An Extended Data-Flow Architecture for Data Analysis and Visualization
VIS '95 Proceedings of the 6th conference on Visualization '95
Visualization in Grid Computing Environments
VIS '04 Proceedings of the conference on Visualization '04
Efficient scheduling and execution of scientific workflow tasks
SSDBM'2005 Proceedings of the 17th international conference on Scientific and statistical database management
Scientific workflow management and the Kepler system: Research Articles
Concurrency and Computation: Practice & Experience - Workflow in Grid Systems
Pegasus: A framework for mapping complex scientific workflows onto distributed systems
Scientific Programming
Self-Adaptive Configuration of Visualization Pipeline Over Wide-Area Networks
IEEE Transactions on Computers
Trident: Scientific Workflow Workbench for Oceanography
SERVICES '08 Proceedings of the 2008 IEEE Congress on Services - Part I
Towards Quality of Service in Scientific Workflows by Using Advance Resource Reservations
SERVICES '09 Proceedings of the 2009 Congress on Services - I
CloudVista: visual cluster exploration for extreme scale data in the cloud
SSDBM'11 Proceedings of the 23rd international conference on Scientific and statistical database management
Hi-index | 0.00 |
Science is becoming data-intensive, requiring new software architectures that can exploit resources at all scales: local GPUs for interactive visualization, server-side multi-core machines with fast processors and large memories, and scalable, pay-as-you-go cloud resources. Architectures that seamlessly and flexibly exploit all three platforms are largely unexplored. Informed by a long-term collaboration with ocean scientists, we articulate a suite of representative visual data analytics workflows and use them to design and implement a multi-tier immersive visualization system. We then analyze a variety of candidate architectures spanning all three platforms, articulate their tradeoffs and requirements, and evaluate their performance. We conclude that although "pushing the computation to the data" is generally the optimal strategy, no one single architecture is optimal in all cases and client-side processing cannot be made obsolete by cloud computing. Rather, rich visual data analytics applications benefit from access to a variety of cross-scale, seamless "client + cloud" architectures.