Keynote Lectures

Abstract
In recent years we have witnessed an increasing interest in monitoring crowd flows by capturing the signals from devices such as smartphones. Notably WiFi packet sniffing is popular, in particular by identifying devices by means of their transmitted MAC address. Problematic is the situation that this method infringes upon privacy, even after applying secure one-way encryption techniques, for the simple reason that the European General Data Protection Regulation stipulates that data captured from a device may never be processed if there is any way that this data may lead to the identification of an individual.

Considering the privacy sensitivity of crowd monitoring by capturing signals from smartphones, the question is whether one can design a cloud-based system that is guaranteed to preserve the privacy of individuals while at the same time providing accurate data on crowd flows. In this talk, I will focus on the architecture and design of such a system, which can even withstand malfunctioning cloud-hosting providers.

Abstract
Blockchain has emerged as a decentralized platform for managing digital assets and executing 'smart contracts', i.e., user-defined code. While blockchain's suitability for a given use case should always be scrutinized, it does have the potential to disrupt many of the connection points between individuals, companies, and government entities.
In this keynote talk, I will provide an overview of what architects and developers need to know in order to build blockchain-based applications, and how it relates to the cloud and software services. Among others, I will cover blockchain-as-a-service concepts, as well as architectural concerns and model-driven engineering for blockchain applications, the latter also in relation to collaborative business processes. To highlight some of the challenges, I will discuss insights from a project on "programmable money", i.e., blockchain-based money for conditional payments where the money itself checks whether it can be spent in a certain way at the point of payment.
Finally, I will touch on insights into current adoption of blockchain.

Abstract
Artificial Intelligence based on Machine Learning, and in particular Deep Learning, is today the fastest growing trend in software development, and literally used in all other research disciplines, with a very high impact on the modern society. However, a wide use of AI in many systems, in particular dependable systems, is still far away. The challenges for managing  AI-based complex and dependable systems through  the entire lifecycle  are enormous. Some aspects of these challenges are based on management of resources, including computational, data storage capacity. Others are related to dependability itself, as the AI has by its nature a probabilistic approach, and dependable systems require justifiable quality assurance. This talk discusses some of these challenges, illustrate a case of Cyber-physical systems, and gives some ideas for new research in software engineering for AI development, i.e. for AI engineering.

Abstract
For a long time benchmarking has been considered a practical discipline driven by industry's interest in marketing their products and services. Benchmarking evolved from the need to be able to compare different systems fairly in order to make an informed purchasing decision. In the past decade, benchmarking has gained increasing interest in the scientific community and its scope has expanded significantly. Systems benchmarking is now considered in a broad sense, including rating tools and research benchmarks (benchmark frameworks), and benchmarks are employed in many fields of testing going beyond competitive system comparisons, for example, regulatory programs, research evaluation, and testing during system design and development. Further, the scope of systems benchmarking has expanded to other system attributes in addition to classical performance aspects, for example, system energy efficiency, reliability, or security. The conception, design, and development of benchmarks requires a thorough understanding of the benchmarking fundamentals including statistics, measurement methodologies, metrics, and relevant workload characteristics. Benchmarking involves deciding what to measure, how to measure it, and how to aggregate the measurement results into meaningful metrics. The aggregation of metrics into scoring systems as well as the design of workloads, including workload characterization and modeling, are further challenging topics.

In this keynote talk, I will first provide an overview of the foundations of benchmarking as a discipline, covering the three fundamental elements of each benchmarking approach: metrics, workloads, and measurement methodology. I will discuss relevant topics and respective literature sources based on extensive experience gained from teaching these topics in the past 15 years at several leading European universities. I will then present several case studies from different fields with a focus on cloud benchmarking. The case studies illustrate the unique challenges that arise in the conception and development of benchmarks for specific systems or subsystems.