Information and Software Technology最新文献

{"title":"MLOps best practices, challenges and maturity models: A systematic literature review","authors":"Mohammad Zarour,&nbsp;Hamza Alzabut,&nbsp;Khalid T. Al-Sarayreh","doi":"10.1016/j.infsof.2025.107733","DOIUrl":"10.1016/j.infsof.2025.107733","url":null,"abstract":"<div><h3>Context:</h3><div>Agile and DevOps methodologies have revolutionized software development, leading to increased efficiency and reliability in product delivery. Building on this success, Machine Learning Operations (MLOps) has emerged to streamline the development and deployment of machine learning (ML) models, addressing challenges unique to ML workflows.</div></div><div><h3>Objectives:</h3><div>This study aims to explore the complexities organizations face when adopting MLOps, focusing on three main challenges: the lack of standardized practices, difficulties in maintaining model consistency and scalability, and ambiguities in assessing MLOps maturity. The study also aims to identify best practices and common pitfalls, contributing to a clearer understanding and standardization of MLOps.</div></div><div><h3>Methods:</h3><div>A comprehensive literature review was conducted, analyzing 45 articles that address MLOps best practices, challenges, maturity models, and lessons from previous implementations. This review categorizes findings to provide insights into successful and unsuccessful MLOps applications.</div></div><div><h3>Results:</h3><div>The study identifies nine best practices, eight common challenges, and five maturity models relevant to MLOps adoption. Key lessons from successful and unsuccessful MLOps implementations are outlined, with a focus on improving standardization and reducing ambiguity in MLOps practices.</div></div><div><h3>Conclusion:</h3><div>The findings highlight the importance of establishing standardized MLOps practices to address the unique challenges of machine learning workflows. By categorizing best practices, maturity models, and lessons learned, this study aims to contribute to a robust MLOps framework that enhances the reliability and scalability of machine learning in production environments.</div></div>","PeriodicalId":54983,"journal":{"name":"Information and Software Technology","volume":"183 ","pages":"Article 107733"},"PeriodicalIF":3.8,"publicationDate":"2025-03-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143738109","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Reengineering software systems into microservices: State-of-the-art and future directions","authors":"Thakshila Imiya Mohottige ,&nbsp;Artem Polyvyanyy ,&nbsp;Colin Fidge ,&nbsp;Rajkumar Buyya ,&nbsp;Alistair Barros","doi":"10.1016/j.infsof.2025.107732","DOIUrl":"10.1016/j.infsof.2025.107732","url":null,"abstract":"<div><h3>Context:</h3><div>With the acknowledged benefits of microservices architectures, such as scalability, flexibility, improved maintenance, and deployment, legacy software systems are increasingly being reengineered into microservices. Recently, a plethora of methods, techniques, tools, and evaluation criteria for reengineering software systems into microservices have been proposed without being systematized.</div></div><div><h3>Objectives:</h3><div>The objective of this work is to conduct an in-depth systematic literature review to identify and analyze methods, techniques, and tools for reengineering software systems into microservices and the ways for evaluating such reengineering initiatives and their results.</div></div><div><h3>Methods:</h3><div>A systematic literature review of works on reengineering software systems into microservices was performed, yielding 117 primary studies. The review focused on addressing key research questions concerning the evolution of microservices reengineering, methodologies employed, tools available, and the challenges faced in the reengineering process. We used a taxonomy development method to systematize knowledge in these areas.</div></div><div><h3>Results:</h3><div>The analysis revealed multiple reengineering approaches: static, dynamic, hybrid, and artifact-driven. Significant evaluation criteria identified include coupling, cohesion, and modularity. Key paradigms for microservices reengineering, such as domain-driven design and interface analysis, were identified and discussed. The study also highlights that incremental and iterative transitions are favored in practice.</div></div><div><h3>Conclusion:</h3><div>This study provides a structured overview of the current state of research on reengineering software systems into microservices. It highlights challenges in existing reengineering methodologies. Future directions include validating behavioral equivalence of original and reengineered systems, automating microservices generation, and refining database layer partitioning. The findings emphasize the need for further work to enhance the reengineering process and evaluation of the transition between monolithic and microservices architectures.