TY - CHAP A1 - Kruber, Nico A1 - Schintke, Florian A1 - Berlin, Michael T1 - A Relational Database Schema on the Transactional Key-Value Store Scalaris T2 - Proceedings of 2nd Workshop on Scalable Cloud Data Management Y1 - 2014 U6 - https://doi.org/10.1109/BigData.2014.7004441 ER - TY - JOUR A1 - Kramer, Tobias A1 - Läuter, Matthias A1 - Rubin, Martin A1 - Altwegg, Kathrin T1 - Seasonal changes of the volatile density in the coma and on the surface of comet 67P/Churyumov-Gerasimenko JF - Monthly Notices of the Royal Astronomical Society N2 - Starting from several monthly data sets of Rosetta’s COmetary Pressure Sensor we reconstruct the gas density in the coma around comet 67P/Churyumov-Gerasimenko. The underlying inverse gas model is constructed by fitting ten thousands of measurements to thousands of potential gas sources distributed across the entire nucleus surface. The ensuing self-consistent solution for the entire coma density and surface activity reproduces the temporal and spatial variations seen in the data for monthly periods with Pearson correlation coefficients of 0.93 and higher. For different seasonal illumination conditions before and after perihelion we observe a systematic shift of gas sources on the nucleus. Y1 - 2017 U6 - https://doi.org/10.1093/mnras/stx866 VL - 469 SP - S20 EP - S28 ER - TY - JOUR A1 - Zander, Thomas A1 - Dreischarf, Marcel A1 - Timm, Anne-Katrin A1 - Baumann, Wolfgang A1 - Schmidt, Hendrik T1 - Impact of material and morphological parameters on the mechanical response of the lumbar spine - A finite element sensitivity study JF - Journal of Biomechanics KW - finite element modelling KW - lumbar spine KW - sensitivity analysis KW - model validation KW - coefficient of determination Y1 - 2016 U6 - https://doi.org/10.1016/j.jbiomech.2016.12.014 VL - 53 SP - 185 EP - 190 ER - TY - JOUR A1 - Ritter, Zully A1 - Belavy, Daniel A1 - Baumann, Wolfgang A1 - Felsenberg, Dieter T1 - In vivo bone remodeling rates determination and compressive stiffness variations before, during 60 days bed rest and two years follow up: a micro-FE-analysis from HR-pQCT measurements of the Berlin bed rest study-2 JF - Acta Astronautica KW - finite-element analysis KW - bone structure KW - compressive stiffness KW - bed rest KW - HR-pQCT measurement Y1 - 2017 U6 - https://doi.org/10.1016/j.actaastro.2016.12.002 VL - 132 SP - 67 EP - 77 ER - TY - JOUR A1 - Kramer, Tobias A1 - Rodriguez, Mirta T1 - Two-dimensional electronic spectra of the photosynthetic apparatus of green sulfur bacteria JF - Scientific Reports N2 - Advances in time resolved spectroscopy have provided new insight into the energy transmission in natural photosynthetic complexes. Novel theoretical tools and models are being developed in order to explain the experimental results. We provide a model calculation for the two-dimensional electronic spectra of Cholorobaculum tepidum which correctly describes the main features and transfer time scales found in recent experiments. From our calculation one can infer the coupling of the antenna chlorosome with the environment and the coupling between the chlorosome and the Fenna-Matthews-Olson complex. We show that environment assisted transport between the subunits is the required mechanism to reproduce the experimental two-dimensional electronic spectra. Y1 - 2017 U6 - https://doi.org/10.1038/srep45245 VL - 7 SP - 45245 ER - TY - JOUR A1 - Ablyazimov, T. A1 - Abuhoza, A. A1 - Adak, R. P. A1 - Reinefeld, Alexander A1 - Schintke, Florian A1 - et al., T1 - Challenges in QCD matter physics – The scientific programme of the Compressed Baryonic Matter experiment at FAIR JF - The European Physical Journal A Y1 - 2017 U6 - https://doi.org/10.1140/epja/i2017-12248-y N1 - This paper has 587 authors. For the complete list of authors use the DOI, please. ER - TY - JOUR A1 - Salem, Farouk A1 - Schintke, Florian A1 - Schütt, Thorsten A1 - Reinefeld, Alexander T1 - Supporting various interconnects in FLESnet using Libfabric JF - CBM Progress Report 2016 Y1 - 2017 UR - https://repository.gsi.de/record/201318 SN - 978-3-9815227-4-7 SP - 159 EP - 160 ER - TY - JOUR A1 - Orgis, Thomas A1 - Läuter, Matthias A1 - Handorf, Dörthe A1 - Dethloff, Klaus T1 - Baroclinic waves on the β plane using low-order Discontinuous Galerkin discretisation JF - Journal of Computational Physics N2 - A classic Discontinuous Galerkin Method with low-order polynomials and explicit as well as semi-implicit time-stepping is applied to an atmospheric model employing the Euler equations on the β plane. The method, which was initially proposed without regard for the source terms and their balance with the pressure gradient that dominates atmospheric dynamics, needs to be adapted to be able to keep the combined geostrophic and hydrostatic balance in three spatial dimensions. This is