Introducing Falcore and Timber

BY DAVE GRIJALVA and SCOTT WHITE

We're happy to announce that we've open-sourced some of the reusable tools we've built to ramp up our Go infrastructure.

Embedding V8, Part One: Building Your Application

BY CHRIS JIMISON

Welcome to our tutorial on embedding V8; this is the first post in a series.

Agile Distilled

BY ERIC GRUNDSTROM

There are many great ideas in agile methodologies. Often, though, it feels like you have to drink the Kool-Aid, not really knowing what parts of the process are critical and which are supplemental.

Testing with Apache Bench

BY SCOTT WHITE

We've recently been writing a bunch of new server software to scale up our Mobage service. We already have a high traffic service that works well, but we've basically hit a wall with the current architecture.

Meet Eli Delventhal

BY STEPHANIE LIN

(Like what you've seen? Come work with us, and change the world.)

Hammer that nail!

BY ERIC GRUNDSTROM

You may have heard the expression, “If all you have is a hammer, everything looks like a nail.”

A Tree-based Approach to Persistence with MongoDB

BY HERNAN SILBERMAN

Persistence is always the biggest challenge for our large game deployments. Obviously, we want persistence that's fast, reliable, scalable and a pleasure to use. Historically, we've handled persistence with MySQL and we’ve done all sorts of things with it: we've sliced our data sets vertically and horizontally, we've learned how to gracefully do schema migrations, and we’ve learned how to make our MySQL deployments as highly available as possible.

How do you cache your data?

BY ASHLEY MARTENS

We all know that caching is the secret scaling sauce. Your database caches recent data and queries. You use memcache to store recent page fragments or rendered data. You use in-process memory to cache connections and other data. You might even use a GEM or some other logic to cache ActiveRecord models in memcache/redis/membase/tokyo/ etc. Whatever caching strategy you're using, you probably couldn't keep your service up without it.

For the past two years I've maintained a fork of the cache-money GEM and I'd like to share some thoughts about the GEM, its design, some of the issues and what the future might hold.

Connection Pooling Sucks

BY DAVE GRIJALVA

Connection pooling sucks and we shouldn't be doing it anymore.

Modern server stacks are handling multiple requests concurrently at every tier of the infrastructure. Connection pooling is an artifact of older times. It introduces many more problems than it solves. Any developer or operations expert who's spent hours, days, or weeks tuning their connections pools is likely to agree with me.