How Microsoft's engineers created the world's first broadband-connected game console, based on interviews with the people who made it.
There was a dream that was a box.
Part sci-fi, part nostalgia, this box would unite people from around the world and provide a platform for entertainment that would drive the future. In young computer scientist J Allard's dream, this box would change the world.
The dream started with an early 1980s television commercial for the Atari 2600. In it, entire families huddle around a plastic-and-wood-grain machine, playing versions of popular arcade games, cheering, laughing and stopping to stare in awe. It was a ludicrous and obnoxious marketing vision and it sold those plastic-and-wood-grain machines by the truckload. Its message: People unite through play.
But it was limited, this machine. It was local. To unite the world, Allard's dream box would have to connect the world. That's where the book came in.
Part two of Allard's dream came in the form of the 1992 novel by then little-known visionary author and journalist Neal Stephenson. In the novel, called Snow Crash, Stephenson's protagonist is a sword-wielding, high-stakes, pizza-delivering ninja by day, and an internet-dwelling master sword fighter by night — on the internet. He is sarcastically named "Hiro Protagonist."
Before the internet was a part of the cultural consciousness, this book described it as a lived-in universe filled with interaction. Before there was a Second Life, a Facebook or an Xbox, Stephenson wrote of a world people could visit and enjoy over fiber wires, across oceans. It was a vision of a world united. For young computer geniuses like Allard, it was mind-bending.
Allard, fresh from Boston University, latched on to Snow Crash like it was an oxygen tank. The book inspired him to make its vision of the future a reality. It bounced off the Atari commercial in his mind, shook loose the fantasies of his youth and the two halves of his dream recombined into a plan: Find a job at a place where he could build a box to make his dream a reality.
And that's where Microsoft comes in.
X marks the box
In the 1990s Microsoft was known for one thing: Windows. By 1999, Microsoft's flagship operating systems, Windows 95 and Windows 98, resided on roughly three quarters of the 100 million PCs sold worldwide, making Windows the most ubiquitous PC operating system for a world of consumers quickly discovering that computers were the window to the future.
All of those PCs running Windows also needed word processing programs, accounting software and an array of what we now call apps. Microsoft made the operating system, and then slowly and inexorably began making everything else. And if it found someone who was making something it didn't make, it would make that, too. And if the other company's was better, Microsoft would simply buy that company. Easy.
It was a business strategy many likened to Star Trek's Borg; Microsoft didn't acquire, it assimilated. And resistance was futile. To say Microsoft dominated the PC market would be an understatement. The company owned the business software market, and it had the cash to ensure it always would. And then it sets its sights on home entertainment.
In the 1990s, Microsoft was either the most well known software developer on the planet, or the least well known video game publisher, depending on who you asked.
"We were clearly leaders in building operating system technologies [at] the lowest levels of how that hardware worked," says Microsoft's current Corporate VP of Xbox Live, Marc Whitten, one of the first employees of the Xbox team. Much of the software driving Xbox and Live came from Whitten's hand. "And we were building really deep infrastructure around how software servers worked at the time. ... And we had a game studio."
Microsoft had been tinkering with games for over a decade. The company shipped inveterate time-wasters like Solitaire and Minesweeper with every Windows PC, but in the mid-90s it started to get serious about wasting serious time. The company acquired MechWarrior creator FASA, and used FASA's talent to forge a killer internal game studio.
"DirectX was really important as a starting point. Sony, Sega and Nintendo would struggle long-term given their lack of software and service experience. The future of entertainment was in our DNA."
Meanwhile, on the platform side, the company was making moves to consolidate PC game development under Windows, which developers and gamers had traditionally abhorred. The company began rolling out sets of software tools (called APIs) designed to form a bridge between software and hardware, allowing developers to program only once for the Windows APIs instead of dozens of times for every available piece of hardware. These APIs were eventually bundled together as one product, which Microsoft shipped to consumers for free with Windows 95. It was a Trojan horse designed to spur development for the Windows platform, and it was called DirectX. And it wasn't the only gift horse in the Windows 95 box.
