Seminar on Modern Software Engineering and Database Concepts Gunter - - PowerPoint PPT Presentation
Databases D B and Software S E Engineering Seminar on Modern Software Engineering and Database Concepts Gunter Saake, Jacob Kr uger, David Broneske, Xiao Chen, Gabriel Campero Durand, Bala Gurumurthy, Sandro Schulze, Sabine Wehnert
D S E B
Databases Software Engineering and
Gunter Saake, Jacob Kr¨ uger, David Broneske, Xiao Chen, Gabriel Campero Durand, Bala Gurumurthy, Sandro Schulze, Sabine Wehnert
Arbeitsgruppe Datenbanken und Software Engineering
D S E B
Databases Software Engineering and
Pro-Seminar, Bachelor, 3 CP Delivieries:
age Wissenschaftliches Seminar, Bachelor, 3CP Delivieries:
Saake et al. Seminar on Modern Software Engineering and Database Concepts 2
D S E B
Databases Software Engineering and
uhrungstermin (heute)
Thema)
Pr¨ asentationen an Einzelterminen
age als Blockseminar – Termin wird am Ende festgelegt
ussen eine Woche vor dem Vortrag eingereicht werden → Feedback vom Betreuer
Wochen vor dem Vortrag abgegeben werden
Saake et al. Seminar on Modern Software Engineering and Database Concepts 3
D S E B
Databases Software Engineering and
usselkompetenzen erlernen
uben
→ M¨
ur Bachelor-Arbeit / Team-Projekt
Saake et al. Seminar on Modern Software Engineering and Database Concepts 4
D S E B
Databases Software Engineering and
Uberziehen: Redner wird abgew¨ urgt
uh: Mehr Fragen (ggf. mehr Kritik)
asentationen testen!
Saake et al. Seminar on Modern Software Engineering and Database Concepts 5
D S E B
Databases Software Engineering and
D S E B
Evaluationsbogen Vortrag
Arbeitsgruppe Datenbanken und Software Engineering
Titel des Vortrags:
Sehr Gut . . . Neutral . . . Schlecht Pr¨ asentation Wertung: 1 2 3 4 5 6 7 Auftreten, z.B.
ahigkeit, ¨ Uberzeugungskraft Sprache und Stimme, z.B.
arke, Modulation, Klarheit
ussigkeit Mimik und Gestik, z.B.
Fachlicher Eindruck, z.B.
at
Folien Wertung: 1 2 3 4 5 6 7 Design, z.B.
Ubersichtlichkeit
Visualisierungen (Tabellen, Grafiken,...), z.B.
arungskraft, Klarheit
Vortragskonzept Wertung: 1 2 3 4 5 6 7 Gliederung, z.B.
Argumentation, z.B.
ussigkeit, Roter Faden
andigkeit Inhalt Wertung: 1 2 3 4 5 6 7 Erl¨ auterung der Rahmenbedingungen, z.B.
undung)
Erl¨ auterung der Vorgehensweise, z.B.
Bemerkungen:
Saake et al. Seminar on Modern Software Engineering and Database Concepts 6
D S E B
Databases Software Engineering and
Saake et al. Seminar on Modern Software Engineering and Database Concepts 7
D S E B
Databases Software Engineering and
Bekanntgeben von neuen Errungenschaften/Erfahrungen
Andere (z.B. Community) ¨ uber die eigene Arbeit informieren
F¨ ur euch: → ¨ Uben f¨ ur die Bachelor-Arbeit
Saake et al. Seminar on Modern Software Engineering and Database Concepts 8
D S E B
Databases Software Engineering and
D S E B
Evaluationsbogen wissenschaftliche Ausarbeitung
Arbeitsgruppe Datenbanken und Software Engineering
Titel der Ausarbeitung: Autor: Gutachter:
Sehr Gut . . . Neutral . . . Schlecht Titel, Abstract, Einleitung Wertung: 1 2 3 4 5 6 7 Diskussionspunkte (z.B.):
at der Zusammenfassung
Struktur Wertung: 1 2 3 4 5 6 7 Diskussionspunkte (z.B.):
Uberschriften zum Inhalt Inhalt Wertung: 1 2 3 4 5 6 7 Diskussionspunkte (z.B.):
undungen
von Beispielen
Saake et al. Seminar on Modern Software Engineering and Database Concepts 9
D S E B
Databases Software Engineering and
Saake et al. Seminar on Modern Software Engineering and Database Concepts 10
D S E B
Databases Software Engineering and
CPU “smaller than”-selection
i n t pos = 0; f o r ( i n t i =0; i < a r r a y s i z e ; ++i ){ i f ( a r r a y [ i ] < comp val ) r e s u l t [ pos++]=i ; }
GPU “smaller than”-selection
i n t t i d = t h r e a d I d x . x + b l o c k I d x . x ✯ blockDim . x ; w h i l e ( tid<a r r a y s i z e ){ bitmask [ t i d ] = ( a r r a y [ t i d ] < comparison value ) ; t i d += blockDim . x ✯ gridDim . x ; }
Code Optimizations (Broneske)
1.
