Topic 3: Scheduling and Load Balancing

Parallel and distributed systems available today are becoming more complex and powerful, however, they are still not fully exploited. Scheduling and load balancing issues are crucial for a more efficient and transparent use. Related techniques provided at both application and system levels are of interest for this topic. At the application level, the mapping of applications onto the underlying computing platforms, and the development of dynamic algorithms that are able to adapt to the particular characteristics and the actual utilization of the systems are of particular relevance. At the system level, areas of interest include the support of modern multi-core and many-core architectures, huge data centers, and virtual systems like cloud infrastructures.

This topic covers all aspects related to scheduling and load balancing from theoretical foundations for modeling and designing efficient and robust strategies to experimental studies, applications and practical tools. This applies to multi-core processors, servers, heterogeneous systems, HPC clusters as well as distributed systems such as computational grids, clouds and global computing platforms.

• Scheduling algorithms for homogeneous or heterogeneous platforms
• Theoretical foundations of scheduling algorithms
• Robustness of scheduling algorithms
• Multi-objective scheduling
• Decentralized or hierarchical scheduling
• Scheduling at extreme scale
• On-line scheduling
• Resource management and awareness
• Evaluation and analysis of load balancing and scheduling techniques
• Implementation issues of scheduling
• Workload characterization and modeling
• Workflow and job scheduling
• Performance models for scheduling and load balancing
• Power-aware and thermal-aware methods in scheduling and load balancing
• Energy efficient scheduling
• Concurrent workflow scheduling