Throughout the early 90s J Allard had been steadily working his way up the ladder at Microsoft and into the awareness of its leadership. While Win95 was still in development, he sent a now-famous email to Microsoft founder Bill Gates extolling the virtues of the internet. Windows, he advised Gates, could become the internet's "killer app," the program that would own the technology for years to come. All Microsoft needed to do, Allard argued, was bundle internet connectivity into Windows 95, making it "internet-ready" off the shelf. Gates agreed and the result was astounding.
Arguably the first modern operating system, Win95 opened PC computing and the internet to the consumer market, and DirectX, combined with Allard's bundled internet connectivity software, helped usher in an explosion in PC gaming that remains legendary to this day. And then, Microsoft decided, it was time to reinvent the game console. Allard was tapped to head up the effort, to bring his dream of a box to fruition, and it was dubbed "Xbox" after "DirectX."
"I believed that entertainment would become a software business," says Allard. "Just like [the industry of] business tools ... and that someone would lead the 'digital entertainment platform.' DirectX was really important as a starting point because a) we had credibility, and b) the medium of gaming was the closest to "programming" and "software" as anything was in the late 90s. If you buy that — that entertainment would become a software business that would center on the cloud — then Microsoft was very natural to play a role in gaming, and Sony, Sega and Nintendo would struggle long-term given their lack of software and service experience. The future of entertainment was in our DNA."
The building of Xbox would be a monumental engineering and software development challenge for Microsoft. The company not only had to prove — to consumers as well as game makers — that it knew the game business, but it would have to leverage hardware production and distribution, taking big gambles on the direction of console gaming and entertainment in general. And it would have to do so in less than two years from concept to delivery.
It was a risky proposition for Microsoft, which had never created such an ambitious consumer product. And Allard extended that risk even further, deciding early on that his Xbox would not use the then-common dial-up method of connecting to the internet. Xbox wouldn't even have a modem port, as would other game consoles of its generation. Instead, Xbox would have an Ethernet port. Xbox would be broadband.
Bet on broadband
Bill Gates loved the idea of of an internet-connected game console. He understood what going online would mean for the box and for Microsoft. He got it. But he didn't like the Ethernet port. And he wasn't alone.
Over a decade later, every Microsoft veteran spoken to for this story remembers the decision to install an Ethernet port in the Xbox as the single most controversial, most fraught with peril risk the Xbox team took. And the most hotly debated.
"A lot of stuff at the time was still dial-up," says Whitten. "But we didn't want to build for a constrained world. We wanted to build for a world where you had higher bandwidth, where you had consistent access to that bandwidth. We knew that then, if you could build a service, that would be really powerful."
Xbox wasn't being designed to compete with modern devices; it was being devised to compete with devices of the future. "Project Orange," for example, was a scheme for selling entire games over the internet, with no physical disc. Something that has only become possible within the past few years, and which (much like the original Ethernet port) is widely controversial.
For Allard, the future of entertainment meant broadband internet and all that the much higher bandwidth could make possible. Bill Gates — and many within Microsoft — didn't agree. They believed dial-up internet still had legs, and the Xbox's Ethernet port became a topic of heated debate.
In 1999, broadband internet was still an unproven technology. It was a toy for the rich or well-connected. It was growing, but slowly. Allard believed that growth would accelerate and he held his ground. Broadband, Allard believed, was coming and would be ready in time for Xbox. To aim lower would be to miss the mark.
"We wanted to create an awesome gameplay experience for users and that meant 'no compromises,'" says Allard. "And having as ambitious an online strategy as we did with graphics, audio, controllers and content, dialup was simply 'beneath' our ambition and our commitment to gamers and the industry to 'take it up a notch.'"
The debate would rage on for months. In October of 1999 Allard made a bet with Xbox General Manager Cam Ferroni. Allard bet Ferroni $1,000 that he would win Gates over on broadband by December of the same year. Allard lost that bet. Ferroni got his check, but Allard would eventually win the debate. Just as he had once convinced Gates of the importance of the internet to Windows, he convinced him again of the importance of broadband to Xbox. But convincing the Xbox team itself would be a longer battle.
"Allard says to me, 'All right, we put the Ethernet port in the box. Go figure out what it talks to.' That was the mission. I was the first person on Live. I hired everybody. I hired the whole dev team."
In August of 2000, Boyd Multerer was hired as development manager for what would become Xbox Live. At the time it was called simply "Xbox Online."