2.
State-of-the-Art Multicore Architectures. SC Saake et al. Seminar on Modern Software Engineering and Database Concepts 11
D S E B
Databases Software Engineering and
Multi-Dimensional Index Structures for Main Memory (Broneske)
Hauptspeicherdatenbanken sind ein heißes Forschungsthema. Aktuell stehen beschleunigte Scans im Fokus, wobei adaptierte, klassische Indexstrukturen jedoch nicht außer Acht gelassen werden sollten. Die Frage ist: Welche klassischen Indexstrukturen machen f¨ ur den Hauptspeicherbereich Sinn? Welche Adaptionen sind f¨ ur klassische Indexstrukturen in Hauptspeicherdatenbanken sinnvoll?
1. Volker Gaede und Oliver G¨
2. Kim, Changkyu, et al. 2010. FAST: Fast architecture sensitive tree search on modern CPUs and GPUs. SIGMOD Saake et al. Seminar on Modern Software Engineering and Database Concepts 12
D S E B
Databases Software Engineering and
Hybrid Storage In Practice (Campero)
Several commercial systems nowadays offer hybrid storage. This commonly means that data can be stored simultaneously in columnar format and in traditional index
mixed workloads, provided that hybrid designs are properly recommended. In this seminar topic we will review recent results and study a prototypical hybrid design
1. Dziedzic, Adam, Jingjing Wang, Sudipto Das, Bolin Ding, Vivek R. Narasayya, and Manoj Syamala. 2018. Columnstore and B+ tree-Are Hybrid Physical Designs Important? ICMD 2. Abadi, Daniel, Peter Boncz, Stavros Harizopoulos, Stratos Idreos, and Samuel Madden. 2013. The design and implementation of modern column-oriented database systems. Foundations and Trends in Databases Saake et al. Seminar on Modern Software Engineering and Database Concepts 13
D S E B
Databases Software Engineering and
Economic Games for Data Management (Campero)
In designing autonomous strategies for multi-user data systems, researchers have found that topics from Game Theory (like Pareto dominance) and from Economic Theory (like Supply/Demand modeling) can be useful to optimize the trade-offs in scheduling a common pool of resources across users with different workloads. In this seminar topic we’ll give a brief overview on basics from both fields, and we’ll study some components of a novel data management system following these ideas (NashDB). Looking forward we’ll consider how models from these fields could hold relevance for self-driving data management beyond the proposed system.
1. Marcus, Ryan, Olga Papaemmanouil, Sofiya Semenova, and Solomon Garber. 2018. NashDB: An End-to-End Economic Method for Elastic Database Fragmentation, Replication, and Provisioning. CDM 2. Pentaris, Fragkiskos, and Yannis Ioannidis. 2006. Query optimization in distributed networks of autonomous database systems. ACM Transactions on Database Systems Saake et al. Seminar on Modern Software Engineering and Database Concepts 14
D S E B
Databases Software Engineering and
Interactive Data Systems (Campero)
While traditional database systems are optimized for throughput or some analytic
with fast query interfaces (e.g. touchscreens or voice), and most optimizations must consider user interactions. In this seminar topic we will establish how these systems assist users in data exploration, we’ll review the challenges they face, some proposed benchmarks, and some system designs. The focus of the research can be determined by the student.
