Scheduling and Adaptive Execution of Data Analysis Workflows across Heterogeneous Infra-structures (B01)

复杂DAW的可伸缩实现需要对深层技术堆栈有深刻的了解，这些技术堆栈包括抽象的DAW描述、逻辑计划的编译、映射到当前可用的基础设施以及执行引擎的适当配置。为一个系统开发的组件通常不适合另一个系统，从而导致不受欢迎的平台锁定。B1的目标是通过将daw的需求映射到可用基础设施的声明性描述，并结合利用这些描述的新颖调度算法，来增强daw的可移植性和适应性。B1是其他几个FONDA子项目的重要合作伙伴，因为它开发的声明性硬件描述是实现可移植性的关键要求之一。B1将由大型分布式系统管理和编程专家Kao教授和具有丰富科学应用经验的高效调度算法和负载平衡专家Meyerhenke教授领导。

The scalable implementation of complex DAWs requires profound knowledge of a deep technology stack consisting of an abstract DAW description, compilation of logical plans, mapping onto the cur-rently available infrastructure, and appropriate configuration of execution engines. Components de-veloped for one system are often unsuitable for another, leading to an undesirable platform lock-in. The goal of B1 is to enhance portability and adaptation of DAWs by mapping their requirements to a declarative description of the available infrastructure combined with novel scheduling algorithms that exploit these descriptions. B1 is an important partner for several other FONDA subprojects, as the declarative hardware descriptions it develops are one of the key requirements for achieving portability. B1 will be led by Prof. Kao, an expert in the management and programming of large distributed sys-tems, and Prof. Meyerhenke, an expert in efficient scheduling algorithms and load balancing with ample prior experience in scientific applications.