"Allard says to me, 'All right, we put the Ethernet port in the box. Go figure out what it talks to,'" Multerer says, laughing at the memory. "That was pretty much the mission. I can say I was the first person on Live. I hired everybody. I hired the whole dev team."
Multerer kept his standards high, pulling in programmers from the core Xbox team as well as from other departments at Microsoft. Turning away people who were "really good" because they weren't "really really good," Multerer assembled a team of roughly 15 people who, over the course of two years, would build Xbox Live from the ground up. Their plan was to sail under the radar as much as possible, keeping costs low and requests small so they could control as much of the destiny of Live as possible.
"We were all kind of low-budget and incubation at some level," says Whitten. "We were reinventing how to think about gaming in the living room and how to think about how a service would play on top of that. We didn't want to bring this really broad-scale effort to it. We wanted to do foundational thinking on something we could build from. We had a pretty small team."
Part of the reason for the small team was to avoid an all-out fight with the core team on Xbox, which had begun to grow suspicious of Live's drain on resources. This effort was not entirely successful. Many at Microsoft, including some on the Xbox team, saw Live as a distraction, at best, and at worst a case of terminal mission creep that could jeopardize the entire Xbox effort. A project doomed to fail.
"I heard an analogy at the time, which I've seen applied to many projects," says Multerer. "The analogy is, any project that you're doing, if it's radical and new, it's sort of like having a horse. At first all the other people look at your horse and say, 'Aw, what an old nag, somebody oughta put that horse down.' Then the horse lives on and you move along and you start making some progress. Then they say, 'Oh, why is that horse getting so much attention? We need to kill the horse and not have it as a competitor.' Then it survives and it moves along, and then it's like, 'Hey, this is actually a business driver. I want to hook my cart up to the horse!' I've seen so many projects go through this. It's not just a Microsoft thing. This is just humans."
The Live team, driven by Allard's vision, was betting that broadband would start to grow faster, get into more homes less expensively and be the horse that would pull Xbox into a new century. And they were working alongside internet service providers to hedge that bet.
Michael Mott was Xbox Live's director of business development. He and Developer Relations Manager Scott Henson would travel the United States, tag-teaming presentations to internet providers, extolling the virtues of Xbox Live. Mott provided the business perspective, Henson the technical vision.
Among broadband providers, the proposition was a hit almost instantly. They could see the future and they had a vested interest in pushing as much fiber wire into as many cities as possible. And, inevitably, there was always one engineer in the room who was also a gamer, who helped sell the idea to his superiors. Not only were broadband providers interested in Live, they were willing to bundle access to Live to new broadband customers, and tweak their protocols, to ensure Live traffic wouldn't get throttled or constrained.
Xbox Live would garner near universal support from internet providers, all of whom were also betting on broadband. Yet the naysayers at Microsoft still weren't convinced.
The greatest sword fighter in the world
It's impossible to overstate the radical proposition posed by Xbox Live at the turn of the century. This was a new millennium, but it was still bound in the thinking of the old.
As the new century dawned, the entire world had waited with anxiety as clocks clicked over to the stroke of midnight, Jan. 1, 2000, for fear that computers the world over would grind to a halt, casting the globe into anarchy, thanks to Y2K.
Y2K was the now infamous computer glitch and attendant paranoia resulting over the inability of many early computer systems, programmed in the late 1940s using the early programming language COBOL, to account for dates with more than two numerical digits. The fear was that these computer systems — managing things like satellites, power grids and Defense Department mainframes — would reach midnight on Dec. 31, 1999 and then suddenly believe they were traveling backward in time, to the year 1900, and shut down completely. These fears were not entirely unfounded.
Internally at many of the large corporations and government institutions responsible for managing these systems, the debate raged only over how severe an impact it would be. Yet by late 1999 almost everyone was convinced there was a problem, compounded by the fact that few programmers working at the turn of the century even understood COBOL.
Imagine a visionary programmer who, in the late 1950s, suggested to his colleagues that computers begin tallying dates using four digits instead of only two. Because some day, this programmer would argue, these mainframe computers and weather satellites would create a virtual web of machines controlling almost every aspect of human society. And in the year 2000 — a date in time so far distant many considered it the stuff of science fiction — those machines would still need to function.
