport. 2018. Abridged Version of Amazon Paper. Retrieved from https:\/\/lamport.azurewebsites.net\/tla\/amazon-excerpt.html. Amazon Web Services has successfully used TLA+ and its PlusCal front-end since 2011 to design and verify complex fault-tolerant systems. Traditional testing missed subtle bugs that TLA+ uncovered such as a rare 35-step failure sequence in DynamoDB. Engineers found it practical to learn within weeks and it improved both correctness and time to market. Beyond bug detection TLA+ provided precise mental models for collaboration supported safe performance optimisations and shifted focus to \u201cwhat must go right.\u201d By 2015 multiple teams were using it with executive support ensuring TLA+ specifications became integral to AWS\u2019s system development and planning."},{"key":"e_1_3_4_110_2","article-title":"SampCert: Verified Differential Privacy","author":"Community Lean Prover","year":"2026","unstructured":"Lean Prover Community. 2026. SampCert: Verified Differential Privacy. GitHub repository, Retrieved from https:\/\/github.com\/leanprover\/SampCert. SampCert repository and documentation; describes deployment in AWS Clean Rooms Differential Privacy. Accessed 17 Jan 2026.","journal-title":"GitHub repository"},{"key":"e_1_3_4_111_2","unstructured":"June Lee Hanno Becker and John Harrison. 2024. Formal Verification Makes RSA Faster \u2014 and Faster to Deploy. Retrieved from https:\/\/www.amazon.science\/blog\/formal-verification-makes-rsa-faster-and-faster-to-deploy. Amazon\u2019s Automated Reasoning Group enhanced RSA signature throughput on Graviton2 processors by 33 to 94 per cent depending on key size through algorithmic and microarchitectural optimisations. They utilised formal verification methods to ensure functional correctness resulting in both improved performance and accelerated deployment.."},{"key":"e_1_3_4_112_2","unstructured":"Colm MacC\u00e1rthaigh. 2024. Improving Security in s2n. Retrieved from https:\/\/shufflesharding.com\/posts\/improving-security-in-s2n. AWS developed s2n an open-source TLS\/SSL implementation. Uses formal verification and automated testing to ensure security. Introduced Sidetrail a tool that verifies that s2n is free of side-channel vulnerabilities. Formal verification combined with manual security reviews helped catch and mitigate TLS 1.3 security flaws."},{"key":"e_1_3_4_113_2","unstructured":"Markus de Medeiros Muhammad Naveed Tancr\u00e8de Lepoint Temesghen Kahsai Tristan Ravitch Stefan Zetzsche Anjali Joshi Joseph Tassarotti Aws Albarghouthi and Jean-Baptiste Tristan. 2025. Verified Foundations for Differential Privacy. Retrieved from https:\/\/assets.amazon.science\/94\/22\/04f3db3340ecac13847fb2ff8b97\/verified-foundations-for-differential-privacy.pdf. Presents SampCert a Lean mechanisation with verified discrete Laplace and Gaussian samplers and supporting proof infrastructure; the authors report deployment of verified sampling algorithms in AWS differential privacy offerings."},{"key":"e_1_3_4_114_2","unstructured":"Felipe R. Monteiro Mikhail R. Gadelha and Lucas C. Cordeiro. 2018. Boost Continuous Formal Verification in Industry. Retrieved from https:\/\/ssvlab.github.io\/lucasccordeiro\/talks\/tapas2019_slides.pdf. Discusses how AWS integrates continuous formal verification (CFV) into industrial software development. It highlights challenges such as scalability modularity and complexity management. The key takeaway is that CFV helps detect software vulnerabilities early by integrating formal verification techniques such as model checking into DevOps workflows. The study also emphasises the importance of automation in verification to improve efficiency and scalability."},{"key":"e_1_3_4_115_2","unstructured":"Yannick Moy. 2018. Amazon Relies on Formal Methods for the Security of AWS. Retrieved from https:\/\/blog.adacore.com\/amazon-relies-on-formal-methods-for-the-security-of-aws. Based on a talk by Byron Cook describes how AWS has extensively adopted formal methods to ensure cloud security. AWS has built a team of over 20 formal methods experts to develop verification tools for cloud security. AWS utilises theorem provers and model checking tools to continuously verify cryptographic protocols hypervisors bootloaders firmware and networks. The key takeaway is that formal methods provide security guarantees crucial to AWS customers driving widespread adoption."