Measurement Science for Cloud Computing - overview
Cloud computing has provided the ability to command nearly unlimited computing power and storage space, with customers able to scale their usage up or down dynamically in a matter of minutes. This is cost-effective for customers, but it presents extremely difficult system management challenges for cloud service providers who must keep track of virtualized resources and the state of the physical infrastructure under extreme dynamics and scale. This situation will become worse as cloud computing becomes more successful, potentially inhibiting its progress and limiting its advantages to society. Addressing this necessitates developing mechanisms to effectively measure the state of cloud resources and manage them. This project will advance the state-of-the-art in cloud measurement science by providing an accurate understanding of the underlying cloud systems behavior by analyzing measurement data, by improving existing service hosting assignment and workflow scheduling techniques, and by proposing extensions that will enable existing policy standards and systems to support cloud measurement data.
Funded by National Institute of Standards and Technology (NIST).
Contact: Kang-Won Lee and Ting He