In the 1950s, such concerns were considered asinine. And the man who actually raised them — the man who created the precursor to the COBOL programming language — was ignored. For nearly 30 years.
Bringing a vision of the future into being takes courage and will. It takes the ability to convince others of the rightness of that vision. It takes the skill and expertise to see clearly the steps from having that vision to making it real, and the gift of motivating others to go against the prevailing wisdom of the day and help forge a new path.
In 1999, the prevailing wisdom of the game industry was that consoles were standalone units. You did not talk to people while playing a console. And even if you did, there was no keyboard for text chat. Games sold consoles, they would say. Not internet or software. Because that is how it had always been. And besides, if you're going to connect a console and sell it to the broadest audience possible, then connect it via a modem with a phone jack. Because everyone already had one of those.
"We had strong opinions. And that mattered. It really mattered for the first time at Microsoft."
Allard believed otherwise, and he was bucking deeply-held, long-established trends. Yet he was no raving lunatic. He was, by all accounts, a visionary. And when he said, "There should be a box," people listened.
"He was fantastic at telling a story. 'This is where we're going and this is why this is important,'" says Xbox Live's Operations Program Manager Eric "e" Neustadter. "I will never forget my first meeting with [Allard]. He blew my mind. ... I knew a lot about him, but I didn't know a lot of the details about him. 'Impressive' would not be a big enough word."
Neustadter was among the first called in to create what would become the Xbox Live "op center," the nerve center for controlling what would become a global system of data centers. Early in the Live project, Neustadter was called in to present his plans to Allard and the Xbox leadership. He put together his PowerPoint slides and walked into the room, not knowing what to expect.
Allard arrived late, wearing a cast on his arm. He'd broken his arm snowboarding, again. A fan of extreme sports, Allard is remembered as having frequently broken something. An occurrence that led many to believe he was more brogrammer than programmer. He would dispel these misconceptions almost instantly.
"So we're waiting," says Neustadter. "He comes in. I've got my laptop with my slides. Here we go. My expectation, honestly, because this was my boss's boss, was that I was going to present at a very high level and [Allard] would nod and ask a couple of questions and that would be the end of that. I was not prepared for his level of knowledge. ... He proceeded to grill me in levels of detail that absolutely blew my mind. I had no idea he had that level of knowledge."
Xbox Account Manager Ric Neil describes another meeting in which Allard was sharing his excitement with the team over the amount of extra memory he had been able to put "on the metal" in the Xbox. Neil argued that extra memory in the retail Xbox wouldn't matter if the test kits they were developing didn't also have it. He pushed back on Allard.
"Without debug memory, more memory didn't matter to anybody," says Neil, recounting the meeting. "If I can't have extra memory on my debug box to test the extra memory, that's not really extra memory. I kinda made that obvious in the meeting. He wanted everybody to be excited about the extra memory, and I was like, 'Thanks, but no thanks.'"
The room turned silent. Neil's boss turned to him and stared. It was one of those moments when the sound of a heartbeat can fill an entire room.
"I called him out in front of everybody and the whole fuckin' room was just silent, looking at me," Neil says. "And I'm like, 'Oh, I must have fucked up.' And [Allard], after the meeting, he's like, 'Hey, you're the only one with balls enough to call me out.'
"We had strong opinions. ... And that mattered. It really mattered for the first time at Microsoft."
Allard shared his vision with the entire Xbox team. He gave copies of Snow Crash to everyone, and made everyone read it. There was no subtlety about his desire to bring that fictional world to fruition nor his desire to eventually become Hiro Protagonist.
There's a scene in Snow Crash, a dialogue between Hiro Protagonist and a character named Juanita. Protagonist is describing a virtual sword fight that took place in the interconnected virtual world he inhabits in his off hours. Allard, to this day, without prompting, is able to quote the scene verbatim.
"Did you win your sword fight?"
"Of course I won the fucking sword fight," Hiro says. "I'm the greatest sword fighter in the world."
"And you wrote the software."
"Yeah. That, too," Hiro says.
"Well, we all read Snow Crash," says Multerer, laughing. "J was certainly going on about it."
Multerer believes that roughly half of what would become Xbox Live was inspired by the novel (or Allard's retelling of it) and the other half were simply practical concerns about the future of gaming on consoles.
