With the Google IO Developer Conference completed there are dozens and dozens of information packed videos now available. While you won't get any of the nifty free swag the attendees rake in, it's great of Google to make these videos available so quickly after the conference.
Let's say you don't want to watch all the videos on the pretense you have a life, here are just a dozen scalability and architecture related videos you might find interesting:
- App Engine Backends by Justin Haugh, Greg Darke. One of this biggest complaints about GAE are the request deadlines. Those are now gone when you rent a new Backend node.
- Scaling App Engine Applications by Justin Haugh, Guido van Rossum. Two parts: how App Engine scales internally and how you as a programmer make it scale using the tools. Good discussion of scaling and why it's hard; 10,000 queries per second is a large app that needs good architecting; a good discussion of GAE's predictive scaling formula of how it decides when to spawn instances; throughput is determined by CPU usage; understand your apps performance; using load testing; strategies for scaling; common pitfalls. The less you do the faster it will be and the less it wil cost.
- App Engine MapReduce by Mike Aizatskyi. Previously you could only write Map jobs, now you can run full Map Reduce jobs on App Engine.
- Coding For The Cloud: How We Write Enterprise Apps for Google on App Engine by Ben Fried, Justin McWilliams, Eric Schoeffler, Justin Fagnani. App Engine is just a backend for Facebook games, Google want to remind you that they do the Enterprise too: performance reviews, help desk, course scheduling, expense reporting, payroll, etc.
- Fireside Chat with the App Engine Team with Max Ross, Alon Levi, Sean Lynch, Greg Dalesandre, Guido van Rossum, Brett Slatkin, Peter Magnusson, Mickey Kataria, Peter McKenzie. Get the marshmallows, the fire is hot. The team talks about the pricing changes, HR datastore, and lots of great questions that lead to more great things to work on.
- Full Text Search by Bo Majewski, Ged Ellis. A quirky fact about a Google App Engine is that it did not support search. Now it will.
- More 9s Please: Under The Covers of the High Replication Datastore by Alfred Fuller, Matt Wilder. GAE is replacing their original Master/Slave Datastore with the High Replication Datastore. Why such a big step? Higher availability. The cost? Latency and consistency. Excellent discussion of the different tradeoffs and why they were made.
- Putting Task Queues to Work by Nicholas Verne, Vivek Sahasranaman. Always a great feature, task queues were restricted by being pushed based, now pull based queues make it possible to process tasks in a VM by pulling from queues using a REST API.
- Large-scale Data Analysis Using the App Engine Pipeline API by Brett Slatkin. Think Big Data and you think Google. Here's how you do Big Data on GAE: build multi-phase Map Reduce workflows; how to merge multiple large data sources with "join" operations; and how to build reusable analysis components.
- Using The Google Docs APIs To Store All Your Information In The Cloud by Vic Fryzel. Not just another file store: Storage is per user; users control storage quota; Data is inherently structured; All entries have the same metadata; Documents currently use zero quota.
- Highly Productive GWT: Rapid Development with App Engine, Objectify, RequestFactory, and gwt-platform by David Chandler, Philippe Beaudoin, Jeff Schnitzer. GAE+GWT+objectify+ other tools is an awesomely productive tool chain. Nice examples with code on how to make it work.
These talks are generally of high quality and provide insight you won't get elsewhere. Highly recommended to take a look around.