achieved inside the discretisation through a polynomial mapping of both source and flux terms without imposing filters between time steps. After introduction and verification of this balancing, the realistic development of barotropic and baroclinic waves in the model is demonstrated, including the formation of a retrograde Rossby wave pattern. A prerequesite is the numerical solution of the thermal wind equation to construct geostrophically balanced initial states in z coordinates with arbitrary prescribed zonal wind profile, offering a new set of test cases for atmospheric models employing z coordinates. The resulting simulations demonstrate that the balanced low-order Discontinous Galerkin discretisation with polynomial degrees down to k=1 can be a viable option for atmospheric modelling. Y1 - 2017 U6 - https://doi.org/10.1016/j.jcp.2017.03.029 VL - 339 SP - 461 EP - 481 ER - TY - CHAP A1 - Coutinho, Jose A1 - Stillwell, Mark A1 - Argyraki, Katerina A1 - Ioannidis, George A1 - Iordache, Anca A1 - Kleineweber, Christoph A1 - Koliousis, Alexandros A1 - McGlone, John A1 - Pierre, Guillaume A1 - Ragusa, Carmelo A1 - Sanders, Peter A1 - Schütt, Thorsten A1 - Yu, Teng A1 - Wolf, Alexander T1 - The HARNESS Platform: A Hardware- and Network-Enhanced Software System for Cloud Computing T2 - Software Architecture for Big Data and the Cloud Y1 - 2017 SN - 9780128054673 SP - 323 EP - 351 PB - Morgan Kaufmann ER - TY - GEN A1 - Wende, Florian A1 - Laubender, Guido A1 - Steinke, Thomas T1 - Integration of Intel Xeon Phi Servers into the HLRN-III Complex: Experiences, Performance and Lessons Learned N2 - The third generation of the North German Supercomputing Alliance (HLRN) compute and storage facilities comprises a Cray XC30 architecture with exclusively Intel Ivy Bridge compute nodes. In the second phase, scheduled for November 2014, the HLRN-III configuration will undergo a substantial upgrade together with the option of integrating accelerator nodes into the system. To support the decision-making process, a four-node Intel Xeon Phi cluster is integrated into the present HLRN-III infrastructure at ZIB. This integration includes user/project management, file system access and job management via the HLRN-III batch system. For selected workloads, in-depth analysis, migration and optimization work on Xeon Phi is in progress. We will report our experiences and lessons learned within the Xeon Phi installation and integration process. For selected examples, initial results of the application evaluation on the Xeon Phi cluster platform will be discussed. T3 - ZIB-Report - 14-15 KW - Performance and usage measurement KW - System management KW - System integration KW - Xeon Phi cluster Y1 - 2014 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:0297-zib-49990 UR - https://cug.org/proceedings/cug2014_proceedings/includes/files/pap194-file2.pdf SN - 1438-0064 ER - TY - JOUR A1 - Hermann, Frank A1 - Ehrig, Hartmut A1 - Golas, Ulrike A1 - Orejas, Fernando T1 - Formal Analysis of Model Transformations Based on Triple Graph Grammars JF - Mathematical Structures in Computer Science Y1 - 2014 U6 - https://doi.org/10.1017/S0960129512000370 VL - 24 IS - 4 SP - 1 EP - 57 ER - TY - JOUR A1 - Ehrig, Hartmut A1 - Golas, Ulrike A1 - Habel, A. A1 - Lambers, L. A1 - Orejas, Fernando T1 - M-Adhesive Transformation Systems with Nested Application Conditions. Part 1: Parallelism, Concurreny and Amalgamation JF - Mathematical Structures in Computer Science Y1 - 2014 U6 - https://doi.org/10.1017/S0960129512000357 VL - 24 IS - 4 SP - 1 EP - 48 ER - TY - JOUR A1 - Gabriel, Karsten A1 - Braatz, Benjamin A1 - Ehrig, Hartmut A1 - Golas, Ulrike T1 - Finitary M-Adhesive Categories JF - Mathematical Structures in Computer Science Y1 - 2014 U6 - https://doi.org/10.1017/S0960129512000321 VL - 24 IS - 4 SP - 1 EP - 40 ER - TY - CHAP A1 - Kleineweber, Christoph A1 - Reinefeld, Alexander A1 - Schütt, Thorsten T1 - QoS-aware Storage Virtualization for Cloud File Systems T2 - Proceedings of the 1st ACM International Workshop on Programmable File Systems Y1 - 2014 U6 - https://doi.org/10.1145/2603941.2603944 SP - 19 EP - 26 ER - TY - CHAP A1 - Dreßler, Sebastian A1 - Steinke, Thomas T1 - An Automated Approach for Estimating the Memory Footprint of Non-linear Data Objects T2 - Euro-Par 2013: Parallel Processing Workshops Y1 - 2014 U6 - https://doi.org/10.1007/978-3-642-54420-0_25 VL - 8374 SP - 249 EP - 258 ER - TY - THES A1 - Fischer, Jens V. T1 - A Gossiping Framework for Scalaris Y1 - 2014 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:0297-zib-50685 ER - TY - JOUR A1 - Golas, Ulrike A1 - Habel, A. A1 - Ehrig, Hartmut T1 - Multi-Amalgamation of Rules with Application Conditions in M-Adhesive Categories JF - Mathematical Structures in Computer Science Y1 - 2014 U6 - https://doi.org/10.1017/S0960129512000345 