Serverless computing that runs functions with auto-scaling is a popular task execution pattern in the cloud-native era. By connecting serverless functions into workflows, tenants can achieve complex functionality. Pri...
详细信息
ISBN:
(纸本)9798400703942
Serverless computing that runs functions with auto-scaling is a popular task execution pattern in the cloud-native era. By connecting serverless functions into workflows, tenants can achieve complex functionality. Prior research adopts the control-flowparadigm to orchestrate a serverless workflow. However, the control-flowparadigm inherently results in long response latency, due to the heavy data persistence overhead, sequential resource usage, and late function triggering. Our investigation shows that the data-flow paradigm has the potential to resolve the above problems, with careful design and optimization. We propose dataflower, a scheme that achieves the data-flow paradigm for serverless workflows. In dataflower, a container is abstracted to be a function logic unit and a data logic unit. The function logic unit runs the functions, and the data logic unit handles the data transmission asynchronously. Moreover, a host-container collaborative communication mechanism is used to support efficient data transfer. Our experimental results show that compared to state-of-the-art serverless designs, dataflower reduces the 99%-ile latency of the benchmarks by up to 35.4%, and improves the peak throughput by up to 3.8X.
Serverless computing that runs functions with auto-scaling is a popular task execution pattern in the cloud-native era. By connecting serverless functions into workflows, tenants can achieve complex functionality. Pri...
详细信息
ISBN:
(纸本)9798400703942
Serverless computing that runs functions with auto-scaling is a popular task execution pattern in the cloud-native era. By connecting serverless functions into workflows, tenants can achieve complex functionality. Prior research adopts the control-flowparadigm to orchestrate a serverless workflow. However, the control-flowparadigm inherently results in long response latency, due to the heavy data persistence overhead, sequential resource usage, and late function *** investigation shows that the data-flow paradigm has the potential to resolve the above problems, with careful design and optimization. We propose dataflower, a scheme that achieves the data-flow paradigm for serverless workflows. In dataflower, a container is abstracted to be a function logic unit and a data logic unit. The function logic unit runs the functions, and the data logic unit handles the data transmission asynchronously. Moreover, a host-container collaborative communication mechanism is used to support efficient data transfer. Our experimental results show that compared to state-of-the-art serverless designs, dataflower reduces the 99%-ile latency of the benchmarks by up to 35.4%, and improves the peak throughput by up to 3.8X.
This paper describes the development of an integrated and interactive 'quick' design environment for concept designs utilizing application builder visualization tools for the components of shipbuilding enginee...
详细信息
This paper describes the development of an integrated and interactive 'quick' design environment for concept designs utilizing application builder visualization tools for the components of shipbuilding engineering. The design environment consists of modules inter-linked within a visual network editor. Each module (or filter) operates on a data set using a data-flow paradigm. The primary modules consist of a Read/CAD input module, a surface (geometry) generator and render modules. Large variations in surface geometry for the designs may be found using the solution to certain sets of partial differential equations (PDEs) whilst varying the associated boundary conditions. The interactive nature of the environment is highlighted by the fact that changes in the boundary conditions of the designs for the Read/CAD module are immediately reflected by changes in surface geometry modeling the design through the module links. Here the designer may, in a single session, investigate a broad range of designs each one providing important geometry information. The representation of a simple hull, a propeller blade geometry and impeller blade geometry for an axial-flow pump is demonstrated. It should be noted that standard techniques of surface representation, such as B-spline interpolation, require a large number of control points to achieve this. The PDEs method approaches the representations of the designs as a boundary-valued problem which ensures that a fair surface is generated and, secondly, that a small set of design parameters is needed. The small parameter set is of importance, since we can first manipulate the design with ease, and second, use it to great advantage in the task of functional design.
The increasing demand in HPC to utilize accelerators has motivated the development of pragma-based directives to target these devices. OmpSs-2 and OpenACC are both directive-based solutions that allow application prog...
详细信息
ISBN:
(纸本)9781450392044
The increasing demand in HPC to utilize accelerators has motivated the development of pragma-based directives to target these devices. OmpSs-2 and OpenACC are both directive-based solutions that allow application programmers to utilize accelerators. The two leverage distinct types of parallelism: task parallelism and data parallelism, respectively. Non-trivial scientific applications can benefit from both types of available parallelism. However, the combination of pragma-based models is difficult to coordinate, as both assume full control and are unaware of each other at runtime. We propose an interoperation mechanism to enable novel composability across pragma-based programming models. We study and propose a clear separation of duties and implement our approach by augmenting the OmpSs-2 programming model, compiler and runtime to support OmpSs-2 + OpenACC programming.
暂无评论