Learn the foundations of serverless computing and how to innovate across its deep distributed software-hardware stack.
The tutorial is co-located with ASPLOS 2022 but is planned to be delivered in a hybrid, semi-virtual format.
Room 5B (9:00-18:00) and on Zoom
March 1st, 2022.
Register for in-person or remote participation (mandatory).
Join the tutorial's Slack space (recommended).
Sign up below to receive updates (recommended).
This tutorial provides an overview of serverless cloud computing and introduces the vHive ecosystem, a full-stack open-source framework for serverless experimentation and innovation. The tutorial seeks to educate the community about serverless computing architecture and benchmarking methodology, and teach the researchers from the computer architecture and computer systems community to use vHive for their research. The tutorial includes a number of hands-on sessions on writing serverless applications, analyzing their performance in production and open-source serverless clouds as well as instrumenting and optimizing serverless infrastructure across the whole stack.
vHive is a full-stack open-source framework for serverless experimentation, integrating production technologies.
Arrive early to allow enough time to set up your laptop for the morning hands-on sessions.
- Tutorial goals and contents
- Evolution from cloud to serverless computing
- Key concepts and deep dive into serverless infrastructure
- Introduction to AWS Lambda cloud
- Developing and deploying functions in AWS Lambda
- Developing complex applications and function composition
- Error handling
- vSwarm benchmarking methodology
- vSwarm workloads
- Running vSwarm in AWS Lambda
- Analyzing vSwarm's video analytics workload' response time with AWS X-Ray
Enjoy your lunch meal and get ready for the afternoon session.
Arrive early to allow enough time to set up your laptop for the afternoon hands-on sessions.
- vHive architecture and technologies
- vHive tools for performance analysis
- Guidance for research in computer architecture and distributed systems
- Deploying functions and autoscaling demo
- End-to-end system benchmarking
- Comparing various isolation technologies (containers, Firecracker and gVisor MicroVMs)
- Distributed tracing with OpenTelemetry and Zipkin
Using Intel Top-Down methodology, based on CPU counters
- Generating docker images to be run in the Gem5 simulator
- Analyzing performance numbers obtained from Gem5 simulator
Start a free site with Mobirise