"For me, at the roots of Xbox Live, it is a way to find friends and get into games with them," Multerer says. "That means you have to have an identity. That's super important. And you need to be able to find people and see what they're doing and then invite them to play with you. Those are the key ingredients that made it work."
This "social layer," although few at the time called it that, would be the glue holding the entire service together. Years before Facebook, the Xbox team was working on what would become one the most widely used social networks in the world.
And when Xbox Live goes online, Allard's gamertag, the name by which all users are known on the service, will be HiroProtagonist.
I work at Millennium E
In 2001, Microsoft employed tens of thousands of people. Ask most of those where they worked, and they'd tell you "Microsoft." But if you asked the 30 or so people who formed the core of the Xbox team, they'd say "I work for Xbox."
"We had a culture that was unique," says Henson, "but at the same time we weren't so unique that we went and said, 'Yeah, we'll just go reinvent everything.' That would have just been completely crazy. And it wouldn't have allowed us to scale the way we did."
Most of the Xbox team people came from within Microsoft and knew what the company had to offer. Xbox wasn't about building a team from scratch and reinventing the wheel; it was about building a rocket using rocket parts made by a company that was already building rockets. And of course, it also had rocket scientists. Real ones.
"[Lead Technical Manager Jeff Pobst] is a rocket physicist!" says Neil. "I was working for a rocket scientist. I graduated from Woodside High School. And Jeff would be like, 'Dude, you know more about GPUs and CPUs than I do. Whatever. Shut up.' Once you close the doors at Microsoft, it's that level of respect. I might have argued with J and Jeff. I made Jeff's life impossible sometimes. But there was still a level of respect. We still liked each other at the end of the day, and we made something great happen. Because of that, because of the passion. We all cared about making it happen.
"I don't think it was cowboyish so much as it was just super passionate people."
"Smart people is not the word for it," says Neustadter. "These were people who were so much smarter than the people I was used to working with it was scary."
Xbox veterans refer to the days of the original box and the birth of Xbox Live as "the Golden Age of Xbox."
"It was awesome," says Jeff Pobst. Now president of Hidden Path Entertainment, in the early days of Live, Pobst was the lead technical game manager for Xbox's Advanced Technology Group, the team of programming and game development veterans Microsoft would send to third-party developers to help solve problems and make their games run smoothly with Xbox. "It was so fun to join this team. ... It had a startup culture. It had really passionate people. It had everyone wanting to do a lot of the right things."
The opportunity to take on such a bold, risky project and potentially change the way games were played attracted a particular kind of talent. The result was a sort of skunkworks inside of the traditionally corporate Microsoft, where a lot of the traditional ways of making products were thrown out the window.
Xbox producer Jeff Henshaw, drawing on an analogy widely used inside of Microsoft, describes the original Xbox team members as "pioneers."
"There are pioneers, and there are settlers," says Henshaw. "The pioneers are the people that take that risk to create something new. Not just a new product, but even a new culture or a new philosophy around how you build products. The settlers are the people that come in after things are established, after the product is built, after the business is established. They grow it and nurture it in a safe, conservative way. But pioneering is where all of the fun is in technology.
"Everyone that joined Xbox to do that was actually making a conscious decision to give up a successful career somewhere else in Microsoft. There were safer things to do at Microsoft. ... People consciously gave those roles up because they wanted to work on something new, where they felt like they could have a tremendous impact and really satisfy this lust to make video gaming better than it had ever been before."
"There were safer things to do at Microsoft. People consciously gave those roles up because they wanted to work on something new, where they could satisfy this lust to make video gaming better."
The Xbox Live team of pioneers was small and nimble, but perhaps most important, it was headquartered away from Microsoft. One of Allard's earliest (and continual) fights was keeping Xbox separate from the main campus in Redmond. Instead, it was housed down the road in the rented set of buildings known as Microsoft's "Millennium Campus." Xbox was in Millennium E.
In terms of actual physical separation, Millennium E was only a five minute drive away from Microsoft. In terms of some functional separation, the distance didn't matter at all. Email and telephones still worked just as well in Millennium E as in any of Microsoft's other buildings. But in terms of psychic separation, that five minutes down the road was critical to Xbox's early success. And maintaining that separation became a passion for Allard.