},{"key":"e_1_3_4_116_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-662-43652-3_3"},{"key":"e_1_3_4_117_2","doi-asserted-by":"publisher","DOI":"10.1145\/2699417"},{"key":"e_1_3_4_118_2","unstructured":"Yan Peng. 2024. Formal Verification of AWS-LibCrypto. Retrieved from https:\/\/sos-vo.org\/. Discusses AWS-LibCrypto an open-source cryptographic library optimised for AWS. It describes formal verification techniques applied to ensure memory safety and functional correctness. The paper outlines AWS\u2019s use of SAW Coq and SMT solvers to verify cryptographic primitives and optimise performance."},{"key":"e_1_3_4_119_2","doi-asserted-by":"publisher","DOI":"10.1145\/3243734.3243775"},{"key":"e_1_3_4_120_2","unstructured":"Andrey Satarin. 2022. Formal Methods at Amazon S3. Retrieved from https:\/\/asatarin.github.io\/talks\/2022-02-formal-methods-at-amazon-s3\/. Amazon S3 uses lightweight formal methods to validate ShardStore a key-value storage node. By decomposing correctness into independent properties utilising an executable reference model and automating checks they prevented 16 production issues including consistency and concurrency bugs thereby ensuring reliability at scale."},{"key":"e_1_3_4_121_2","unstructured":"Steve Schmidt. 2017. s2n Is Now Handling 100 Percent of SSL Traffic for Amazon S3. Retrieved from https:\/\/aws.amazon.com\/blogs\/security\/s2n-is-now-handling-100-percent-of-of-ssl-traffic-for-amazon-s3\/. This AWS blog post announces that s2n has replaced OpenSSL for handling all SSL traffic for Amazon S3. It describes s2n\u2019s security benefits including its smaller codebase use of formal verification and integration of automated reasoning tools."},{"key":"e_1_3_4_122_2","unstructured":"K. Shanmugam and S. Nath. 2024. Automated reasoning and generative AI: Harness creativity with formal verifications. Retrieved from https:\/\/dev.to\/aws\/automated-reasoning-and-generative-ai-harness-creativity-with-formal-verifications-o6. A practitioner-oriented AWS Community Builders article explaining how automated reasoning can be combined with generative AI to check outputs against formal logical constraints. It is a useful accessible secondary source on AWS\u2019s positioning of automated reasoning as a guardrail for reducing hallucinations ensuring policy compliance and advancing trustworthy AI particularly through Amazon Bedrock Guardrails. Accessed: 2025-08-28."},{"key":"e_1_3_4_123_2","volume-title":"Costly Code: The Price Of Software Errors","author":"Simonova Margarita","year":"2023","unstructured":"Margarita Simonova. 2023. Costly Code: The Price Of Software Errors. Retrieved from https:\/\/www.forbes.com\/councils\/forbestechcouncil\/2023\/12\/26\/costly-code-the-price-of-software-errors\/. Practitioner-oriented discussion of the economic impact of software errors, using the language of cost, risk, and business consequence to motivate investment in software quality, testing, and assurance."},{"key":"e_1_3_4_124_2","unstructured":"Ricardo Sueiras. 2024. AWS open source newsletter #205. Retrieved from https:\/\/community.aws\/content\/2pbgDjylg23XWvy5RpWZ3BcsoNA\/aws-open-source-newsletter-205?lang=en. This AWS Open Source Newsletter #205 highlights recent developments in AWS\u2019s open-source initiatives including new project releases community contributions and upcoming events. It showcases AWS\u2019s commitment to fostering open-source collaboration and innovation across various domains."},{"key":"e_1_3_4_125_2","unstructured":"Emma Tosch. 2014. Formal Methods in Practice. Retrieved from https:\/\/blog.emmatosch.com\/formal%20methods\/2014\/09\/29\/formal-methods-in-practice.html. Blog post discussing AWS\u2019s practical use of formal methods especially model checking Alloy SAT solving and TLA+ for debugging distributed-system designs whose edge-case failures were not exposed by conventional testing."}],"container-title":["Formal Aspects of Computing"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3815784","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,21]],"date-time":"2026-10-21T21:00:10Z","timestamp":1792616410000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3815784"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,10,16]]},"references-count":124,"journal-issue":{"issue":"4","published-print":{"date-parts":[[2026,12,31]]}},"alternative-id":["10.1145\/3815784"],"URL":"https:\/\/doi.org\/10.1145\/3815784","relation":{},"ISSN":["0934-5043","1433-299X"],"issn-type":[{"value":"0934-5043","type":"print"},{"value":"1433-299X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,10,16]]},"assertion":[{"value":"2025-10-08","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-05-05","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-10-16","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}