VL - 24 IS - 4 SP - 1 EP - 68 ER - TY - GEN A1 - Wende, Florian A1 - Steinke, Thomas A1 - Cordes, Frank T1 - Multi-threaded Kernel Offloading to GPGPU Using Hyper-Q on Kepler Architecture N2 - Small-scale computations usually cannot fully utilize the compute capabilities of modern GPGPUs. With the Fermi GPU architecture Nvidia introduced the concurrent kernel execution feature allowing up to 16 GPU kernels to execute simultaneously on a shared GPU device for a better utilization of the respective resources. Insufficient scheduling capabilities in this respect, however, can significantly reduce the theoretical concurrency level. With the Kepler GPU architecture Nvidia addresses this issue by introducing the Hyper-Q feature with 32 hardware managed work queues for concurrent kernel execution. We investigate the Hyper-Q feature within heterogeneous workloads with multiple concurrent host threads or processes offloading computations to the GPU each. By means of a synthetic benchmark kernel and a hybrid parallel CPU-GPU real-world application, we evaluate the performance obtained with Hyper-Q on GPU and compare it against a kernel reordering mechanism introduced by the authors for the Fermi architecture. T3 - ZIB-Report - 14-19 KW - GPGPU KW - Hyper-Q KW - Concurrent Kernel Execution Y1 - 2014 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:0297-zib-50362 SN - 1438-0064 ER - TY - GEN A1 - Noack, Matthias T1 - HAM - Heterogenous Active Messages for Efficient Offloading on the Intel Xeon Phi N2 - The applicability of accelerators is limited by the attainable speed-up for the offloaded computations and by the offloading overheads. While GPU programming models like CUDA and OpenCL only allow to optimise the application code and its speed-up, the available low-level APIs for the Intel Xeon Phi provide opportunity to address the overheads, too. This work presents an Heterogeneous Active Message (HAM) layer that minimises software overheads for offloading on Intel’s Xeon Phi. It provides the basis for an offload API with similar semantics as the Intel Language Extensions for Offload (LEO). In contrast to LEO, HAM works within the C++ language and needs no additional compiler support. We evaluated HAM on top of SCIF and MPI as communication backends. While the SCIF backend offers the best performance, the MPI backend allows for inter-node offloads which are not possible with other offload solutions. Benchmark results show that the cost for offloading a function call can be decreased by a factor up to 18 compared with LEO. T3 - ZIB-Report - 14-23 KW - Intel Xeon Phi KW - MIC KW - Active message KW - Offloading KW - SCIF KW - MPI KW - LEO KW - HAM Y1 - 2014 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:0297-zib-50610 SN - 1438-0064 ER - TY - CHAP A1 - Schäfer, Patrick T1 - Towards Time Series Classification without Human Preprocessing T2 - MLDM 2014 Y1 - 2014 U6 - https://doi.org/10.1007/978-3-319-08979-9_18 VL - 8556 SP - 228 EP - 242 ER - TY - JOUR A1 - Potamitis, Ilyas A1 - Schäfer, Patrick T1 - On Classifying Insects from their Wing-Beat: New Results JF - Ecology and acoustics: emergent properties from community to landscape, Paris, France Y1 - 2014 ER - TY - JOUR A1 - Krüger, Jens A1 - Grunzke, Richard A1 - Gesing, Sandra A1 - Breuers, Sebastian A1 - Brinkmann, André A1 - de la Garza, Luis A1 - Kohlbacher, Oliver A1 - Kruse, Martin A1 - Nagel, Wolfgang A1 - Packschies, Lars A1 - Müller-Pfefferkorn, Ralph A1 - Schäfer, Patrick A1 - Schärfe, Charlotta A1 - Steinke, Thomas A1 - Schlemmer, Tobias A1 - Warzecha, Klaus Dieter A1 - Zink, Andreas A1 - Herres-Pawlis, Sonja T1 - The MoSGrid Science Gateway - A Complete Solution for Molecular Simulations JF - Journal of Chemical Theory and Computation Y1 - 2014 U6 - https://doi.org/10.1021/ct500159h VL - 10 IS - 6 SP - 2232 EP - 2245 ER - TY - THES A1 - Lorenz, Christian T1 - Leistungsverbesserung der Datenübertragung bei der Replikation von nicht veränderlichen Dateien in einem verteilten Dateisystem Y1 - 2009 ER - TY - THES A1 - Hoffmann, Marie T1 - Ein agglomeratives, gossipbasiertes Clustering-Verfahren für verteilte Systeme implementiert in Scalaris Y1 - 2009 ER - TY - THES A1 - Münnich, Bert T1 - Kürzeste-Wege-Suche mit Grafikkarten Y1 - 2011 ER - TY - THES A1 - Langner, Felix T1 - Überlastungsschutz mit Anfragepriorisierung für das Internetdateisystem XtreemFS Y1 - 2012 ER - TY - THES A1 - Beick, Matthias T1 - On-Line Processing of Streams and Time Series – An Evaluation Y1 - 2010 ER - TY - THES A1 - Borchert, Sebastian T1 - Generische Templates zur Kommunikation in heterogenen Manycore-Systemen Y1 - 2011 ER - TY - THES A1 - Merkulow, Igor T1 - Parameter selection and evaluation for programming of heterogeneous architectures Y1 - 2013 ER - TY - JOUR A1 - Härtig, Hermann A1 - Matsuoka, Satoshi A1 - Mueller, Frank A1 - Reinefeld, Alexander T1 - Resilience in Exascale Computing (Dagstuhl Seminar 14402) JF - Dagstuhl Reports Y1 - 2015 U6 - https://doi.org/10.4230/DagRep.4.9.124 VL - 4 IS - 9 SP - 124 EP - 139 ER - TY - JOUR A1 - Tuma, Christian A1 - Sauer, Joachim T1 - Quantum chemical ab initio prediction of proton exchange barriers between CH4 and different H-zeolites JF - The Journal of Chemical Physics Y1 - 2015 U6 - https://doi.org/10.1063/1.4923086 VL - 143 IS - 10 SP - 102810-1 EP - 102810-13 ER - TY - CHAP A1 - Wende, Florian A1 - Steinke, Thomas A1 - Reinefeld, Alexander ED - Gray, A. ED - Smith, L. ED - Weiland, M. T1 - The Impact of Process Placement and Oversubscription on Application Performance: A Case Study for Exascale Computing T2 - Proceedings of the 3rd International Conference on Exascale Applications and Software, EASC 2015 Y1 - 2015 SN - 978 -0-9 926615 -1-9 SP - 13 EP - 18 PB - The University of Edinburgh ER - TY - CHAP A1 - Noack, Matthias A1 - Wende, Florian A1 - Oertel, Klaus-Dieter ED - Reinders, James ED - Jeffers, Jim T1 - OpenCL: There and Back Again T2 - High Performance Parallelism Pearls Y1 - 2015 SN - 978-0-12-803819-2 VL - 2 SP - 355 EP - 378 PB - Morgan Kaufman, Elsevier ER - TY - JOUR A1 - Kramer, Tobias A1 - Noack, Matthias T1 - Prevailing Dust-transport Directions on Comet 67P/Churyumov–Gerasimenko JF - The Astrophysical Journal Letters N2 - Dust transport and deposition behind larger boulders on the comet 67P/Churyumov–Gerasimenko (67P/C–G) have been observed by the Rosetta mission. We present a mechanism for dust-transport vectors based on a homogeneous surface activity model incorporating in detail the topography of 67P/C–G. The combination of gravitation, gas drag, and Coriolis force leads to specific dust transfer pathways, which for higher dust velocities fuel the near-nucleus coma. By distributing dust sources homogeneously across the whole cometary surface, we derive a global dust-transport map of 67P/C–G. The transport vectors are in agreement with the reported wind-tail directions in the Philae descent area. Y1 - 2015 U6 - https://doi.org/10.1088/2041-8205/813/2/L33 VL - 813 IS - 2 SP - L33 ER - TY - JOUR A1 - Schäfer, Patrick T1 - Scalable time series classification JF - Data Mining and Knowledge Discovery N2 - Time series classification tries to mimic the human understanding of similarity. When it comes to long or larger time series datasets, state-of-the-art classifiers reach their limits because of unreasonably high training or testing times. One representative example is the 1-nearest-neighbor dynamic time warping classifier (1-NN DTW) that is commonly used as the benchmark to compare to. It has several shortcomings: it has a quadratic time complexity in the time series length and its accuracy degenerates in the presence of noise. To reduce the computational complexity, early abandoning techniques, cascading lower bounds, or recently, a nearest centroid classifier have been introduced. Still, classification times on datasets of a few thousand time series are in the order of hours. We present our Bag-Of-SFA-Symbols in Vector Space classifier that is accurate, fast and robust to noise. We show that it is significantly more accurate than 1-NN DTW while being multiple orders of magnitude faster. Its low computational complexity combined with its good classification accuracy makes it relevant for use cases like long or large amounts of time series or real-time analytics. Y1 - 2015 U6 - https://doi.org/10.1007/s10618-015-0441-y SP - 1 EP - 26 ER - TY - CHAP A1 - Kruber, Nico A1 - Lange, Maik A1 - Schintke, Florian T1 - Approximate Hash-Based Set Reconciliation for Distributed Replica Repair T2 - 2015 IEEE 34th International Symposium on Reliable Distributed Systems (SRDS) Y1 - 2015 U6 - https://doi.org/10.1109/SRDS.2015.30 SP - 166 EP - 175 PB - IEEE ER - TY - BOOK A1 - Ehrig, Hartmut A1 - Ermel, Claudia A1 - Golas, Ulrike A1 - Hermann, Frank T1 - Graph and Model Transformation BT - General Framework and Applications N2 - This book is a comprehensive explanation of graph and model transformation. It contains a detailed introduction, including basic results and applications of the algebraic theory of graph transformations, and references to the historical context. Then in the main part the book contains detailed chapters on M-adhesive categories, M-adhesive transformation systems, and multi-amalgamated transformations, and model transformation based on triple graph grammars. In the final part of the book the authors examine application of the techniques in various domains, including chapters on case studies and tool support. The book will be of interest to researchers and practitioners in the areas of theoretical computer science, software engineering, concurrent and distributed