</div></div>","PeriodicalId":54983,"journal":{"name":"Information and Software Technology","volume":"183 ","pages":"Article 107732"},"PeriodicalIF":3.8,"publicationDate":"2025-03-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143760685","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"BTAL: An imbalance software bug report triage approach based on BERT-TextCNN","authors":"Yanmei Zhang,&nbsp;Yuhang Sun,&nbsp;Yi Shi,&nbsp;Shujuan Jiang,&nbsp;Guan Yuan","doi":"10.1016/j.infsof.2025.107731","DOIUrl":"10.1016/j.infsof.2025.107731","url":null,"abstract":"<div><div>With the expansion of software project scale, a large number of software bug reports have been generated. Bug triage is an indispensable task in software development and maintenance, which directly affects the efficiency of software bug fixing and maintenance cost. Existing bug triage methods often fail to make full use of the useful information in defect reports, resulting in the neglect of some auxiliary information that is critical to the defect triage task. Meanwhile, they also fail to take into account the uncertainty of developers’ work and the gap in their activity, thus leading to the imbalance in software defect report datasets. To address this issue, we propose an imbalance software bug report triage method, BTAL, based on BERT-TextCNN. Firstly, the method utilizes multiple information from software bug reports and employs BERT and TextCNN models for vector representation and feature extraction. Then, the output feature vectors are input into the softmax function to obtain the probabilities of bug reports being triaged to developers. To solve the problem of imbalanced datasets, we propose an adaptive loss function that can adaptively adjust the loss weights based on different categories of samples. This helps the network reduce its focus on majority classes and increase its focus on minority classes, thereby improving triage accuracy. Experimental results on five large-scale open-source software projects, namely GCC, NetBeans, Eclipse, Mozilla, and OpenOffice, demonstrated the effectiveness of the BTAL method in solving software bug triage problems, outperforming current state-of-the-art models.</div></div>","PeriodicalId":54983,"journal":{"name":"Information and Software Technology","volume":"183 ","pages":"Article 107731"},"PeriodicalIF":3.8,"publicationDate":"2025-03-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143760686","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A study of Scrum @ S&P Global in the post-COVID-19 era: Unsuitable for remote work or just flawed implementation?","authors":"Adam Przybyłek ,&nbsp;Daniel Belter ,&nbsp;Kieran Conboy","doi":"10.1016/j.infsof.2025.107728","DOIUrl":"10.1016/j.infsof.2025.107728","url":null,"abstract":"<div><h3>Context:</h3><div>Although the COVID-19 pandemic is behind us, the seminal challenge for agile teams has not gone away, as “Work-From-Home (WFH) is here to stay”. Despite a wealth of literature documenting the challenges of applying agile methods in remote environments, practical guidance for overcoming these challenges remains limited. Furthermore, systematic meta-guidelines to help practitioners develop tailored, context-specific solutions are largely absent.</div></div><div><h3>Method:</h3><div>This article reports on an Action Research project conducted at S&amp;P Global 18 months after the company’s transition to a WFH model.</div></div><div><h3>Objective:</h3><div>The project aimed to evaluate existing adaptations to remote work, identify persistent issues, and develop strategies to address them.</div></div><div><h3>Results:</h3><div>Our investigation revealed that while ad-hoc solutions — such as virtualizing Scrum ceremonies and shifting informal in-person conversations to Slack channels — enabled work continuity during the pandemic, they proved insufficient for tackling more complex challenges. Intriguingly, only two ongoing issues stemmed from misalignments between Scrum and remote settings, while the remainder were rooted in flawed Scrum implementation. With our guidance, the participating team not only rediscovered the “Inspect and Adapt” principle but also mastered a systematic approach to method tailoring, enabling them to devise measures that effectively resolved most of their issues.</div></div><div><h3>Conclusion:</h3><div>Our study reinforces prior findings indicating the applicability of Scrum beyond its traditional “home ground”, while also highlighting the need for remote teams to exhibit even greater discipline in adhering to the Scrum framework to navigate the additional challenges of remote work. We also observed a heightened necessity for more comprehensive documentation in remote settings to compensate for limited immediate access to colleague assistance, which consequently restricts knowledge sharing. Finally, the structured approach we employed offers valuable guidance for software companies seeking to identify and address operational inefficiencies in both remote and hybrid work environments.