Cloud Control with Distributed Rate Limiting.PDF

Cloud Control with Distributed Rate Limiting Barath Raghavan, Kashi Vishwanath, Sriram Ramabhadran, Kenneth Yocum, and Alex C. Snoeren Department of Computer Science and Engineering University of California, San Diego ABSTRACT script together third-party components to deliver a sophisticated Today’s cloud-based services integrate globally distributed re- service [27, 29]. These specialized services are just the beginning: sources into seamless computing platforms. Provisioning and ac- flagship consumer and enterprise applications are increasingly be- counting for the resource usage of these Internet-scale applications ing delivered in the software-as-a-service model [9]. For example, presents a challenging technical problem. This paper presents the Google Documents, Groove Office, and Windows Live are early ex- design and implementation of distributed rate limiters, which work amples of desktop applications provided in a hosted environment, together to enforce a global rate limit across traffic aggregates at and represent the beginning of a much larger trend. multiple sites, enabling the coordinated policing of a cloud-based Aside from the functionality and management benefits Web- service’s network traffic. Our abstraction not only enforces a global based services afford the end user, hosted platforms present sig- limit, but also ensures that congestion-responsive transport-layer nificant benefits to the service provider as well. Rather than de- flows behave as if they traversed a single, shared limiter. We ploy eac