Track 4: Advanced CI Coordination Ecosystem: Monitoring and Measurement Services

点击翻译按钮获取中文摘要

The SUNY at Buffalo team provides monitoring and measurement services as part of the Advanced Cyberinfrastructure Coordination Ecosystem: Services & Support (ACCESS) program. Advanced cyberinfrastructure (CI), which includes high performance computing (HPC), cloud computing and advanced networks, plays a pivotal role in ensuring U.S. leadership in science and engineering and serves as a powerful tool to help drive the U.S. economy. Indeed, CI performance is essential in all technology fields from research and design to manufacturing. Given the importance of CI, as well as its substantial cost, it is important that all stakeholders, i.e., the public, end-users, system support personnel, application developers and decision makers, have at their disposal tools to help optimize and measure CI performance, as well as the applications that use it.The University at Buffalo is developing CI measurement and optimization tools, namely XDMoD (XD Metrics on Demand) and its open-source counterpart, Open XDMoD, which are in use worldwide for advanced CI monitoring and reporting in academia and industry. This project builds upon and expands this foundation for measuring and monitoring services for the large NSF investment in the Advanced Cyberinfrastructure Coordination Ecosystem: Services & Support program. Highlights of this project include: (1) development of a flexible framework that allows stakeholders to use analytical tools they desire on performance data from national and local CI resources; (2) development of a comprehensive view of the CI ecosystem, including network, data, and public clouds, to facilitate CI planning from the local to the national level; (3) development of a CI simulator to predict the response of the CI ecosystem to proposed new systems or changes in existing systems enabling efficient deployment and use; and (4) development of automated tools to monitor application performance metrics (including energy) on established and novel CI architectures. The relationship between time-to-solution versus energy consumed is important to determine the feasibility of adopting novel CI architectures which promise lower energy consumption.Open XDMoD has a broad impact through its use by institutions worldwide to better manage their CI assets, including academic, industrial, and governmental research centers. Development and deployment of the tools include a range of activities to train the future CI workforce in an inclusive manner promoting diversity and equity.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.