From Traditional to Digital: How software, data and AI are transforming the embedded systems industry
0 likes49 views
The document discusses the transformative impact of digitalization, software, and AI on the embedded systems industry, emphasizing the shift from traditional to digital business models that prioritize continuous value delivery and data-driven decision-making. It highlights the necessity for companies to adapt their operational practices and engage in ecosystem evolution to maintain competitiveness in a rapidly changing digital landscape. Ultimately, the conclusions stress the need for ongoing innovation and reinvention in business strategies to thrive in this digital age.
From Traditional to Digital: How software, data and AI are transforming the embedded systems industry
- 1. From Traditional to Digital: How software, data and AI are transforming the embedded systems industry Helena Holmström Olsson Professor of Computer Science Dept. of Computer Science and Media Technology Malmö University, Sweden
- 2. • Who am I? And what is Software Center? • Digitalization and Digital Transformation • From Traditional to Digital: Business models: Continuous value delivery and recurring revenue Ways-of-working: Data-driven decision making and Ops practices Ecosystems: Transformation and co-existence of existing and emerging • Conclusion Overview
- 4. Software Center Mission: To significantly improve the digitalization capability of the European Software-Intensive industry
- 7. Stairway to Heaven: Speed R&D teams R&D teams V&V R&D teams V&V Release Cust. Sup. R&D teams V&V Release Cust. sup. Prod. mgmt. Sales & mrkt Olsson, H.H., Alahyari, H. and Bosch, J. (2012). Climbing the “Stairway to Heaven”. A Mulitiple-Case Study Exploring Barriers in the Transition from Agile Development towards Continuous Deployment of Software. In Proceedings of the 38th Euromicro Conference on Software Engineering and Advanced Applications (pp. 392-399). IEEE.
- 8. • Who am I? And what is So>ware Center? • Digitaliza?on and Digital Transforma?on • From Tradi?onal to Digital: Business models: Con<nuous value delivery and recurring revenue Ways-of-working: Data-driven decision making and Ops prac<ces Ecosystems: Transforma<on and co-existence of exis<ng and emerging • Conclusion Overview
- 9. Digitalization Digitalization is the use of digital technologies to change a business model and provide new revenue and value- producing opportunities; it is the process of moving to a digital business. - Gartner
- 10. Digitalization
- 11. • …”is an effect that changes the fundamental expectations and behaviors in a culture, market, industry or process that is caused by, or expressed through, digital capabilities, channels or assets.” (Gartner.com) • …the change that happens when new digital technologies, services, capabilities, and business models affect and change the value of the industry's existing services and goods.” (Simplilearn.com) Digital disruption ZipCar, Lyft, Uber P2P commerce, e.g., Airbnb Online payment, digital wallets, cryptocurrency 3D printing Apple, Google etc. Voice, video, media (Skype, Reddit, Discord)
- 12. Technology Evolution mechanics electronics so;ware data artificial intelligence digitalization
- 13. Business Evolution Transactional model(s) Subscription-based model(s) Value-based and continuous model(s) Business model evolu9on New partner(s) New partner(s) New partner(s) Ecosystem evolution DevOps DataOps MLOps Ways-of- working evolu9on Olsson, H. H., & Bosch, J. (2020). Going digital: Disruption and transformation in software-intensive embedded systems ecosystems. Journal of Software: Evolution and Process, e2249.
- 14. Product sales (limited service revenue) Product-as-a- service sales Complementary services around products Customer KPI’s- based business model Multi-sided ecosystem model Focus is on the product as a mechanical and physical item Product sold ”as-is” Focus is on the product but with a few services in areas such as e.g. support Reactive SW updates Focus is on expanding service offerings around the product Proactive and continuous SW updates Focus shifts to customer outcomes and data is monetized with the primary customer base Continuous upgrade of SW and periodic upgrade of electronics and HW Data for product performance Data for QA and diagnostics Data for primary customer base Data for secondary customer base Data from one customer is used for that customer Data for human interpretation Static ML models trained on fixed data sets AI integrated into DevOps Mass customization with ML models adjusting to products and users Autonomous system experimentation Product updates and changes because of the potential to monetize the secondary customer base Full digitalized offering AI/ML/DL dimension Business model dimension Data exploitation dimension Product upgrade dimension From Traditional to Digital: The evolution path along four dimensions Bosch, J., and Olsson, H.H. (2021). "Digital for real: A multicase study on the digital transformation of companies in the embedded systems domain.” Journal of Software: Evolution and Process 33, no. 5: e2333.