systems, and visual modelling. Y1 - 2015 SN - 978-3-662-47979-7 U6 - https://doi.org/10.1007/978-3-662-47980-3 PB - Springer-Verlag Berlin Heidelberg ER - TY - THES A1 - Schäfer, Patrick T1 - Scalable Time Series Similarity Search for Data Analytics Y1 - 2015 UR - http://edoc.hu-berlin.de/docviews/abstract.php?id=42117 ER - TY - THES A1 - Lange, Maik T1 - Effiziente Reparatur von Repliken in Distributed Hash Tables N2 - Eine distributed hash Table (DHT) stellt eine über multiple Rechner strukturiert verteilte Tabelle von Schlüssel-Wert-Paaren dar. Ein verteiltes System unterliegt zumeist einer Teilnehmerdynamik in Form von ausfallenden oder beitretenden Knoten, wodurch Daten verloren gehen können. Zur Sicherstellung der Verfügbarkeit aller Schlüssel-Wert-Paare ist Redundanz notwendig. Die Einführung einer Redundanzschicht erhöht die Verfügbarkeit der Daten erheblich auf Kosten einer verringerten Leistungsfähigkeit. Der Fokus dieser Arbeit liegt im Erhalt der Redundanz, was die Regeneration verlorener Kopien sowie eine Aktualisierung veralteter Kopien umfasst. Die Regeneration verlorener Kopien ist notwendig um die Verfügbarkeit langfristig zu sichern. Die Aktualisierung ist optional und verbessert die Leselatenz in geografisch weit verteilten Systemen. Die vorgestellte Lösung besteht aus drei Komponenten. Der Set Reconciliation Dienst ermöglicht den effizienten Vergleich zweier Knoten zur Identifikation verlorener und veralteter Kopien. Der Aktualisierungsdienst synchronisiert die verglichenen Knoten durch die Übertragung der Differenzen. Die dritte Komponente ist ein Anti-Entropie Protokoll, dass die Auswahl eines jeden Knotens als Synchronisationspartner sicherstellt. Das Ziel dieser Arbeit liegt in dem Vergleich dreier Set Reconciliation Algorithmen, deren Eignung und Leistungsfähigkeit für die Erkennung verlorener und veralteter Kopien evaluiert werden soll. Ausgewählt wurden der Bloom Filter, der Merkle-Tree sowie deren Kombination der Approximate Reconciliation Tree. Die Evaluation zeigte, dass der Merkle-Tree die höchste Genauigkeit besitzt und spätestens nach einer initiierten Synchronisation eines jeden Knotens zur vollständigen Synchronität in der DHT führt. Der Bloom Filter besitzt bei der Erkennung verlorener Repliken nur die halbe Genauigkeit gegenüber veralteten Repliken, wodurch die Konvergenzgeschwindigkeit dementsprechend geringer ist gegenüber dem Merkle-Tree. Dafür sind die Übertragungskosten des Bloom Filter um bis zu 80% geringer als beim Merkle-Tree und weisen kaum nennenswerte Spitzen auf. Die Übertragungskosten des Merkle-Tree sind abhängig von der Anzahl der Differenzen, sodass bei einer Differenz von Null nur eine Signatur mit weniger als einem kByte übertragen werden muss. Ist die Differenz größer, steigen die Kosten je nach Bucket-Größe an. Der Approximate Reconciliation Tree ist ein als Bloom Filter codierter Merkle-Tree, der je nach Parameterkonstellation die Eigenschaften einer der beiden Strukturen aufweist. Die Bloom Filter ähnlichen Konfigurationen bieten keinen Vorteil, wohingegen die Merkle-Tree ähnlichen nur zwei Mitteilungsrunden benötigen bei einer leichten Genauigkeitsverringerung gegenüber dem Merkle-Tree. Y1 - 2012 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:0297-zib-42275 ER - TY - THES A1 - Keidel, Stefan T1 - Snapshots in Scalaris N2 - Eines der größten Hindernisse beim praktischen Einsatz von Scalaris, einer skalierbaren Implementierung einer verteilten Hashtabelle mit Unterstützung für Transaktionen, ist das Fehlen eines Verfahrens zur Aufnahme eines konsistenten Zustandes des gesamten Systems. Wir stellen in dieser Arbeit ein einfaches Protokoll vor, dass diese Aufgabe erfüllt und sich, auf Grund der von uns gewählten Herangehensweise, leicht implementieren lässt. Als Ausgangspunkt dafür wählen wir aus einer Reihe von „klassischen“ Snapshot-Algorithmen ein 1993 von Mattern entworfenes Verfahren, welches auf dem Algorithmus von Lai und Yang basiert, aus. Diese Entscheidung basiert auf einer gründlichen Analyse der Protokolle unter Berücksichtigung der Architektur der existierenden Software. Im nächsten Arbeitsschritt benutzen wir unser vollständiges Wissen über die Interna des Transaktionssystems von Scalaris und vereinfachen damit das Verfahren hinsichtlich Benutzbarkeit und Implementierungskomplexität, ohne die Anforderungen an den aufgenommenen Zustand aufzuweichen. Statt einer losen Anhäufung lokaler Zustände der einzelnen Teilnehmerknoten können wir am Ende eine große Schlüssel-Wert-Tabelle als Ergebnis erzeugen, die konsistent ist, sich leicht weiterverarbeiten lässt und die einem Zustand entspricht, in dem sich das System einmal befunden haben könnte. Nachdem wir das Verfahren dann in Software umgesetzt haben, werten wir die Ergebnisse hinsichtlich des Einflusses auf die Performanz