Serverless Meetup: Sandboxing With V8 Isolates, Serverless Video Sharing, and Serverless Transforms

Date
March 21, 2019
Location
Austin, US
Details

Join us for three serverless talks in at Cloudflare’s new office. This event is in partnership with ATX Serverless Meetup.

Talk 1: Fine-Grained Sandboxing With V8 Isolates - Kenton Varda

In an ideal world, code would always run as close as possible to the things with which it communicates – end users, storage, external APIs, etc. However, most software stacks are only cost-effective if applications are long-running, handling many events in one place. We need a new approach that can efficiently load code to handle just one event, so that it can be executed wherever is best for that one event, across a distributed network.

Kenton explains how Cloudflare solved this problem by building a compute platform using V8 isolates instead of containers or VMs, achieving 10x-100x faster cold starts and lower memory footprints. We’ll go through technical details of embedding V8, distributing code, scheduling isolates, resource management, and security risks.

Speaker Bio:

Kenton Varda is the tech lead and architect for the Cloudflare Workers project, which allows Cloudflare customers to run JavaScript code directly on Cloudflare’s edge network. Before he joined Cloudflare, Kenton served as co-founder and author of Sandstorm.io and created Cap’n Proto, both open source projects. Before that, Kenton spent eight years at Google, where he was responsible for open sourcing Protocol Buffers.

Talk (Workshop) 2: From CSV to DynamoDB - Serverless Transforms That Don’t Break The Bank - Julia Jacobs

A common use case for mobile / e-commerce apps is consuming delimited product data from an external source. Some of the AWS ETL tools, like Data Migration Service or Data Pipeline use other services (Elastic Map Reduce) that can be costly. Importing delimited data into DynamoDB, also currently being used in mobile apps, is not trivial.

To address some of these issues, this workshop will show how to create a data pipeline that will:

Run a nightly scheduled lambda to pull a headerless pipe delimited file from an external FTP drive to an S3 bucket Run lambdas that execute Athena queries to clean up data Run a lambda to import csv to DynamoDB

Julia Jacobs is an AWS Certified Associate Architect with over 20 years of experience working as both a business owner with Fortune 500 clients and Full Stack Consultant engagements. Her latest work has involved using serverless infrastructure in ETL solutions.

Talk 3: Building a Serverless Video Sharing Site - Jesse Kipp

There is an incredible variety of Cloud-computing and SAAS resources available in the market. Until recently, developers had to configure and deploy a web server to integrate them together in novel ways. Serverless computing provides a way to combine hosted services and APIs without the need to provision and run an origin server. Jesse will show how to build a serverless video sharing site by combining several API services within serverless compute platforms.

Speaker Bio:

Jesse Kipp is a Systems Engineer at Cloudflare. He helped build Cloudflare Stream and now works on using machine learning models to identify and block attacks on customers’ websites.

Agenda:

6:00pm - Doors open 6:30pm - Talk 1: Fine-Grained Sandboxing With V8 Isolates - Kenton Varda (Cloudflare) 7:10pm - Talk (Workshop) 2: From CSV to DynamoDB - Serverless Transforms That Don’t Break The Bank 8:00pm - Talk 3: Building a Serverless Video Sharing Site - Jesse Kipp (Cloudflare) 8:30pm - Networking over food and drinks