</div></div>","PeriodicalId":54983,"journal":{"name":"Information and Software Technology","volume":"183 ","pages":"Article 107728"},"PeriodicalIF":3.8,"publicationDate":"2025-03-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143725357","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A systematic mapping study on graph machine learning for static source code analysis","authors":"Jesse Maarleveld,&nbsp;Jiapan Guo,&nbsp;Daniel Feitosa","doi":"10.1016/j.infsof.2025.107722","DOIUrl":"10.1016/j.infsof.2025.107722","url":null,"abstract":"<div><div><strong>Context:</strong> In recent years, graph machine learning and particularly graph neural networks have seen successful and widespread applications in many fields, including static source code analysis. Such machine learning techniques enable learning on rich information networks capable of representing different relations and entities.</div><div>However, there have been no comprehensive studies investigating the use of graph machine learning for static source code analysis. There is no complete systematic picture of what techniques may be considered tried and tested, and where opportunities for future improvements can still be found.</div></div><div><h3>Objective:</h3><div>The main goal of this study is to provide a broad overview of the state of the art of static source code analysis using graph machine learning.</div></div><div><h3>Methods:</h3><div>A systematic mapping was performed covering 4499 studies, presenting a final selection of 323 primary studies.</div></div><div><h3>Results:</h3><div>Among the selected studies, seven major sub-domains were identified. The use and combinations of artefacts, different graph representations, different features, and different machine learning models used were collected and categorised.</div></div><div><h3>Conclusions:</h3><div>The use of graph learning, and in particular graph neural networks, has increased significantly since 2018. Although a wide variety of methods is used, across every dimension we investigated (artefacts, graphs, features, models), we found small sets of technologies which are used in the vast majority of studies. Future opportunities lie in exploring under-explored domains more thoroughly, exploring the use of additional artefacts alongside source code, and paying more attention to interpretability and explainability.</div></div>","PeriodicalId":54983,"journal":{"name":"Information and Software Technology","volume":"183 ","pages":"Article 107722"},"PeriodicalIF":3.8,"publicationDate":"2025-03-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143738110","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"White-box structure analysis of pre-trained language models of code for effective attacking","authors":"Chongyang Liu,&nbsp;Xiaoning Ren,&nbsp;Yinxing Xue","doi":"10.1016/j.infsof.2025.107730","DOIUrl":"10.1016/j.infsof.2025.107730","url":null,"abstract":"<div><h3>Context:</h3><div>Pre-trained language models of code (PLMs-C for short) have dramatically improved the state-of-the-art on various programming language processing tasks.</div></div><div><h3>Objective:</h3><div>Due to these well-performed models being easily disturbed by slight perturbations, several black-box approaches have been proposed for attacking them. However, these studies have presented two challenges: (1) black-box attacks lack interpretability in generating adversarial examples and are inefficient in attacking; (2) white-box analysis methods in natural language processing (NLP) have not yet been applied to PLMs-C, incurring a gap in this field.</div></div><div><h3>Methods:</h3><div>To address these challenges, we make the first attempt to perform a white-box structure analysis for PLMs-C, followed by a grey-box attack for PLMs-C named <span>MindAC</span> based on derived linguistic structures. Specifically, referring to the probing tasks for analyzing the linguistic property of text in NLP, we first design eight novel probing tasks for code to perform white-box structure analysis. We derive three types of linguistic structures from PLMs-C named <em>SurStruct</em>, <em>SyntStruct</em>, and <em>SemStrcut</em> which correspond to the surface, syntactic and semantic structures of code, respectively. Subsequently, <span>MindAC</span> perturbs the code snippets through variable replacement, variable redefinition, and equivalent transformation of loop statements. Besides, in linguistic structures, <span>MindAC</span> introduces the angular distance of hidden output (<em>ADHO</em>) and the Euclidean distance of attention output (<em>EDAO</em>) to guide the generation of adversarial examples.</div></div><div><h3>Results:</h3><div>Our experiments reveal that: (1) for the first time, it has been demonstrated that PLMs-C possess three linguistic structures; (2) <span>MindAc</span> outperforms the state-of-the-art baselines on attack success rate by 2.47% <span><math><mo>∼</mo></math></span> 25.45%, reduces the execution time by 25 m <span><math><mo>∼</mo></math></span> 36h16 m, and achieves a significantly lower number of queries. Furthermore, we perform adversarial fine-tuning on the training sets and recover the <em>Accuracy</em> and <em>F1</em> of the victim models by at least 57.76% and 60.13%, respectively.