des Gesamtsystems aus und diskutieren mögliche Weiterentwicklungen. KW - scalaris KW - dht KW - algorithm Y1 - 2013 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:0297-zib-42282 ER - TY - THES A1 - Hupfeld, Felix T1 - Causal weak-consistency replication – a systems approach Y1 - 2009 U6 - http://nbn-resolving.de/urn/resolver.pl?nbn:de:kobv:11-100100235 UR - http://edoc.hu-berlin.de/docviews/abstract.php?id=29970 ER - TY - THES A1 - Hoffmann, Marie T1 - Approximate Algorithms for Distributed Systems N2 - Peer-to-peer (P2P) systems form a special class of distributed systems. Typically, nodes in a P2P system are flat and share the same responsabilities. In this thesis we focus on three problems that occur in P2P systems: the storage of data replicates, quantile computation on distributed data streams, and churn rate estimation. Data replication is one of the oldest techniques to maintain stored data in a P2P system and to reply to read requests. Applications, which use data replication are distributed databases. They are part of an abstract overlay network and do not see the underlying network topology. The question is how to place a set of data replicates in a distributed system such that response times and failure probabilities become minimal without a priori knowledge of the topology of the underlying hardware nodes? We show how to utilize an agglomerative clustering procedure to reach this goal. State-of-the-art algorithms for aggregation of distributed data or data streams require at some point synchronization, or merge data aggregates hierarchically, which does not accompany the basic principle of P2P systems. We test whether randomized communication and merging of data aggregates are able to produce the same results. These data aggregates serve for quantile queries. Constituting and maintaining a P2P overlay network requires frequent message passing. It is a goal to minimize the number of maintenance messages since they consume bandwidth which might be missing for other applications. The lower bound of the frequency for mainte- nance messages is highly dependent on the churn rate of peers. We show how to estimate the mean lifetime of peers and to reduce the frequency for maintenance messages without destabilizing the infrastructure of the constituting overlay. KW - peer-to-peer, machine learning, approximate, clustering, quantile, linear regression Y1 - 2013 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:0297-zib-42370 ER - TY - THES A1 - Berlin, Michael T1 - Popularitätsbasierte Anpassung des Replikationsfaktors in verteilten Dateisystemen T1 - Popularity-based Adjustment of the Replication Factor in Distributed File Systems N2 - Verteilte Dateisysteme erlauben es zusätzliche Replikate von Dateien anzulegen, um Client-Zugriffe auf mehrere Server zu verteilen. Bei einer einheitlichen Festlegung des Replikationsfaktors für alle Dateien wird der beschränkte verfügbare Speicherplatz jedoch nicht effizient genutzt, da es nur einen geringen Anteil von Dateien gibt, die häufig angefragt werden, während die restlichen Dateien unnötig Speicherplatz belegen. Der erste Teil dieser Arbeit beschreibt daher ein Verfahren den Replikationsfaktor einer Datei in Abhängigkeit ihrer Popularität festzulegen. Im Kontext des verteilten Dateisystems XtreemFS wurde dazu ein zentraler Dienst für die dynamische Replikation implementiert und mit einem Trace simuliert. Die Auswertung zeigt, dass die popularitätsbasierte Anpassung ein geeignetes Verfahren ist den Replikationsfaktor dynamisch festzulegen, da bei einer wöchentlichen Ermittlung der Popularität weniger als durchschnittlich 3,1% zusätzliche Replikate angelegt wurden. Da der Dienst selbst nicht mit einer beliebigen Anzahl von Dateien skaliert, muss er in mehrere Partitionen unterteilt und dezentral ausgeführt werden. Darum entwickelt der zweite Teil der Arbeit eine dezentrale Variante des Dienstes und zeigt, dass der Replikationsfaktor identisch zum zentralen Fall festgelegt werden kann. Nur im Fall eines Netzwerkausfalls oder nicht synchronisierter lokaler Uhren tritt eine Abweichung der Replikationsfaktoren auf, deren Ausmaß in der Simulation untersucht wurde. N2 - Distributed file systems allow the creation of additional file replicas in order to spread requests across several servers. However, setting a uniform number of replicas for every file would not utilize the available storage space in an efficient manner. As there are only few popular files, the majority of files would unnecessarily use up the limited space. Therefore the first part of this thesis presents a method to adjust the replication factor depending on the popularity of the files. This method was implemented as a central service for dynamic replication in the context of the distributed file system XtreemFS. The service was then