Zynga Case Study

Zynga is one of the world’s most successful mobile game companies, with a collection of hit games—including Words With Friends, Zynga Poker, and FarmVille—that to date have been installed by more than one billion players worldwide. In 2011 the company made a strategic decision to move from AWS to its own private cloud infrastructure. Four years later it returned certain workloads to AWS, taking advantage of the continuous technology innovation that characterizes the AWS cloud so Zynga employees can focus on what they do best: delivering great games.

Early on, Zynga’s games experienced massive growth on the web. FarmVille, which launched in 2009, went from zero to 10 million daily users in a few weeks and exceeded 80 million monthly users just months after its launch. The immense popularity of FarmVille and the anticipated demand for upcoming games created what Zynga Chief Information Officer Dorion Carroll calls “an insatiable demand” for Amazon Elastic Compute Cloud (Amazon EC2) instances.

“In the summer of 2009, it would not have been possible for us to acquire enough hardware for our meteoric growth, let alone provision it or even fit it into our data center footprint. We were running over 10,000 instances and wanted more. By working with AWS at that time, we could accumulate a lot more hardware than if we had purchased, racked, and stacked it on our own. It was a huge value to us.”

Zynga soon determined that its game architecture, which was targeted to desktops and channels like Facebook, the predominant social gaming platform at the time, should run on an internally designed cloud architecture. Zynga added data centers on both the West Coast and East Coast and launched “zCloud” in 2011.

Zynga ran zCloud for several years but, as Carroll notes, the realities of legacy equipment began setting in just as mobile platforms began emerging as a huge new market for the game industry.

“We were maintaining equipment over a typical three-year lifecycle. In our East Coast data center, which had the oldest equipment, we were replacing between 80 and 100 hard drives a month,” he says. “The move to mobile gaming had begun, suggesting a very different way of doing things. Our data center infrastructure was originally built for massive Facebook games that gradually declined as players shifted to mobile. Mobile clients allowed for a richer and smarter experience and we realized that we didn’t need nearly as much equipment infrastructure to serve the emerging mobile game market.”

Zynga learned a lot and became proficient at running its data center infrastructure, but the pace of technology innovation and the new and different demands of mobile gaming brought the company full circle.

“AWS is innovating on technology at a pace that we simply cannot keep up with, while the flexibility and the learnings we achieved from operating our own cloud allowed us to find a path back to AWS,” says Carroll. “It made a lot of sense for us to reengage.”

Zynga’s first big AWS migration project was its analytics platform, which is critical for a player-centric, data-driven company. With assistance from AWS, Zynga conducted proofs of concept (PoCs) to test the analytics platform infrastructure against modern Amazon EC2 instance classes. The result was an analytics management solution that initially reduced a 230-node zCloud cluster to a 115-node cluster running in AWS. Additional PoC work increased efficiency and performance even further by enabling the solution to run on 70-node clusters.

“If we had chosen to go our own route for the migration, I would have picked hardware I was familiar with, including platter drives instead of SSDs,” Carroll says. “By the time we completed the migration of certain services to AWS, our architecture for these services would already have been obsolete. In addition to that, AWS released two new generations of hardware in that same year. So we would have been three generations behind and just starting a three-year capital depreciation on our equipment. Instead, by migrating to AWS we shrunk the footprint for those workloads thanks to hardware advances that AWS made that we didn’t have to.”

Zynga’s move back to AWS began slowly, allowing it to transition at a pace that would support a successful migration without interrupting the player experience or business operations. By using AWS, Zynga could carefully plan, test, and develop PoCs without needing to commit to long-term fixed IT assets. This resulted in reduced costs and lower risk as it adapted to technology demands in real time, instead of relying on forecast models.

Zynga has tapped into a broad range of AWS services to manage processes for IT and the larger business. The company is using Amazon Virtual Private Cloud (Amazon VPC) for individual game team operations, which provides access control as well as billing and management functions for better insight into who is using the AWS pay-as-you-go services. Carroll points to a couple of games as examples of the broad benefits gained by using AWS. For example, it re-implemented the Words With Friends Community Matchmaking feature, reducing the number of database servers by 97 percent, from 100 to 3, on AWS. It also updated that game to Amazon Linux AMI, which helped reduce the number of application server instances by 33 percent.

“With Zynga Poker, we moved a MySQL farm, which required dedicated in-house resources to manage, over to Amazon DynamoDB, which is a fully managed service. It’s resulted in dramatically reduced operational overhead,” he says, adding that AWS Support provided additional assistance with AWS Infrastructure Event Management. “And separately, we’ve gotten a massive performance boost on a Zynga Poker database cluster, with queries that used to take 30 seconds now taking one second. That’s just by taking advantage of the architecture’s modern instance classes--and more importantly, leveraging the continual innovation and investments that AWS makes in systems and the constant discounts it provides.”