Game On!

Sun Labs' "Project Darkstar" Helps Game Developers Take Advantage of the Burgeoning Market for Massively Scalable, Multi-player Online Games

March 20, 2006 - The explosive growth of online video games means somebody out there is having fun. The $28 billion video game market is now the second largest media market in the world-beating out music, books, and every other form of entertainment except films.

The people who play the games are having a good time. The people who market and sell the games are enjoying life. The shareholders of video game makers are living large. But for developers, writing online games that must scale to dozens, hundreds, or thousands of players—and that must accommodate a diverse assortment of net-connected client devices—is anything but fun and games. It is an expensive, time-consuming, risky, and frequently frustrating venture.

Sun Labs aims to change that. Project Darkstar, released publicly at the 2006 Game Developer Conference in San Jose, California, is harnessing Sun's expertise in development tools, Java technology, and massively scalable back-end infrastructure to simplify the process of developing games that can be deployed on a massive scale to players using virtually any client device. A freely downloadable, early access SDK for writing client-side and server-side code is now available from Sun Labs at http://www.projectdarkstar.com.

This article takes a closer look at the problems Project Darkstar is helping developers overcome—and the new business opportunities the Sun Labs technology can open up for game development companies of all types and sizes.

Online Game Development Today: Risky Business

The inherent complexity of multi-player games can make them prohibitively risky and expensive for most game developers today. In her highly respected book, "Developing Online Games: Insiders Guide," Jessica Mulligan estimates that taking a single-player game project to massively multi-player triples the minimum development costs from $10 million to $30 million. These costs are driven by four key factors:

• Server development costs: Today the vast majority of massively multi-player games employ game servers written specifically for their game. So writing the game server is essentially a second development project on top of the base project of developing the game client software.

• Cost of building a server room: Since the game software is custom, the game developer typically must build a custom server room in which to run it. And since CPU workloads are extremely hard to predict, excessive CPU resources are often purchased and then underutilized, or insufficient resources are purchased and the game can't scale.

• Server administration costs: When you're running custom software in a custom server room, you need to train and deploy your own server management team, further driving up costs.

• Customer service and support: Custom software means additional support issues. Moreover, quality-of-service issues inherent in the architectures in use today create a heavy support load. Beyond the cost issues, there are additional risks to be addressed. These include (among others):

• Risk of outright failure to launch: Of more than 100 games announced two years ago, fewer than 10% have actually come to market. Many game projects fail in the server development phase.

• Technical and market risk: Some games that do launch are market failures due to technical problems. They may not scale to large numbers of users; they may be frequently unavailable due to lack of fault tolerance in the underlying infrastructure; or they may be plagued by excessive customer service issues.

• Subscriber loss: Gamers are notoriously impatient. Sluggish game play, buggy servers, or poor quality of service leads to boredom or frustration, which translates—often instantly—into a dissatisfied customer and a permanently lost subscriber.

Project Darkstar: "We're Doing a Lot of the Hard Work for You"

Led by video game industry experts Chris Melissinos and Jeffrey Kesselman under the direction of Karl Haberl, Director, Sun Labs is developing technologies that attack the cost and risk of online game development on several fronts.

"Sun is not a video game company and we won't be producing our own video games," said Chris Melissinos, "but we can give game developers access to powerful technologies that will do a lot of the heavy lifting for them. There's no need to build a game server from scratch and then figure out how you're going to scale it for massive multi-player scenarios. We're doing a lot of the hard work for you so you can focus on your game, not the infrastructure."

Examples of the technologies the Project Darkstar team has produced or is working on include:

Sun Game Server (SGS): Introduced as a prototype at the 2005 Game Developers Conference, the SGS is the industry's only platform that can scale vertically and horizontally to enable companies to host their games and add server resources incrementally. The unique architecture of the SGS enables several different games to coexist on the same infrastructure, thereby lowering the total cost of ownership (TCO) for any game company deploying online games. The SGS lowers development risks and costs in the following ways:

• It makes it easy to write massively multi-player game servers. The game programmer just writes event handling code; the system does all the rest. Game server coding takes months rather than years.
  
