Geek Gestalt

SAN FRANCISCO--It will likely come as no surprise to anyone familiar with virtual worlds and online games that they can be hacked. But what might come as a shock is the sheer breadth of types of exploits that are possible.

That was the broad message of a Thursday panel called, appropriately, "Exploiting Online Games" at the RSA 2009 security conference here.

Moderated by Gary McGraw, CTO of software security consulting firm Cigital and an author of several books, the panel took the audience on a deep dive into the diverse ways that hackers and others have figured out to either skim real money or to gain game play advantages not available to normal players.

McGraw opened the panel with a brief explanation of the fact that there are real, functioning economies in virtual worlds and online games, and that players cash in their virtual goods for real money, to the tune of more than $1 billion a year. This, of course, is old news to those in game playing circles, but for many of the security experts in the room, it may well have been eye-opening.

And, McGraw said, it's the very fact that real money is at stake that often gets otherwise uninterested game players to pay attention to the security risks they face every day.

"There's a whole bunch of normals (those not steeped in knowledge about computers) using games, and they don't care about security," McGraw said. "But they like their stuff, (and) when their stuff gets taken, that really hurts the hell out of them. That's a way to start a conversation about computer security with normals, because almost everybody knows somebody who plays online games."

The first panelist to present was Greg Hoglund, the founder of Rootkit.com and the CEO of the consulting firm, HBGary. He explained that online games are regularly under attack by two discrete types of cheats: exploits--actual bugs in games that clever hackers have figured out how to mine in various ways, and bots, which are essentially automated macros that can be used to perform mundane tasks again and again and again, and very profitably.

The bugs, Hoglund said, often exist "at the borders of systems," and are used for things such as duplicating gold, or leveraging poor synchronization between back-end databases to extract money out of a game economy or even to gain teleportation powers that otherwise don't exist.

Hoglund also recalled a security expert who figured out a hack that allowed him not only to filch Second Life users' virtual currency--which is directly convertible to US dollars--but also to get ahold of users' credit card information and then use it to buy more of the currency to trade in. That exploit, Hoglund explained, was done only to prove that it could be done, but it underlined some of the significant risks facing players of online games and virtual worlds with functioning economies, as well as the publishers of those titles.

He also talked about bots, and explained that they, too, are often employed to gain an advantage most players don't have. They are almost universally prohibited, but Hoglund said creating them and using them is remarkably easy for those who know what they're doing. And he talked about one he had written to use in World of Warcraft that allowed his character to stay safe from attack from the rear, while also luring in loot-bearing enemies to kill. Once killed, the enemies would be regenerated by the bot, allowing Hoglund's character to kill them and pick off all their loot over and over again, a process that netted him significant profit, he hinted.

Similarly, he explained that games like World of Warcraft have vulnerabilities that allow savvy hackers to tap into the games' code, allowing for all kinds of new abilities, like being able to perform 15 charms at once, not available to the public at large.

Hoglund said companies like WoW publisher Blizzard are always actively trying to stop players from employing bots and ban those they catch, but added that for those who know what they're doing, detection is not something to worry about. And that, of course, is one of the explanations behind the so-called gold "farmers," often teams working in third-world countries whose job it is to run multiple accounts simultaneously, usually employing bots to perform gold-earning tasks and essentially just making sure that their in-game characters don't get "lodged in a tree."

Courts weigh in
Next up was Sean Kane, a partner with the New York law firm of Drakeford & Kane, and a leading voice on issues surrounding the law and virtual worlds.

Kane talked about two specific cases, one that is several years old and one that is much more recent.

The older case, Bragg v. Linden Research, focused on whether Linden, the publisher of the virtual world Second Life, was right to shut down the account of a user who had discovered an exploit allowing him to buy virtual land at below-market prices. Mark Bragg, the plaintiff, demanded $8,000 in restitution and eventually won a settlement from Linden in which his account was reinstated. But that only happened, Kane pointed out, after a federal judge ruled that the arbitration clause in the Second Life terms of service was onerous and one-sided.

At the time, the entire virtual world community had been watching the case closely, as many thought it would be the case that for the first time established the real-world value of virtual goods (and despite the fact that Bragg, himself a lawyer, had filed his suit in state court with a hand-written form), However, the settlement, not long after the federal judge's ruling, side-stepped that outcome.

But what many found interesting at the time was that Bragg had argued his hack was fair game, since all he did was exploit a feature hidden in the Second Life code. In effect, Bragg argued, code is law, and anything that players can do with the tools at their disposal is legitimate. Linden obviously disagreed, but ended up settling anyway.

Kane also focused on another case, MDY Industries v. Blizzard, in which MDY had created a bot, called Glider, that allowed players to level-up their characters without even having to be playing.

Blizzard sued for copyright infringement, arguing that bots like Glider were prohibited under its end-user license agreement (EULA) and that only that license actually allowed players to run WoW. In essence, the argument said that by running WoW under circumstances that violated the EULA, Glider was supporting copyright infringement.

Ultimately, though many argued that Blizzard's argument was beyond specious, the courts ruled in favor of the publisher, awarding it $6 million. But, not surprisingly, the outcome is on appeal.

Hacking Disney
Aaron Portnoy, a researcher with Tippingpoint security research, took the microphone next and talked briefly about his experiences hacking the Python code of the Disney online game, Pirates of the Caribbean. He explained that because Python is a dynamic language, he and a colleague had needed just a couple of days to reverse-engineer all of the game's code, and were able to use their exploit to get their in-game characters to do things that were otherwise impossible.

During a panel on exploiting online games, Tippingpoint's Aaron Portnoy talked about how he and a colleague discovered that Disney's online game Pirates of the Caribbean was written in Python, a language that allowed them to reverse-engineer the game's code in just two days. The result was that Portnoy's character was able to fly high in the sky, whereas everyone else in the game was limited to jumps of just four feet high.

(Credit: Daniel Terdiman/CNET Networks)

For example, Portnoy said, he was able to easily get his character to jump high in the air, while the standard maximum jump was just about four feet. Or, to jump out of a pirate ship, walk on water at a speed faster than sailing ships in the game could travel, and attack at will.

"Everybody could see my guy jumping over buildings for miles," Portnoy said.

And, given how easy he and his colleague found it to reverse-engineer the code, Portnoy said, "It's almost like (Disney) didn't even consider security."

Gaming the games
Last up was Avi Rubin, a professor of computer science at Johns Hopkins. He talked, also relatively briefly, about how easy it is for some cheaters to exploit the game of online poker.

Essentially, Rubin argued, a hack called a Sybil attack--which employs fake people participating in games--makes it possible for online poker players to gain a big advantage over their opponents. That works, he said, by making it possible for a single player to control multiple hands in a game, allowing that person to see more cards than they would otherwise, and get a better handle on the odds of their own hand.

For example, he said, in a game of Texas Hold'em, a player employing a Sybil attack on an online poker game could control multiple hands and see things like whether the fives or eights they need to complete a full house and beat an opposing player's flush had already been played.

Rubin's point, then, was that game operators need to work harder at identity management, in order to keep players from employing such exploits. He didn't, however, offer any solutions as to how to do that.

All told, the panelists made it clear that just about any kind of online game or virtual world--especially those where money is on the line--is subject to some kind of hack or exploit, and that for those with the skills to launch such attacks, the barriers stopping them are easily surmountable.

The lesson, then, is that publishers of such games need to think harder about how to manage their players' actions and expectations. Otherwise, players may find themselves in games that are so compromised that the economies collapse and the fun disappears.