"We definitely needed that head space to get Xbox right," says Henshaw. "We needed that head space to go engage with a different type of partner, with game developers, than we'd engaged with before. And honestly we needed some of that head space to tell Microsoft to leave us alone with the other agendas and let us go build something that's focused and does a few things around gaming really magically."
Neustadter believes there were decisions made at Millennium E that would not have been possible had Xbox been built closer to home.
"We were a little isolated," Neustadter says. "That slowly vanished over time. I mean, now we're on this shiny new campus across the street from main campus and we're very integrated with the rest of the company, down to the point of being able to log on to Xbox.com on my phone, or get achievements on PC and Windows Phone and iOS and all these other places. But we had to build those relationships later. We didn't have any of them, really, at start time.
"But at the same time, given the goals that we had when we launched, I think it was the right thing to do. I think it helped us be successful. … There were a lot of decisions that we made that probably, had we made them in full view of the people in the rest of the company, they would have said, 'You can't do that.' ... So we didn't worry about that kind of stuff. We just worried about, is this going to make Xbox Live amazing? If so, we'll do it. If not, don't care."
Over time the Millennium E building became bedecked with green lights (the brand color of Xbox) and other decorations to physically differentiate it from the main campus. It was furnished like a game development studio, with games to play, televisions to watch and comfortable places to sit, and as the team worked through 2000 and 2001, its status within Microsoft grew legendary.
"Xbox was cool," says Pobst. "And to have a part of Microsoft that was cool had not really happened before."
When new positions opened on the Xbox team, the response from within Microsoft could be overwhelming. Xbox was the place to be inside Microsoft.
"It was the greatest thing ever," says Whitten. "To be clear, it was awesome. ... The way I describe Xbox, especially back then when we were a much smaller team, is I say we were kind of like the world's best funded startup. We had great talent and great passion around what we were doing, and yet we're at this company that did two things really right for the launch of Xbox, I believe."
The first was that Microsoft funded the project well. The second was that it recognized it had no idea how to build a console and stepped aside to let Team Xbox figure it out.
"[Microsoft] was like, 'We're not telling you how to do this. We're putting this team together to go learn how to build this,'" Whitten says. "We were given a lot of latitude to have a real vision and to have a long runway to go get this right. That was pretty special."
Disney World not Coney Island
In 1999 Allard and his Xbox leadership team identified the principles Xbox Live had to embody in order to be successful. They were:
Users had to feel as if the service was secure, and that their information and identities would be protected. The service had to work the first time and every time. It had to reward good behavior and punish those who took advantage. It had to be effortless to connect with and able to adapt to the needs of different games and experiences. It had to be a level playing field, free from exploitation by hackers. And it had to continue to provide a premium experience that users would be willing to pay for, year after year, game after game.
These seven "pillars," as former Live team members call them, were enunciated by Henshaw in a PowerPoint presentation to Microsoft in the Summer of 2000.
"I remember it very clearly," Henshaw says. "It was a meeting that was held by [President of Entertainment and Devices] Robbie Bach ... We had the whole leadership team present there. We had J Allard ... We had Ed Fries, who was at the time running Microsoft Game Studios. And then a handful of other leaders from around the organization.
"It was me pitching this vision, and a slide deck of what we thought would make a truly innovative, truly differentiating online experience on Xbox that people would not be able to get anywhere else, including on PCs."
Chief among the seven pillars, Henshaw believed, was simplicity. Xbox Live had to "just work," and it had to be consistent across all games. The worst crime it could commit, the Xbox team believed, would be to ask gamers to create separate accounts and logins from game to game to game. Xbox Live had to be one service from top to bottom, and the makers of games had to fall in line.
"As we laid out these principles," says Henshaw, "everyone looked at the vision and said, 'This is absolutely what we should build. These principles are solid. We believe that these are principles for an online service that is designed for the long term, designed for the future, and designed for growth. We will build this.'"
The team adopted a mantra: "Disney World, not Coney Island."
"Disney was able to deliver a world class experience for their customers [at Disney World in Florida] because of infrastructure," says Allard. "What happens underground is more impressive than what happens above ground. Miles of tunnels route through the park to allow them to keep the environment clean, safe and functional unlike any other park. Because of this, they are able to preserve more of the illusion they wanted to deliver to their customers.