1. Crotty, Andrew, Alex Galakatos, Emanuel Zgraggen, Carsten Binnig, and Tim Kraska. 2016. The case for interactive data exploration accelerators (IDEAs). HLDA 2. Jiang, Lilong, Protiva Rahman, and Arnab Nandi. 2018. Evaluating Interactive Data Systems: Workloads, Metrics, and Guidelines. ICMD 3. El-Hindi, Muhammad, Zheguang Zhao, Carsten Binnig, and Tim Kraska. 2016. Vistrees: Dast indexes for interactive data exploration. HLDA 4. de Paula R.A. 2018. Visualization Techniques. Encyclopedia of Big Data Technologies Saake et al. Seminar on Modern Software Engineering and Database Concepts 15
D S E B
Databases Software Engineering and
In-Database Machine Learning (Campero)
Businesses have an ongoing interest in Machine Learning (ML), since it provides means to extracting value from large amounts of data with limited effort. However, for enterprise ML, there are several practical challenges, among them the need to move data from the operational databases to the ML systems. In-database ML (IDBML) is a relatively novel approach that solves such challenge. In this seminar topic we’ll list existing IDBML offerings, and we’ll study some underlying techniques such as leveraging UDFs or learning over joins.
1. Kumar, Arun, Matthias Boehm, and Jun Yang. 2017. Data management in machine learning: Challenges, techniques, and systems. ICMD 2. Hellerstein, Joseph M., Christoper R´ e, Florian Schoppmann, Daisy Zhe Wang, Eugene Fratkin, Aleksander Gorajek, Kee Siong Ng et al. 2012. The MADlib analytics library: or MAD skills, the SQL. VLDB Saake et al. Seminar on Modern Software Engineering and Database Concepts 16
D S E B
Databases Software Engineering and
Skew Handling for Parallel Entity Resolution (Chen)
Entity resolution (ER) is a process to identify records that refer to the same real-world
techniques and parallel computation should be both applied for it. To ease parallel computation, big data processing frameworks are used to implement a parallel entity resolution process. However, the sizes of blocks often vary much, which leads to an uneven workload among different nodes based on the default load balancing strategy
been designed for parallel ER to balance the workload of different nodes and reduce execution time.
1. Kolb, Lars, Andreas Thor, and Erhard Rahm. 2012. Load balancing for mapreduce-based entity resolution. ICDE 2. Yan, Wei, Yuan Xue, and Bradley Malin. 2013. Scalable load balancing for mapreduce-based record linkage. IPCCC 3. Mestre, Demetrio Gomes, and Carlos Eduardo Santos Pires. 2013. Improving load balancing for mapreduce-based entity matching. ISCC Saake et al. Seminar on Modern Software Engineering and Database Concepts 17
D S E B
Databases Software Engineering and
Active Learning for Entity Resolution (Chen)
Learning-Based entity resolution (ER) classifies record pairs into matches and non-matches using the classifier which is trained on a training dataset. However, the training dataset has to be labeled by human. In order to reduce human effort for labeling, active learning techniques are proposed, which requires much less training data to reach the same result quality compared to normal machine learning method. So far, different active learning approaches have been proposed for ER:
1. Sarawagi, Sunita, and Anuradha Bhamidipaty. ¨ Interactive deduplication using active learning. 2002. SIGKDD 2. Qian, Kun, Lucian Popa, and Prithviraj Sen. ¨ Active Learning for Large-Scale Entity Resolution. 2017. CIKM 3. De Freitas, Junio, et al. ¨ Active learning genetic programming for record deduplication. 2010. CEC Saake et al. Seminar on Modern Software Engineering and Database Concepts 18
D S E B
Databases Software Engineering and
Query optimization in heterogeneous hardware systems (Gurumurthy)
Current generation hardware are being exploited for processing queries efficiently. However, there is not one standard query processing technique present at the moment. In this topic, we explore the optimization strategies of different systems and summarize them for complete view of query processing in these environment.