- 15. • Who am I? And what is Software Center? • Digitalization and Digital Transformation • From Traditional to Digital: Business models: Continuous value delivery and recurring revenue Ways-of-working: Data-driven decision making and Ops practices Ecosystems: Transformation and co-existence of existing and emerging • Conclusion Overview
- 17. Continuous Value Delivery continuous software deployment
- 18. service revenue product revenue secondary customer base total revenue (2) The business model dimension Product sales (limited service revenue) Product-as-a- service sales Complementary services around products Customer KPI’s- based business model Multi-sided ecosystem model 100% product sales 20% product sales Time Business model dimension Bosch, Jan, and Helena H. Olsson. "Digital for real: A multicase study on the digital transformation of companies in the embedded systems domain.” Journal of Software: Evolution and Process 33, no. 5 (2021): e2333.
- 19. Product- oriented value Outcome- oriented value Comparative value Two - (multi) sided markets Product performance, product health, preventive maintenence Customer insights (productivity, efficiency, quality), life cycle management, device and asset management Comparison to compeAtors, benchmarking, trend analysis, acAons and recommendaAons Data from one customer is used for another (new) customer/ (and vice versa) Commodity Differentiating Innovation Few opportunities for monetization and no generation of new (recurring) revenue streams Customer 1 Customer 2 data data Premium priced product with product-oriented value “as a package” Value is paid for separately (on top of existing sales) Value is paid for separately (beyond existing sales) Data is monetized, e.g., data-as-a-service Opportunities for monetization and generation of new (recurring) revenue streams New (recurring) revenue streams Value provided to stakeholders: Value Offerings Framework Olsson, H.H., and Bosch, J. (2022). Living in a Pink Cloud or Fighting a Whack-a-Mole? On the Creation of Recurring Revenue Streams in the Embedded Systems Domain. SEAA, August 31 – September 2, Maspalomas, Gran Canaria, Spain, 2022.
- 20. • Who am I? And what is Software Center? • Digitalization and Digital Transformation • From Traditional to Digital: Business models: Continuous value delivery and recurring revenue Ways-of-working: Data-driven decision making and Ops practices Ecosystems: Transformation and co-existence of existing and emerging • Conclusion Overview
- 21. “Featuritis”
- 22. Experimentation Data is what keeps us alive. Everything we do ends up with, or start with, an experiment. We have passed the point where we make decisions on what to ship or not. Everything is instrumented and we base what to ship on the results we get from experimentation”. (Product Manager, Microsoft) “We know how much you click, how fast you click, how long you stay, when you come back… This translates into metrics that span across page level metrics, script errors, performance, frequency of pop-up’s, queries per user, click through rate, sessions per user, re-visits of users...”. (Product Manager, Microso4)
- 23. • Due to increasing connectivity and data collection from products in the field, data-driven practices are being adopted also in software-intensive embedded systems companies • Experiments are run on selected instances of the system or as comparisons of previously computed data to ensure value delivery to customers, improve quality and explore new value propositions What we see happening… Performance Stability Efficiency Feedback time Flexibility Security
- 24. Olsson, H.H, and Bosch, J. (2014). The HYPEX model: From opinions to data- driven software development. In Continuous software engineering, pp. 155- 164. Springer, Cham, 2014. Strategic product goal Feature: expected behavior (Bexp) select implement MVF actual behavior (Bact) generate Bexp Experimentation relevant gap (Bact ≠ Bexp) no gap (Bact = Bexp) Business strategy and goals Feature backlog Gap analysis Develop hypotheses implement alternative MVF Product extend MVF abandon DevOps: The HYPEX Model
- 25. The ‘Experimentation Evolution Model’ Fabijan, A., Dmitriev, P., Olsson, H. H., and Bosch J. (2017). The Evolution of Continuous Experimentation in Software Product Development: From Data to a Data-driven Organization at Scale. In Proceedings of the 39th International Conference on Software Engineering (ICSE), May 20 – 28th, Buenos Aires, Argentina.
- 26. • Process support for continuous experimentation in mission-critical systems • Includes the different experimentation objectives and provides a classification of the different experimentation techniques and when these can be used Development Pre-study CI Internal laboratory evaluation Simulation Customer laboratory evaluation Passive launch Restricted launch One-customer gradual rollout Customer laboratory evaluation Internal Feedback channel Gradual rollout Customer request Market R&D goals R&D organization Internal validation Single customer validation Multiple customer validation Ideas General availability Customer Feedback channel A/B testing, Crossover exp. A/B testing, Quasi- experiments, Crossover exp. A/B testing, Crossover experiments New features Software corrections New configurations CE in mission-critical B2B systems David Issa MaVos, Anas Dakkak, Jan Bosch, and Helena Holmström Olsson. 2020. ExperimentaYon for Business-to- Business Mission-CriYcal Systems: A Case Study. In InternaHonal Conference on SoIware and Systems Process (ICSSP ’20), October 10–11, 2020, Seoul, Republic of Korea.