</div></div><div><h3>Conclusion:</h3><div>The results show that based on the derived linguistic structures, the proposed <span>MindAC</span> is more interpretable, effective, and efficient in attacking the PLMs-C compared with the state-of-the-art baselines. Besides, the generated adversarial examples can help to enhance the robustness of PLMs-C.</div></div>","PeriodicalId":54983,"journal":{"name":"Information and Software Technology","volume":"183 ","pages":"Article 107730"},"PeriodicalIF":3.8,"publicationDate":"2025-03-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143738112","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Data-Driven Agility: Assessing Agile Culture transformation in a technology organisation","authors":"Chukwudi Uwasomba ,&nbsp;Advait Deshpande ,&nbsp;Helen Sharp ,&nbsp;Peggy Gregory ,&nbsp;Rod Willis ,&nbsp;Leonor Barroca ,&nbsp;Maduka Uwadi ,&nbsp;Katie Taylor","doi":"10.1016/j.infsof.2025.107729","DOIUrl":"10.1016/j.infsof.2025.107729","url":null,"abstract":"<div><h3>Context:</h3><div>Gaining an understanding of the state of an organisation’s culture during Agile transformation is important because culture underpins all aspects of an organisation’s way of working and can indicate how successful the transformation has been.</div></div><div><h3>Objective:</h3><div>This paper explores the impact of Agile transformation on various dimensions of organisational culture over time within a technology organisation. Additionally, it demonstrates how datasets collected using the Pulse survey instrument, a tool for assessing an organisation’s culture, can be analysed to provide actionable insights to support organisations in their cultural transformation efforts.</div></div><div><h3>Methods:</h3><div>This paper employs a mixed research method to conduct a post-hoc analysis of the datasets obtained from a technology organisation that utilised the survey instrument in 2021 and 2022 to assess its transformation agenda. The collected data were analysed using descriptive and inferential statistics. We also assess the internal reliability and validity of the instrument using Cronbach’s Alpha, Composite Reliability, factor loadings and Average Variance Extracted.</div></div><div><h3>Results:</h3><div>Results show that all the Alpha values of the instrument fall between 0.744 and 0.901, which are higher than the satisfactory value of 0.700, indicating acceptable to excellent reliability. After the intervention, the targeted cultural area, that is, Trust and Transparency (TT) improved significantly, and there was a general improvement across almost all areas. The organisation found the insights provided by the survey instrument aided their understanding of the change process.</div></div><div><h3>Conclusion:</h3><div>This study presents an analysis framework to support organisations using or seeking to use the Pulse survey instrument in their efforts to transform culture. The findings validate the use of statistical analysis and data-driven approaches to track shifts in various dimensions of organisational culture over time. The study concludes that targeted efforts on culture elements can lead to corresponding improvements in many areas including those not targeted, emphasising the interconnectedness of Agile culture elements.</div></div>","PeriodicalId":54983,"journal":{"name":"Information and Software Technology","volume":"183 ","pages":"Article 107729"},"PeriodicalIF":3.8,"publicationDate":"2025-03-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143734788","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A low-code assessment platform for urban digital twins","authors":"Martina De Sanctis ,&nbsp;Ludovico Iovino ,&nbsp;Maria Teresa Rossi ,&nbsp;Manuel Wimmer","doi":"10.1016/j.infsof.2025.107726","DOIUrl":"10.1016/j.infsof.2025.107726","url":null,"abstract":"<div><h3>Context:</h3><div>Digital twins, as virtual models reflecting physical entities, serve multiple purposes, supporting the understanding, design, and management of systems, across various domains. The concept of digital twins has been also applied to smart cities, coining the term “Urban Digital Twin”. Indeed, smart cities continuously produce data that can serve as a real-time feed for their digital twin representations. However, developing urban digital twins is challenging as the domain is complex and evolving rapidly. Moreover, existing digital twin platforms are mostly generic platforms, limiting their adoption for a specific context such as smart cities.</div></div><div><h3>Objectives:</h3><div>The aim of this study is to overcome the limitation of existing general-purpose digital twin platforms, particularly in the context of smart cities, by supporting the development and evolution of urban digital twins.</div></div><div><h3>Method:</h3><div>To this aim, we propose a low-code assessment platform for smart cities represented as urban digital twins by leveraging distributed runtime models to implement and deploy service-based quality evaluation systems. The innovative platform and its architecture enable the creation of efficient urban digital twins, allowing for their continuous evolution without requiring redeployment whenever the digital twin definitions are modified.