Call for papers
The continuous demand for high processing power at reasonable cost led to
the evolution of cluster computing putting more demand on the I/O subsystem,
and specifically storage. A key challenge is how to transfer large amounts of
data in and out of large-scale systems, like clusters, Grids and peer-to-peer
systems. The aim of the session is to identify, discuss, and share the barriers
and workarounds that have been discovered.
Particular areas of interest include, but are not limited to:
- Storage Technology and protocols including both hardware and software elements.
- Magnetic storage, network storage, optical storage, fibre channel, iSCSI, SAN, NASD, Storage security, disaster recovery, Environmental (power, cooling, structural, etc) concerns/issues, etc.
- Storage issues in Parallel and distributed systems.
- Massively parallel storage architectures, data and meta-data consistency, allocation and utilization strategies for exploiting parallel and distributed memory, cluster file systems, Ad-hoc parallel file systems, parallel I/O architectures, distributed file systems, etc.
- Storage issues in Grid computing environments.
- I/O interfaces and formats, new topics on I/O architectures for grids, I/O middleware in grids, parallel and Wide-Area File Systems, global file systems, data replication, reliability, I/O performance, I/O and scheduling, etc.
- Storage issues in peer-to-peer storage environments.
- Unstructured routing and file sharing, distributed file systems, searching and querying, resource location, replication strategies, multimedia streaming, XML stores, etc.
- Storage issues associated with large-scale distributed applications.
- Content management and data mining in large storage systems and networks, new trends in very large clusters file systems, system architectures deployed in very large scale systems, Tools to manage large systems, archive strategies, network connectivity within the data center and with external sites.
- Parallel and distributed databases.
- OLAP, replication schemes, optimal data allocation, multidimensional analysis, access patterns, data declustering, Client Caching Data Management Systems, index distribution schemes, etc.
- Performance evaluation: benchmarking of data management systems.
- Benchmarking of data management systems, I/O benchmarks, performance under faulty conditions, performance tuning, tests of products, I/O optimization techniques for scientific and business applications, etc.
- Reliability and availability of storage systems.
- Fault tolerance mechanisms, impact of replication on reliability, reliability and availability modelling, highly available storage systems, ...
- Data management applications.
- Data mining, bioinformatics, simulations, geographical information systems, image processing, Web retrieval, checkpointing, astrophysics, particle physics, etc.
Important dates
- Papers due: March, 1, 2007.
- Notification to authors: March, 20, 2007.
- Camera ready papers: April, 20, 2007.
- Workshop dates: June, 25-28, 2007.
Academic co-sponsors
Academic co-sponsors of PDPTA include:
- Massachusetts Institute of Technology (MIT) Media Laboratory, MIT (Cambridge, Massachusetts).
- Statistical Genomics and Computational Biology Laboratory, Department of Statistics, Harvard University (Cambridge, Massachusetts).
- Texas Advanced Computing Center, The University of Texas at Austin (Austin, Texas).
- Statistical and Computational Intelligence Laboratory of Purdue University (West Lafayette, Indiana).
- University of Iowa's Medical Imaging HPC Lab (Iowa City, Iowa).
- Institute for Informatics Problems of the Russian Academy of Sciences, Moscow, Russia)
|