1. Sebastian Breß, Max Heimel, Norbert Siegmund, Ladjel Bellatreche, Gunter Saake. 2014. GPU-accelerated Database Systems: Survey and Open Challenges? Transactions on Large-Scale Data- and Knowledge-Centered Systems 2. Tomas Karnagel, Dirk Habich, Wolfgang Lehner. 2015. Local vs. Global Optimization: Operator Placement Strategies in Heterogeneous Environments. EDBT/ICDT Saake et al. Seminar on Modern Software Engineering and Database Concepts 19
D S E B
Databases Software Engineering and
Industrial Interests in Systematic Software Reuse (Kr¨ uger)
Systematic software reuse in terms of software product lines is often only introduced after a larger set of different variants has evolved. For varying reasons, including cost reduction, faster development, or improved management, these variants are merged and integrated into a platform (reverse engineering). While there are several case studies that report on the migration processes and experiences, we still need a detailed analysis of the actual industrial motivations that lead to the adoption of product lines. To this end, we aim to analyze several years of the SPLC industry track to identify industrial case studies that are concerned with such migrations and identify the motivations of the organizations.
Saake et al. Seminar on Modern Software Engineering and Database Concepts 20
D S E B
Databases Software Engineering and
Automated Test Refactoring (Kr¨ uger)
Software is regularly updated or refactored, for example, to remove errors, introduce new features, or migrate towards a new technology. However, any change in the productive software also means that corresponding test cases may break or are not sufficient anymore. The purpose of this survey is to identify and summarize existing techniques on automated test case refacotring, meaning techniques that track code changes and support developers in maintaining the test cases for these artifacts.
A Test Case Refactoring Approach for Pattern-Based Software Development. Software Quality Journal
Extreme Programming Perspectives. Chapter Refactoring Test Code
Saake et al. Seminar on Modern Software Engineering and Database Concepts 21
D S E B
Databases Software Engineering and
Why do Developers use IDEs? (Kr¨ uger)
IDEs have emerged as the major tools used by developers to develop software. Arguably, they provide far more support compared to text editors or command line
suggest that they only use a minority of the IDEs’ functionalities. In this survey, we aim to summarize findings that report research in this direction. The main goal is to summarize what motivates developers to use an IDE and its functionalities. For instance, what refactorings are used to what extent or completely ignored in their daily work?
developers using the Elipse IDE?. IEEE Software
Saake et al. Seminar on Modern Software Engineering and Database Concepts 22
D S E B
Databases Software Engineering and
Comparing Data Stream Processing Architectures for Software Analytics (Schulze)
Software analytics (SA) are of increasing interest, as they allow to improve source code, development processes or even for decision-making. Such analytics heavily rely
such a mount of continuous data, stream processing architectures have been proposed. Task: Review data stream processing frameworks in the light of software analytics. Goal: Pros & cons of such frameworks for SA. Finally, a decision what framework/engine fits best given the inherent requirements of software analytics.
Development Analytics. ICSE
Software
(https://github.com/manuzhang/awesome-streaming)
Saake et al. Seminar on Modern Software Engineering and Database Concepts 23
D S E B
Databases Software Engineering and
Reasoning in Legal Ontologies (Wehnert)
Ontologies are conceptualizations of entities and relationships of a domain. They can be highly formalized and often provide reasoning / inference capabilities. The task is to find out which ontologies exist in the legal domain and which types of reasoning they support. Additionally, they shall be compared regarding their core components.
1. Wyner, Adam. 2008. An ontology in OWL for legal case-based reasoning. Artificial Intelligence and Law 2. Hoekstra, Rinke; Breuker, Joost; Di Bello, Marcello; Boer, Alexander et al. 2007. The LKIF Core Ontology
3. Ajani, Gianmaria; Boella, Guido; Caro, Luigi Di; Robaldo, Livio; Humphreys, Llio; Praduroux, Sabrina; Rossi, Piercarlo; Violato, Andre. 2016. The European Taxonomy Syllabus: A multi-lingual, multi-level
4. El Ghosh, Mirna; Naja, H; Abdulrab, H; Khalil, M. 2016. Towards a middle-out approach for building legal domain reference ontology. International Journal of Knowledge Engineering Saake et al. Seminar on Modern Software Engineering and Database Concepts 24
D S E B
Databases Software Engineering and
Themenwunsch an jkrueger@ovgu.de
Saake et al. Seminar on Modern Software Engineering and Database Concepts 25