simulated using a trace. A weekly observation of the popularity resulted in an average creation rate of 3.1% compared to all replicas. This shows that popularity-based adjustment is a feasible way to set the replication factor dynamically. As the service itself does not scale with the number of observed files, it needs to be partitioned and executed in a distributed manner. Therefore the second part of this thesis develops a distributed version of the service and shows that it determines replication factors equal to the result of the central version. However, the replication factor may vary if there is a network outage or a drift of the local clocks. The degree of variation was analyzed in simulations. Y1 - 2011 N1 - Thesis for the German Diplom degree (comparable to a Master degree) in Computer Science. ER - TY - JOUR A1 - Grunzke, Richard A1 - Breuers, Sebastian A1 - Gesing, Sandra A1 - Herres-Pawlis, Sonja A1 - Kruse, Martin A1 - Blunk, Dirk A1 - de la Garza, Luis A1 - Packschies, Lars A1 - Schäfer, Patrick A1 - Schärfe, Charlotta A1 - Schlemmer, Tobias A1 - Steinke, Thomas A1 - Schuller, Bernd A1 - Müller-Pfefferkorn, Ralph A1 - Jäkel, René A1 - Nagel, Wolfgang A1 - Atkinson, Malcolm A1 - Krüger, Jens T1 - Standards-based metadata management for molecular simulations JF - Concurrency and Computation: Practice and Experience Y1 - 2013 U6 - https://doi.org/10.1002/cpe.3116 ER - TY - CHAP A1 - Weinhold, Carsten A1 - Lackorzynski, Adam A1 - Bierbaum, Jan A1 - Küttler, Martin A1 - Planeta, Maksym A1 - Härtig, Hermann A1 - Shiloh, Amnon A1 - Levy, Ely A1 - Ben-Nun, Tal A1 - Barak, Amnon A1 - Steinke, Thomas A1 - Schütt, Thorsten A1 - Fajerski, Jan A1 - Reinefeld, Alexander A1 - Lieber, Matthias A1 - Nagel, Wolfgang T1 - FFMK: A Fast and Fault-tolerant Microkernel-based System for Exascale Computing T2 - SPPEXA Symposium 2016 Y1 - 2016 U6 - https://doi.org/10.1007/978-3-319-40528-5_18 ER - TY - CHAP A1 - Fajerski, J. A1 - Noack, Matthias A1 - Reinefeld, Alexander A1 - Schintke, Florian A1 - Schütt, Thorsten A1 - Steinke, Thomas T1 - Fast In-Memory Checkpointing with POSIX API for Legacy Exascale-Applications T2 - SPPEXA Symposium 2016 Y1 - 2016 U6 - https://doi.org/10.1007/978-3-319-40528-5_19 ER - TY - THES A1 - Bärhold, Robert T1 - Konzipierung und Implementierung eines Quota-Protokolls in objektbasierten Dateisystemen N2 - Verteilte Dateisysteme, wie sie unter anderem in Cloud-Umgebungen eingesetzt werden, werden häufig von verschiedenen Anwendern und Anwendungen zeitgleich verwendet. Da diese Ressourcen in Cloud-Umgebungen gemietet sind, muss sichergestellt werden, dass ein Kunde nur die Ressourcen verwenden kann, für die er bezahlt hat. Andernfalls könnten andere Kunden beeinträchtigt werden. Solch eine Obergrenze wird auch als Quota bezeichnet. Speziell für objektbasierte Dateisysteme, bei denen die Metadaten einer Datei von den Dateiinhalten getrennt gespeichert und verwaltet werden, erfordert die Forcierung einer Quota ein verteiltes Protokoll. Im Rahmen der vorliegenden Arbeit wurde ein Quota-Protokoll für objektbasierte Dateisysteme entwickelt und implementiert. Bei einer aktiven Quota wird zunächst ein Teil des Speicherplatzes für eine bestimmte Datei blockiert, der als Voucher bezeichnet wird. Anschließend wird dem Client die maximale Dateigröße in signierter Form zugewiesen. Sie ergibt sich aus der Summe der bisherigen Dateigröße und allen erstellten Vouchers. Durch diesen Ansatz können die Speicherserver bei jedem Zugriff lokal prüfen, ob der Schreibzugriff die maximale Dateigröße überschreiten würde. Reicht diese nicht aus, muss ein zusätzlicher Voucher erstellt werden, sofern noch Speicherplatz verfügbar ist. Nicht verbrauchter Speicherplatz wird am Ende wieder freigegeben. Zusätzlich unterstützt das Protokoll die Verteilung der Inhalte einer Datei auf mehrere Server (Striping) sowie Sparse-Dateien, bei denen beliebige Bereiche der Datei leer bleiben. Die Implementierung erfolgte im objektbasierten Dateisystem XtreemFS. Dabei wurden spezielle Funktionen von XtreemFS, wie die Dateireplikation, mit einbezogen. In einer abschließenden Evaluation wurde der Einfluss des implementierten Quota- Protokolls bezüglich der Performanz von XtreemFS untersucht. Dabei zeigte sich zunächst, dass bei einer aktiven Quota die Performanz beim Schreiben großer Daten- mengen abhängig von der eingestellten Voucher-Größe ist. Ist diese zu niedrig gewählt, so muss der Client seine maximale Dateigröße durch zusätzliche Anfragen beim Metada- tenserver erhöhen, was die Performanz negativ beeinträchtigt. Bei einer Voucher-Größe von 100 MiB wurde beim Schreiben einer 10 GiB-Datei ein Performanzverlust von unter 2% gemessen, verglichen mit dem Entwicklungsstand ohne implementiertes Quota- Protokoll. Bei der Erstellung mehrerer 64 KiB-Dateien wurde aufgrund zusätzlicher