Consultant David Cahill has published a ground-breaking case history of online game developer Zynga, famous for such games as “Farmville”, with a focus on how it handles the huge volumes of valuable data it harvests from the games.
What makes this a particularly important piece, based on a presentation Zynga made at the Interop Enterprise Cloud Summit in Las Vegas, is that Zynga moved a significant part of its IT from the cloud to a new data center specifically to save money. That flies in the face of the common belief that cloud services are always cheaper.
Zynga is a huge operation, and Cahill, who is principal of Diligence Technology Advisors in Boston, provides some statistics to show just how big, including that 10% of the world’s Internet population plays Zynga games every month, and that if Farmville were a country it would be the fifth largest in the world.
Obviously this is a very large, very IT-centric organization, rivaling large financial institutions. When it started, its first games ran on a dedicated IT infrastructure, but it could not build that out fast enough to keep up with the explosive growth it experienced – another of the statistics Cahill cites is that five new players sign up for Zynga games every second.
To absorb that kind of growth it moved to Amazon Web Services (AWS). But over time, it decided to build its own infrastructure and internal cloud to compliment its AWS footprint, in part to move opex to capex, something else that may seem to go against common wisdom. This was a purely financial decision based in part on the benefits of the depreciation writeoffs Zynga can achieve. Zynga manages both its private and AWS clouds as a unit using a single RightScale management system. It can provision more than 1,000 physical servers in 24 hours to keep up with its growth. Today it uses the AWS cloud to absorb the vertical growth that it often experiences in membership to new games while moving older games, with slower growth, to its internal cloud.
Cahill provides many interesting details and analysis in the full article. But the message is that from a financial standpoint, the choice between public and private clouds is not always clear-cut, and total cost of ownership, including the cost of capital and the benefits of depreciation, need to be considered.
Latest posts by Bert Latamore (see all)
- Researchers explain why most IoT action will happen at the edge - September 27, 2016
- How big data and DBMS markets fundamentally differ - September 27, 2016
- Wikibon offers 12-point evaluation framework for big data - September 27, 2016