</div></div><div><h3>Results:</h3><div>Experimental evaluations demonstrate the effectiveness and efficiency of the proposed platform, highlighting its potential for evolving urban digital twins in smart cities quality assessment.</div></div><div><h3>Conclusions:</h3><div>Results show evidence that combining distributed runtime models with the advantages of low-code platforms is beneficial within the smart cities domain. Moreover, this study underscores the significance of specialized platforms tailored to the smart cities domain.</div></div>","PeriodicalId":54983,"journal":{"name":"Information and Software Technology","volume":"183 ","pages":"Article 107726"},"PeriodicalIF":3.8,"publicationDate":"2025-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143738111","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A taxonomy of microservice integration techniques","authors":"Georg-Daniel Schwarz ,&nbsp;Andreas Bauer ,&nbsp;Dirk Riehle ,&nbsp;Nikolay Harutyunyan","doi":"10.1016/j.infsof.2025.107723","DOIUrl":"10.1016/j.infsof.2025.107723","url":null,"abstract":"<div><h3>Context:</h3><div>Microservices have become an important architectural style for building robust and scalable software systems. A system’s functionality is split into independent units, the microservices, that communicate over a network and can be deployed independently. The shift of complexity into the integration layer necessitates enhanced collaboration among stakeholders, stressing the importance of effective communication.</div></div><div><h3>Objective:</h3><div>We aim to streamline communication between stakeholders in microservice-based projects by constructing a framework for enhanced clarity, a taxonomy, by answering our research question: <em>“How can microservice integration techniques be classified?”</em></div></div><div><h3>Method:</h3><div>We conducted a thematic analysis of literature and six expert interviews to identify microservice integration techniques and construct a taxonomy.</div></div><div><h3>Results:</h3><div>The results of this study are (i) a taxonomy for microservice integration techniques consisting of five main and ten refined categories, (ii) the classification of 121 found integration techniques, (iii) an illustration of the taxonomy usage based on three selected techniques to demonstrate the procedure in case of classification ambiguity, (iv) a comparison of data gathered from literature with the interviews, and (v) comprehensive supplementary materials.</div></div><div><h3>Conclusion:</h3><div>The taxonomy offers a structured framework to classify microservice integration techniques and enhances the understanding of the diverse landscape of microservice integration techniques, including organizational ones that are often overlooked. Practitioners can discover integration techniques through the taxonomy and apply them with guidance provided in the supplementary materials.</div></div>","PeriodicalId":54983,"journal":{"name":"Information and Software Technology","volume":"183 ","pages":"Article 107723"},"PeriodicalIF":3.8,"publicationDate":"2025-03-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143746827","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An empirical guide to MLOps adoption: Framework, maturity model and taxonomy","authors":"Meenu Mary John ,&nbsp;Helena Holmström Olsson ,&nbsp;Jan Bosch","doi":"10.1016/j.infsof.2025.107725","DOIUrl":"10.1016/j.infsof.2025.107725","url":null,"abstract":"<div><h3>Context:</h3><div>Machine Learning Operations (MLOps) has become a top priority for companies. However, its adoption has become challenging due to the need for proper guidance and awareness. Most of the MLOps solutions available in the market are designed to fit the specific platform, tools and culture of the providers.</div></div><div><h3>Objective:</h3><div>The objective is to develop a structured approach to adopting, assessing and advancing MLOps adoption.</div></div><div><h3>Methods:</h3><div>The study was conducted based on a multi-case study across fourteen companies.</div></div><div><h3>Results:</h3><div>We provide a comprehensive analysis that highlights the similarities and differences in the adoption of MLOps practices among companies. We have also empirically validated the developed MLOps framework and MLOps maturity model. Furthermore, we carefully reviewed the feedback received from practitioners and revised the MLOps framework and maturity model to confirm its effectiveness. Additionally, we develop an MLOps taxonomy for classifying ML use cases based on their context and requirements into the desired stage of the MLOps framework and maturity model.</div></div><div><h3>Conclusion:</h3><div>The findings provide companies with a structured approach to adopt, assess, and further advance the adoption of MLOps practices regardless of their current status.</div></div>","PeriodicalId":54983,"journal":{"name":"Information and Software Technology","volume":"183 ","pages":"Article 107725"},"PeriodicalIF":3.8,"publicationDate":"2025-03-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143704182","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}