Data Distribution Service | EPFL Graph Search

Data Distribution Service ou DDS est une norme (ou standard de jure) spécifiée par l'OMG (Object Management Group), dont le rôle est de proposer une technologie évoluée d'échanges de données via un réseau. Conceptuellement, les données échangées sont regroupées dans une notion d'Espace Global de Données EGD (ou GDS, Global Data Space en anglais). Contrairement aux technologies de base de données, cet espace de données est totalement distribué dans le réseau pour éviter les problématiques liées aux goulots étranglements et aux pannes des gestionnaires de données. Fonctionnellement, cette technologie, s'apparente aux systèmes de base de données relationnelles et distribuées mais elle relâche certaines contraintes d'intégrité jugées incongrus avec le traitement en temps réel, tel que le maintien de cohérence forte entre les copies des données ou les mécanismes d'isolations. Contrairement aux systèmes de base de données, DDS prône plutôt un modèle de cohérence dit modèle de (cohérence à

An Architecture for Load Balance in Computer Cluster Applications

Laurent Bindschaedler

Amid a data revolution that is transforming industries around the globe, computing systems have undergone a paradigm shift where many applications are scaled out to run on multiple computers in a computing cluster. As the storage and processing capabilities of a single machine are unable to keep pace with the amount of data, companies turn to distributed solutions to organize, persist, and analyze this Big Data. These new software solutions divide datasets into partitions that are processed in parallel on separate machines. A common problem in current cluster computing frameworks is load imbalance and limited parallelism due to skewed data distributions, processing times, and machine speeds. Load imbalance occurs when a few machines have more work and take longer to finish while the others remain idle, resulting in reduced overall performance and low resource utilization. The underlying cause for these issues is that data locality, where machines process the data stored locally, leads to tight coupling between a partition and the machine on which it is placed. This dissertation proposes a novel scatter architecture for computer cluster applications that effectively addresses the load imbalance problem and improves resource utilization. Scatter systems abandon data locality in favor of load balance. Existing systems first target locality, bringing computation to the data, and only then attempt to minimize load imbalance. Instead, we disregard any locality concerns and focus solely on achieving load balance by scattering all data across machines and bringing data to the computation as needed. The scatter architecture disaggregates compute and storage resources and pools all resources together across machines. We ensure storage load balance by spreading the data for each partition in small blocks across all storage devices and allow machines to retrieve data using an efficient, decentralized scheme. Scatter systems achieve compute load balance at runtime by dynamically adjusting the parallelism within a partition, either through work-sharing or by offloading background tasks to other machines. We design and implement three cluster applications inspired by the scatter architecture: a graph processing system that scales out using secondary storage, a general-purpose analytics framework, and a filesystem substrate that mitigates load imbalance for existing distributed databases. We demonstrate that each application can gracefully handle significant load imbalance with minimal performance degradation and that these applications can perform up to an order of magnitude faster than existing systems.

EPFL2020

Rock You like a Hurricane: Taming Skew in Large Scale Analytics

Laurent Bindschaedler, Ashvin Goel, Jasmina Malicevic, Nicolas Schiper, Willy Zwaenepoel

Current cluster computing frameworks suffer from load imbalance and limited parallelism due to skewed data distributions, processing times, and machine speeds. We observe that the underlying cause for these issues in current systems is that they partition work statically. Hurricane is a high-performance large-scale data analytics system that successfully tames skew in novel ways. Hurricane performs adaptive work partitioning based on load observed by nodes at runtime. Overloaded nodes can spawn clones of their tasks at any point during their execution, with each clone processing a subset of the original data. This allows the system to adapt to load imbalance and dynamically adjust task parallelism to gracefully handle skew. We support this design by spreading data across all nodes and allowing nodes to retrieve data in a decentralized way. The result is that Hurricane automatically balances load across tasks, ensuring fast completion times. We evaluate Hurricane’s performance on typical analytics workloads and show that it significantly outperforms state- of-the-art systems for both uniform and skewed datasets, because it ensures good CPU and storage utilization in all cases.

2018

Status of the ALICE CERN Analysis Facility

Marco Meoni

The ALICE experiment at CERN LHC is using a PROOF-enabled cluster for fast physics analysis, detector calibration and reconstruction of small data samples. The current system (CAF - CERN Analysis Facility) consists of some 120 CPU cores and about 45 TB of disk space distributed across the CAF hosts. One of the most important aspects of the data analysis on the CAF is the speed with which it can be carried out. The system is particularly aimed at the prototyping phase of analyses that need a high number of iterations and thus require a short response time. Quasi-online quality assurance of data can be obtained. The paper describes the design principles of the PROOF frame work and presents the current setup, performance tests and usage statistics. Subsets of selected data can be automatically staged in CAF from the Grid storage systems, therefore data distribution and staging techniques are described in depth. A fairshare algorithm to adjust the priorities of concurrently running sessions is also examined. Further more, the adaptation of PROOF to the AliEn/gLite Grid middleware is described. This approach enables a dynamic startup of PROOF nodes worldwide with the purpose to process much larger physics datasets.

Iop Publishing Ltd, Dirac House, Temple Back, Bristol Bs1 6Be, England2010

An Architecture for Load Balance in Computer Cluster Applications

Laurent Bindschaedler

EPFL2020