Nachrichten ein Verlust von rund 8% gemessen. KW - Quota KW - Protokoll KW - Verteiltes Dateisystem Y1 - 2016 ER - TY - GEN A1 - Clausecker, Robert T1 - Notes on the Construction of Pattern Databases N2 - This report aims to explain and extend pattern databases as a tool in finding optimal solutions to combination puzzles. Using sliding tile puzzles as an example, an overview over techniques to construct and use pattern databases is given, analysed, and extended. As novel results, zero aware additive pattern databases and PDB catalogues are introduced. This report originally appeared as a bachelor thesis at Humboldt University of Berlin under the title Notes on the Construction of Tablebases. T3 - ZIB-Report - 17-59 KW - pattern-database 15-puzzle 24-puzzle ida* heuristic-search pdb-catalogue zpdb Y1 - 2017 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:0297-zib-65586 SN - 1438-0064 ER - TY - GEN A1 - Skrzypczak, Jan T1 - Weakening Paxos Consensus Sequences for Commutative Commands N2 - Consensus (agreement on a value) is regarded as a fundamental primitive in the design of fault tolerant distributed systems. A well-known solution to the consensus problem is Paxos. Extensions of the Paxos algorithm make it possible to reach agreement on a sequence of commands which can then be applied on a replicated state. However, concurrently proposed commands can create conflicts that must be resolved by ordering them. This thesis delivers an in-depth description of a Paxos-based algorithm to establish such command sequences, called Paxos Round Based Register (PRBR). In contrast to conventional approaches like Multi-Paxos, PRBR can manage multiple command sequences independently. Furthermore, each sequence is established in-place, which eliminates the need for managing multiple Paxos instances. PRBR is extended as part of this thesis to exploit the commutativity of concurrently proposed commands. As a result, conflict potential can be greatly reduced which increases the number of commands that can be handled by PRBR. This is shown for a number of workloads in an experimental evaluation. T3 - ZIB-Report - 17-64 KW - Consensus KW - Paxos Y1 - 2017 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:0297-zib-65741 SN - 1438-0064 ER - TY - THES A1 - Skrzypczak, Jan T1 - Weakening Paxos Consensus Sequences for Commutative Commands N2 - Consensus (agreement on a value) is regarded as a fundamental primitive in the design of fault tolerant distributed systems. A well-known solution to the consensus problem is Paxos. Extensions of the Paxos algorithm make it possible to reach agreement on a sequence of commands which can then be applied on a replicated state. However, concurrently proposed commands can create conflicts that must be resolved by ordering them. This thesis delivers an in-depth description of a Paxos-based algorithm to establish such command sequences, called Paxos Round Based Register (PRBR). In contrast to conventional approaches like Multi-Paxos, PRBR can manage multiple command sequences independently. Furthermore, each sequence is established in-place, which eliminates the need for managing multiple Paxos instances. PRBR is extended as part of this thesis to exploit the commutativity of concurrently proposed commands. As a result, conflict potential can be greatly reduced which increases the number of commands that can be handled by PRBR. This is shown for a number of workloads in an experimental evaluation. KW - Paxos KW - Consensus Y1 - 2017 ER - TY - CHAP A1 - Noack, Matthias T1 - OpenCL in Scientific High Performance Computing: The Good, the Bad, and the Ugly T2 - Proceedings of the 5th International Workshop on OpenCL N2 - For writing a new scientific application, portability across existing and future hardware should be the major design goal, as there is a multitude of different compute devices, and programme codes typically outlive systems by far. Unlike other programming models that address parallelism or heterogeneity, OpenCL does provide practical portability across a wide range of HPC-relevant architectures. Other than that, it has a range of further advantages like being a library-only implementation, and using runtime kernel-compilation. We present experiences with utilising OpenCL alongside C++, MPI, and CMake in two real-world scientific codes. Our targets are a Cray XC40 supercomputer with multi- and many-core (Xeon Phi) CPUs, as well as multiple smaller systems with Nvidia and AMD GPUs. We shed light on practical issues arising in such a scenario, like the interaction between OpenCL and MPI, discuss solutions, and point out current limitations of OpenCL in the domain of scientific HPC from an application developer's and user's point of view. Y1 - 2017 U6 - https://doi.org/10.1145/3078155.3078170 SP - 12:1 EP - 12:3 ER -