Logo des Repositoriums

Auflistung nach Autor:in "Sobe, Peter"

1 - 8 von 8
  • Zeitschriftenartikel
    Coding for Reliable Data Storage on Different Hardware Platforms
    (FERS-Mitteilungen: Vol. 28, No. 1, 2010) Sobe, Peter
    Peter Sobe, University of Luebeck, Germany Institute of Computer Engineering sobe@iti.uni-luebeck.de Abstract
  • Zeitschriftenartikel
    A Distributed Hash Table using One-sided Communication in MPI
    (PARS-Mitteilungen: Vol. 34, Nr. 1, 2017) Sobe, Peter; Graupner, Tom; Hennig, Florian
    The Message Passing Interface (MPI) can be applied to implement data structures that are distributed across process memory, such as distributed arrays or hash tables. In this paper a hash table implementation is described that employs one-sided communication in case of collision-free access. Collisions of data entries within the hash table are treated using dynamic overflow memory and two-sided communication. This leads to a two-level communication architecture that combines one-sided and two-sided operations in a data structure and the related access operations. This approach circumvents the problem of dynamic and unforeseen size and arrangement of data structures in shared memory that would be hard to manage using solely one-sided communication.
  • Konferenzbeitrag
    File sharing using IP-multicast
    (ARCS 2004 – Organic and pervasive computing, 2004) Trojahner, Kai; Sobe, Peter
    File sharing systems cause a huge portion of traffic in the Internet. With respect to the peer-to-peer approach, unicast delivery of content is the common case. Unfortunately, an inherent characteristic of filesharing systems employing unicast communication is that the sum of all download rates cannot exceed the sum of all upload rates. As asymmetric internet connections become increasingly popular, this restriction is clearly noticeable in form of low download rates. Together with the observation that popular files are requested by many users within a short time frame, this motivates the usage of multicast communication. This paper analyzes the effects of using IP multicast in a filesharing system.
  • Zeitschriftenartikel
    An OpenCL-based Parallel Coder for Erasure-tolerant Storage
    (PARS: Parallel-Algorithmen, -Rechnerstrukturen und -Systemsoftware: Vol. 28, No. 1, 2011) Hegen, Peter; Sobe, Peter
    Erasure-tolerant coding protects against data loss (erasure) in storage systems by calculating redundant data elements and spreading original and redundant data across several devices. This allows to reconstruct data in case of device failures. Code calculations can be done on the GPU using OpenCL. This allows a highly parallel execution and moves the computational cost away from the CPU. In this paper we introduce a methodology for coding using OpenCL. It consists in a preparation of XOR equations and their compilation to OpenCL kernels.
  • Konferenzbeitrag
    Parallel coding for storage systems - an OpenMP and OpenCL capable framework
    (ARCS 2012 Workshops, 2012) Sobe, Peter
    Parallel storage systems distribute data onto several devices. This allows high access bandwidth that is needed for parallel computing systems. It also improves the storage reliability, provided erasure-tolerant coding is applied and the coding is fast enough. In this paper we assume storage systems that apply data distribution and coding in a combined way. We describe, how coding can be done parallel on multicore and GPU systems in order to keep track with the high storage access bandwidth. A framework is introduced that calculates coding equations from parameters and translates them into OpenMP- and OpenCL-based coding modules. These modules do the encoding for data that is written to the storage system, and do the decoding in case of failures of storage devices. We report on the performance of the coding modules and identify factors that influence the coding performance.
  • Zeitschriftenartikel
    Parallel coding for storage systems — An OpenMP and OpenCL capable framework
    (PARS: Parallel-Algorithmen, -Rechnerstrukturen und -Systemsoftware: Vol. 29, No. 1, 2012) Sobe, Peter
    Parallel storage systems distribute data onto several devices. This allows high access bandwidth that is needed for parallel computing systems. It also improves the storage reliability, provided erasure-tolerant coding is applied and the coding is fast enough. In this paper we assume storage systems that apply data distribution and coding in a combined way. We describe, how coding can be done parallel on multicore and GPU systems in order to keep track with the high storage access bandwidth. A framework is introduced that calculates coding equations from parameters and translates them into OpenMP- and OpenCL-based coding modules. These modules do the encoding for data that is written to the storage system, and do the decoding in case of failures of storage devices. We report on the performance of the coding modules and identify factors that influence the coding performance.
  • Zeitschriftenartikel
    Parallel Processing for Data Deduplication
    (PARS-Mitteilungen: Vol. 32, Nr. 1, 2015) Sobe, Peter; Pazak, Denny; Stiehr, Martin
    Data deduplication is a technique for detection and elimination of duplicated data blocks in storage systems. It creates a set of unique data blocks and places references accordingly, which allows to access the original data within a reduced amount of data blocks. For deduplication, hashes of data blocks are calculated and compared in order to detect and remove duplicates. It can be seen as an alternative to data compression that allows to save storage capacity in large storage systems. A storage capacity saving is reached at the cost of additional computational effort that originates when data blocks are written and updated. This computational effort increases with the size of the storage system. On a single processor system, deduplication influences the performance in a negative way, particularly the write and update rates drop. The utilization of parallelism is a rewarding task to compensate this performance drop, particularly for hash value calculations and comparisons of hashes. In this paper we explain in which parts of a deduplication system it is worth to parallelize and how. Exemplarily, we show the performance results of two deduplication algorithms and their parallel implementations, based on multithreading and on parallel GPU computations.
  • Zeitschriftenartikel
    A Perfomance Study of Parallel Cauchy Reed/Solomon Coding
    (PARS-Mitteilungen: Vol. 31, Nr. 1, 2014) Sobe, Peter; Schumann, Peter
    Cauchy-Reed/Solomon coding is applied to tolerate failures of memories and data storage devices in computer systems. In order to obtain a high data access bandwidth, the calculations for coding must be fast and it is required to utilize parallelism. For a software-based system, the most promising approach is data parallelism which can be easily implemented with OpenMP on a multicore or multiprocessor computer. A beneficial aspect is the clear mathematical nature of coding operations that supports functional parallelism as well. We report on a storage system application that generates the encoder and decoder as C-code automatically from a parametric description of the system and inserts OpenMP directives in the code automatically. We compare the performance in terms of achieved data throughput for data parallelism and for functional parallelism that is generated using OpenMP.