  
• It reduces investment risk on the server side to zero. There is no longer anything to buy upfront. The SDK is free. For serious developers, the Project Darkstar team is currently investigating an optional pre-installed turn-key development system that could be deployed on an existing office network and that includes Sun support, all at a low price. Stay tuned for details.
  
  
• It ensures massive scalability of the game. Scaling problems have doomed otherwise promising games. The SGS has been designed, from the bottom up, to be more scalable beyond other game server solutions currently in the market.
  
  
• It makes the entire game persistent. There is no database code to worry about; players can make permanent changes as they play. Every object in the system is innately persistent.
  
  
• The game never goes down. In server parlance this is "5-nines" reliability or "fault-tolerance." Even if a server completely fails, that failure is hidden from the user and will not interrupt game play. You can add or remove servers, patch server code, or even roll-back patches all without interrupting players.
  
  
• It is client agnostic. Sun provides APIs for C++, Java2 SE, and Java2 ME (cell-phone Java.) The connection API itself is developer extensible, making it possible to add unique clients and connection strategies.

A freely downloadable, early access SDK: This development kit gives game developers what they need to write all their game code for both client side and server side. It allows them to test the game on their own LAN or WAN connection with between 20 and 50 players (the exact number depends on what they are doing in their server-side code and how big a machine they install the server on).

The SDK consists of the following components:

Battle Trolls

• Sun Game Server: Developer Edition: A single-machine version of the Sun Game Server. It is like the full version in every way except that it will not scale across multiple servers.
  
  
• Documentation: Complete and thorough docs on how to code for the SGS.
  
  
• Examples: Three games are included in the SDK. Battletrolls is a simple Massively Multiplayer On-line Roleplay Game (MMORPG). Stomping Grounds is a demo version of a commercial multiplayer casual game from Mind Control Software. The Cosmic Birdie Racing Combat game is a demo version of a commercial fast-action game from Immediate Mode Interactive.

Full source to both the Battletrolls client and server is included with the SDK. Additionally, the networking source code from Stomping Grounds is also included.

Stomping Grounds

Changing the Way the Game is Played

Project Darkstar is about more than making sophisticated online video games more easily available to more players. It is also about changing the rules of current online video game business models.

The video game industry, like most media, is a hits-driven marketplace. The fickle nature of consumer preferences increases the risk for all industry players, but it often makes it prohibitively expensive for smaller game developers to give it a go, because one miss could spell financial disaster.

Project Darkstar can do a great deal to level the playing field among game developers of all sizes. This technology can be instrumental in helping great games come to market—games that would otherwise have been shut out or co-opted by larger, stronger enterprises. The net result is more competition, better products, and lower prices, the three cornerstones of Sun's business model for almost 25 years.

Cosmic Birdie

In addition, the technologies being developed under Project Darkstar will give game developers a new level of flexibility in how they manage resources, and new options for generating revenue. For example:

• The pervasive use of Java technology can help developers create games that are truly client-agnostic, so they can broaden the market for their games.

• Sun's "utility computing" model enables game developers to avoid purchasing and maintaining special-purpose hardware. Instead, they just pay for compute resources on demand as and when needed. Similarly, game developers can take advantage of Sun's "developer utility," which provides simple, on-demand access to commonly used tool sets, such as Rapid Application Development (RAD) tools, the Java Enterprise System stack, testing resources, and load management tools, accelerating the develop/edit/ test/debug cycle and speeding time-to-market for new games and content delivery services.

• Game developers will have a full spectrum of choices in how they license and derive revenue from their games. For example, they could elect to develop and deploy their games under a variety of different models: from standing up their own NOC to completely handing their title off to a fully functioning GNOC.

"What's exciting about this project is that it could have huge impact on everyone involved in online gaming—game players, game developers, Sun, and our partners," said Karl Haberl, project lead. "And in this game, everyone comes out ahead."