TY - CHAP A1 - Hillenbrand, Andrea A1 - Levchenko, Maksym A1 - Störl, Uta A1 - Scherzinger, Stefanie A1 - Klettke, Meike ED - Boncz, Peter ED - Manegold, Stefan ED - Ailamaki, Anastasia ED - Deshpande, Amol ED - Kraska, Tim T1 - MigCast : Putting a Price Tag on Data Model Evolution in NoSQL Data Stores T2 - Proceedings of the 2019 International Conference on Management of Data (SIGMOD/PODS '19) June 2019, Amsterdam, Netherlands N2 - We demonstrate MigCast, a tool-based advisor for exploring data migration strategies in the context of developing NoSQL-backed applications. Users of MigCast can consider their options for evolving their data model along with legacy data already persisted in the cloud-hosted production data-base. They can explore alternative actions as the financial costs are predicted respective to the cloud provider chosen. Thereby they are better equipped to assess potential consequences of imminent data migration decisions. To this end, MigCast maintains an internal cost model, taking into account characteristics of the data instance, expected work-load, data model changes, and cloud provider pricing models. Hence, MigCast enables software project stakeholders to remain in control of the operative costs and to make informed decisions evolving their applications. KW - Data Migration Strategies KW - latency KW - migration costs KW - NoSQL databases KW - Predictive Migration KW - schema evolution Y1 - 2019 SN - 9781450356435 U6 - https://doi.org/10.1145/3299869.3320223 SP - 1925 EP - 1928 PB - ACM CY - New York, NY, USA ER - TY - JOUR A1 - Scherzinger, Stefanie T1 - Build your own SQL-on-Hadoop Query Engine A Report on a Term Project in a Master-level Database Course JF - ACM SIGMOD Record N2 - This is a report on a course taught at OTH Regensburg in the summer term of 2018. The students in this course built their own SQL-on-Hadoop engine as a term project in just 8 weeks. miniHive is written in Python and compiles SQL queries into MapReduce workflows. These are then executed on Hadoop. miniHive performs generic query optimizations (selection and projection pushdown, or cost-based join reordering), as well as MapReduce-specific optimizations. The course was taught in English, using a flipped classroom model. The course material was mainly compiled from third-party teaching videos. This report describes the course setup, the miniHive milestones, and gives a short review of the most successful student projects. Y1 - 2019 U6 - https://doi.org/10.1145/3377330.3377336 VL - 48 IS - 2 SP - 33 EP - 38 PB - ACM ER - TY - CHAP A1 - Scherzinger, Stefanie A1 - Seifert, Christin A1 - Wiese, Lena T1 - The Best of Both Worlds: Challenges in Linking Provenance and Explainability in Distributed Machine Learning T2 - 2019 IEEE 39th International Conference on Distributed Computing Systems (ICDCS), 7-10 July 2019, Dallas, TX, USA N2 - Machine learning experts prefer to think of their input as a single, homogeneous, and consistent data set. However, when analyzing large volumes of data, the entire data set may not be manageable on a single server, but must be stored on a distributed file system instead. Moreover, with the pressing demand to deliver explainable models, the experts may no longer focus on the machine learning algorithms in isolation, but must take into account the distributed nature of the data stored, as well as the impact of any data pre-processing steps upstream in their data analysis pipeline. In this paper, we make the point that even basic transformations during data preparation can impact the model learned, and that this is exacerbated in a distributed setting. We then sketch our vision of end-to-end explainability of the model learned, taking the pre-processing into account. In particular, we point out the potentials of linking the contributions of research on data provenance with the efforts on explainability in machine learning. In doing so, we highlight pitfalls we may experience in a distributed system on the way to generating more holistic explanations for our machine learning models. KW - Computational modeling KW - Data models KW - Decision trees KW - distributed computing KW - Distributed databases KW - explainable machine learning KW - Machine learning Y1 - 2019 U6 - https://doi.org/10.1109/ICDCS.2019.00161 SP - 1620 EP - 1629 PB - IEEE ER - TY - CHAP A1 - Seifert, Christin A1 - Scherzinger, Stefanie A1 - Wiese, Lena T1 - Towards Generating Consumer Labels for Machine Learning Models T2 - 2019 IEEE First International Conference on Cognitive Machine Intelligence (CogMI), 12-14 Dec. 2019, Dallas, TX, USA N2 - Machine learning (ML) based decision making is becoming commonplace. For persons affected by ML-based decisions, a certain level of transparency regarding the properties of the underlying ML model can be fundamental. In this vision paper, we propose to issue consumer labels for trained and published ML models. These labels primarily target machine learning lay persons, such as the operators of an ML system, the executors of decisions, and the decision subjects themselves. Provided that consumer labels comprehensively capture the characteristics of the trained ML model, consumers are enabled to recognize when human intelligence should supersede artificial intelligence. In the long run, we envision a service that generates these consumer labels (semi-)automatically. In this paper, we survey the requirements that an ML system should meet, and correspondingly, the properties that an ML