Brown CS News

Rodrigo Fonseca Wins The USENIX Community Award For Characterizing And Optimizing A Large Serverless Workload


Click the links that follow for more news about Rodrigo Fonseca and other recent accomplishments by our faculty.

Held virtually this year, the USENIX Annual Technical Conference (USENIX ATC) brings together leading systems researchers for the presentation of cutting-edge systems research and the opportunity to gain insight into a wealth of must-know topics, including virtualization, system and network management and troubleshooting, cloud and edge computing, security, privacy, and trust, mobile and wireless, and more.

Brown CS Professor Rodrigo Fonseca was an attendee, and he and colleagues from Microsoft Azure and Microsoft Research (Mohammad Shahrad, Íñigo Goiri, Gohar Chaudhry, Paul Batum, Jason Cooke, Eduardo Laureano, Colby Tresness, Mark Russinovich, and Ricardo Bianchini), have won their Community Award, which is given for providing an important new dataset to the community and for the paper ("Serverless in the Wild: Characterizing and Optimizing the Serverless Workload at a Large Cloud Provider") analyzing it. In presenting the award, the organizers remarked that "when such a release is the first of its kind and on a timely topic, as is the case in this paper, it marks an important milestone for the community".

"Function as a Service (FaaS)," the researchers explain, "has been gaining popularity as a way to deploy computations to serverless backends in the cloud. This paradigm shifts the complexity of allocating and provisioning resources to the cloud provider, which has to provide the illusion of always-available resources (i.e. fast function invocations without cold starts) at the lowest possible resource cost. Doing so requires the provider to deeply understand the characteristics of the FaaS workload. Unfortunately, there has been little to no public information on these characteristics."

In their paper, Rodrigo and his colleagues first characterize the entire production FaaS workload of Azure Functions. They show for example that most functions are invoked very infrequently, but there is an 8-order-of-magnitude range of invocation frequencies. Using observations from their characterization, they then propose a practical resource management policy that significantly reduces the number of function cold starts, while spending fewer resources than state-of-the-practice policies.

The paper and a presentation video are available here.

For more information, click the link that follows to contact Brown CS Communication Outreach Specialist Jesse C. Polhemus.