- 27. BML Loop and MVP Development
- 28. Design Thinking – Lean – Agile (Gartner)
- 29. Continuous practices: DevOps DevOps aims at bringing development and operations together in order to develop and deploy resilient and high-quality software in short release cycles Technical transformations include, e.g., automated deployments using build and continuous integration tools, treating infrastructure as code, and continuous monitoring of infrastructure and system behavior in production On the organizational side, it is crucial to build and strengthen a collaborative culture to successfully establish a straightforward communication and shared responsibilities within and across teams
- 30. Continuous practices: DataOps An approach that accelerates the delivery of high-quality results by automa@on and orchestra@on of data life cycle stages Adopts the best practices, processes, tools and technologies from Agile software engineering and DevOps for governing analytics development, optimizing code verification, building and delivering new analytics Aims to help promoting the culture of collaboration and continuous improvement of data
- 31. Continuous practices: MLOps MLOps involves data and model validations, in the context of CI Processed datasets and trained models are automatically and continuously supplied by data scientists to ML systems engineers The introduction of new data and degradation of model performance require a trigger to retrain model or improve model performance through online methods
- 32. Evolution of AI/ML/DL technology use ExperimentaTon & prototyping Non-critical deployment of ML/DL components Critical deployment of ML/DL components Cascading deployment of ML/DL components Autonomous ML/DL components Each of the steps requires increased activities of ‘AI engineering’. Lwakatare, L., Raj, A.M., Bosch, J., Olsson, H.H., and Crnkovic, I. (2019). A Taxonomy of Software Engineering Challenges for Machine Learning Systems: An empirical investigation. In Proceedings of the International Conference of Agile Software Development (XP), pp. 227 – 243, Springer, Cham.
- 33. • An end-to-end approach that covers all steps from data collection to retraining deployed ML models • The framework supports practitioners in advancing their prototypical analysis to a deployed and continuously retrained ML model Framework for productive use of ML Figalist, I., Elsner, C., Bosch, J. and Olsson, H.H., 2020, November. An End-to-End Framework for Productive Use of Machine Learning in Software Analytics and Business Intelligence Solutions. In International Conference on Product-Focused Software Process Improvement (pp. 217-233). Springer, Cham.
- 34. • Continuous delivery of ML systems: • Development – Training – Deployment – Integration - Evolution • Includes phases • Involves roles • Requires collaboration and iteration • Takes place in the context of a larger system AI driven Business Development John, M.M., Olsson, H.H. and Bosch, J., 2022. Towards an AI-driven business development framework: A multi-case study. Journal of Software: Evolution and Process, p.e2432.
- 35. AI engineering: Research Agenda Generic AI engineering challenges Domain specific AI engineering challenges Data science Bosch, J., Olsson, H.H., and Crnkovic, I. (2021). Engineering AI systems: A research agenda, In Artificial Intelligence Paradigms for Smart Cyber-Physical Systems. IGI Global Strategic focus areas that are related to the four main phases of an ML/DL project.
- 36. • Who am I? And what is Software Center? • Digitalization and Digital Transformation • From Traditional to Digital: Business models: Continuous value delivery and recurring revenue Ways-of-working: Data-driven decision making and Ops practices Ecosystems: Transformation and co-existence of existing and emerging • Conclusion Overview
- 37. Customer Need Evolution Model Olsson, H.H. and Bosch, J., 2020. Going digital: Disruption and transformation in software-intensive embedded systems ecosystems. Journal of Software: Evolution and Process, 32(6), p.e2249.
- 38. Ecosystem transformation Olsson, H.H. and Bosch, J., 2020. Going digital: DisrupTon and transformaTon in soaware-intensive embedded systems ecosystems. Journal of So4ware: Evolu>on and Process, 32(6), p.e2249.
- 39. Strategies used by incumbents Olsson, H.H. and Bosch, J., 2020. Going digital: Disruption and transformation in software-intensive embedded systems ecosystems. Journal of Software: Evolution and Process, 32(6), p.e2249.
- 40. Strategies used by new entrants Olsson, H.H. and Bosch, J., 2020. Going digital: Disruption and transformation in software-intensive embedded systems ecosystems. Journal of Software: Evolution and Process, 32(6), p.e2249.
- 41. • Digitalization is transforming (and disrupting) the embedded systems industry to an extent that we have only seen the beginnings of • To thrive in a digital age – companies need to continuously rethink, reposition and reinvent their: Business models: Continuous value delivery and recurring revenue Ways-of-working: Ops practices and data-driven decision making Ecosystems: Transformation, co-existence and strategic engagement • Transforming from a traditional towards a digital company requires effective use of software, data and AI for continuous delivery and improvement of customer value Conclusions