Managing Thousands of Cloud Instances with Java

Java: Cloud
Location: A107/108
Average rating: *****
(5.00, 2 ratings)

Using real-world examples, learn from the founder and chief engineer of a modern cloud-based platform about what works well in the cloud, what doesn’t work so well, and what will almost certainly surprise you as you scale.

Topics covered will include:

  • The myth of infinite scalability in the cloud, and how to deal with it
  • Designing your Java platform to auto-scale with ease
  • Tips for aligning architecture with cloud pricing models
  • Effectively spinning up 1000’s of instances in minutes – and tracking them!
  • Overview of different cloud instance types (CPU and memory tradeoffs, etc)
  • Techniques for customizing VMs as they launch in the cloud (EC2 User-Data tricks, etc)
  • How to bid on EC2 spot instances – and when it’s not a good idea
  • Example uses of Amazon’s Elastic Block Storage

Throughout this session the audience will see real-world code examples covering things like launching Rackspace and EC2 instances (using popular open source packages such as Typica, for example) and algorithms on how to deal with the nuances of the cloud (such as detecting hung instances and gracefully recovering).

Photo of Patrick Lightbody

Patrick Lightbody

New Relic

Patrick runs product management for New Relic, the leader in software analytics. Prior to that, Patrick started two startups in the website monitoring and load testing space, which were acquired by Gomez and Neustar, respectively.

Patrick is an avid open source contributor, having founded OpenQA, created Selenium Remote Control, and co-created Struts 2.

Patrick has held management and software engineering positions with Jive Software, Spoke Software, and Cisco Systems.