consumer label could capture. We further discuss the feasibility of operationalizing and benchmarking these requirements in the automated generation of ML consumer labels. KW - Artificial-intelligence-machine-learning-consumer-labels-transparency-x-AI KW - Biological system modeling KW - Data models KW - Machine learning KW - Measurement KW - Predictive models KW - Robustness KW - Training Y1 - 2019 U6 - https://doi.org/10.1109/CogMI48466.2019.00033 SP - 173 EP - 179 PB - IEEE ER - TY - CHAP A1 - Holubová, Irena A1 - Scherzinger, Stefanie ED - Groppe, Sven ED - Gruenwald, Le T1 - Unlocking the potential of nextGen multi-model databases for semantic big data projects T2 - Proceedings of the International Workshop on Semantic Big Data - (SBD '19) 05.07.2019 - 05.07.2019, Amsterdam, Netherlands N2 - A new vision in semantic big data processing is to create enterprise data hubs, with a 360° view on all data that matters to a corporation. As we discuss in this paper, a new generation of multi-model database systems seems a promising architectural choice for building such scalable, non-native triple stores. In this paper, we first characterize this new generation of multi-model databases. Then, discussing an example scenario, we show how they allow for agile and flexible schema management, spanning a large design space for creative and incremental data modelling. We identify the challenge of generating sound triple-views from data stored in several, interlinked models, for SPARQL querying. We regard this as one of several appealing research challenges where the semantic big data and the database architecture community may join forces. Y1 - 2019 SN - 9781450367660 U6 - https://doi.org/10.1145/3323878.3325807 SP - 1 EP - 6 PB - ACM Press CY - New York ER - TY - CHAP A1 - Pilven, Matthieu A1 - Scherzinger, Stefanie A1 - d’Orazio, Laurent ED - Guizzardi, Giancarlo ED - Gailly, Frederik ED - Suzana Pitangueira Maciel, Rita T1 - On Complex Value Relations in Hive T2 - Advances in Conceptual Modeling, ER 2020 workshops CMAI, CMLS, CMOMM4FAIR, CoMoNoS, EmpER, Vienna, Austria, November 3-6, 2020, Proceedings N2 - In this paper, we raise the question how data architects model their data for processing in Apache Hive. This well-known SQL-on-Hadoop engine supports complex value relations, where attribute types need not be atomic. In fact, this feature seems to be one of the prominent selling points, e.g., in Hive reference books. In an empirical study, we analyze Hive schemas in open source repositories. We examine to which extent practitioners make use of complex value relations and accordingly, whether they write queries over complex types. Understanding which features are actively used will help make the right decisions in setting up benchmarks for SQL-on-Hadoop engines, as well as in choosing which query operators to optimize for. KW - Complex value relations KW - Empirical study KW - Hive Y1 - 2019 SN - 978-3-030-34145-9 U6 - https://doi.org/10.1007/978-3-030-34146-6_13 SP - 146 EP - 156 PB - Springer International Publishing CY - Cham ER - TY - CHAP A1 - Maiwald, Benjamin A1 - Riedle, Benjamin A1 - Scherzinger, Stefanie ED - Guizzardi, Giancarlo ED - Gailly, Frederik ED - Suzana Pitangueira Maciel, Rita T1 - What Are Real JSON Schemas Like? T2 - Advances in Conceptual Modeling N2 - Recently, the semantics of the JSON Schema format, a de-facto standard for JSON schema declarations, has been formalized. It turns out that JSON Schema is a surprisingly complex schema language based on an open document semantics. In this paper, we present a first empirical analysis of a curated collection of real-world JSON Schemas. Knowing what real JSON Schemas are like (to borrow from a title of a related study on DTDs) helps practitioners and researchers in making realistic assumptions when building tools for JSON Schema processing. Y1 - 2019 SN - 978-3-030-34145-9 U6 - https://doi.org/10.1007/978-3-030-34146-6_9 VL - 11787 SP - 95 EP - 105 PB - Springer International Publishing CY - Cham ER - TY - CHAP A1 - Filho, Edson Ramiro Lucas A1 - de Almeida, Eduardo Cunha A1 - Scherzinger, Stefanie T1 - Don’t Tune Twice: Reusing Tuning Setups for SQL-on-Hadoop Queries T2 - Conceptual Modeling : 38th International Conference, ER 2019, Salvador, Brazil, November 4-7, 2019, Proceedings N2 - SQL-on-Hadoop processing engines have become state-of-the-art in data lake analysis. However, the skills required to tune such systems are rare. This has inspired automated tuning advisors which profile the query workload and produce tuning setups for the low-level MapReduce jobs. Yet with highly dynamic query workloads, repeated re-tuning costs time and money in IaaS environments. In this paper, we focus on reducing the costs for up-front tuning. At the heart of our approach is the observation that a SQL query is compiled into a query plan of MapReduce jobs. While the plans differ from query to query, single jobs tend to be similar between queries. We introduce the notion of the code signature of a MapReduce job and, based on this, our concept of job similarity. We show that we can effectively recycle tuning setups from similar MapReduce jobs already profiled. In doing so, we can leverage any third-party tuning adviser for MapReduce engines. We are able to show that by recycling tuning setups, we can reduce the time spent on profiling by 50% in the TPC-H benchmark. Y1 - 2019 U6 - https://doi.org/10.1007/978-3-030-33223-5_9 SP - 93 EP - 107 PB - Springer CY - Cham ER -