"When you have a Burger King wrapper on the ground and a couple of punks harassing paying customers and your customer's car gets keyed on the on-street parking and your rides break down ... Coney Island is just, well, a distraction. Families save for years to go to Disney World because of the unprecedented experience — they've created a destination."
As it set about creating Live, the core team envisioned it as the digital Disney World. The Xbox team worked with broadband partners and router manufacturers to ensure the service would work reliably on available hardware and that the bandwidth wouldn't get shut down by ISPs. Meanwhile, working with the enterprise server experts at Microsoft, it began sketching out the infrastructure of the service, pulling servers from one place, database tech from another and installing a stringent security layer on top of all it. Staking out a place in the wilderness and creating an unprecedented experience that gamers would — hopefully — be willing to pay for.
"We had found this space that was like the center of Florida," says Whitten. "We were going to start bringing in roads and plumbing and all of this infrastructure so that great experiences, games, would come to life inside of it."
As PC gamers, the Xbox team members were familiar with online experiences, but they were mostly turned off by them. Early PC multiplayer worlds were plagued with cheats and exploits, and finding your friends and getting connected to them was a chore. If you were able to get a game going in the first place, it would often be interrupted or disrupted by inconsistent connections or griefers empowered with hacks.
"I remember playing Diablo. I remember having a blast with Diablo, and then the online hacking and cheating started," says Multerer. "You could see the sales drop off a cliff. It became not so much fun anymore. People didn't go online, because you'd go online and immediately someone bigger than you would show up, kill you, take your loot, and disappear again. It was all being hacked. ... That became a teaching moment for us. We said, security has to be a primary feature. That's a place where we can add value."
"What happens underground is more impressive than what happens above ground."
The solution devised by Multerer was to employ a single security gateway in front of the Live servers. Each Xbox would send encrypted data to the gateway, but behind the gateway the servers would treat the requests for information like normal web traffic. This allowed him to build what, at the time, passed for a fairly standard web server database without the expense of having to encrypt every single server. The Live team would create the largest secure online service ever devised by treating it like a website.
"This way we were able to focus on just the pieces we needed to make the system work and still use standard technology that came from other parts of the company and other parts of the world," says Multerer. "To me it was a neat example of staying focused on what's the enabling piece of technology, while still being able to use all the standards that you possibly can, so you don't have to reinvent everything.
"It was wildly fun. We were looking at expensive equipment, backbones ... We had a big room over in Millennium E where we built out a couple of test versions of the service, and it was a blast to just go back there and look at all the wires and the heat. It was all new. That style of web development was just getting going. "
When Neustadter joined the Xbox team in April of 2001, Xbox Live was just a collection of PCs under people's desks.
"There was no lab," Neustadter says. "There was a room that was designated as, 'this will be our lab,' but the team did not have a single server, a single router, nothing. For a while, until we got the team off the ground, I was the team's only server guy. I was the team's only network guy. I was the team's only security guy."
In just over a year this collection of random computers would morph into a service comprising more than 300 servers, complete with an operations center from which Neustadter and his team could survey and control every aspect of the Live service. Using off-the-shelf technology and Microsoft expertise allowed the Live team to build quickly and cheaply, but it had a downside: It built too much.
"Pick any two services at random. Let's say matchmaking and leaderboards," says Neustadter. "They could have easily lived on the same servers, and there was no reason why they needed their own hardware from a capacity point of view. But we didn't do that, because we wanted to simplify, as much as possible, both the troubleshooting process and also making sure that if we had to go touch matchmaking, to roll out an upgrade, we didn't also have to worry about any implications to leaderboards at the same time."
Today, in the world of iTunes, Google and global data centers, 300 servers sounds quaint, but in the early days of Live it was overkill. Xbox Live had server clusters for every single aspect of the service, no matter how small or how many users it would serve. It was the metaphorical miles of tunnels underground, supporting Live's seven pillars.
"We were using way too many machines," says Multerer. "We didn't need to use that many machines. But we were paranoid. We didn't know how to build these clusters, so we had a cluster for everything."
Eventually Live would be re-scaled to support various services from shared servers, but, for a time, it was the most extravagantly-designed server center in the world. The Live team members were wiz kids in a technology store.
"Back then, we basically had this super magic advantage, which is that we were starting from scratch," says Whitten. "We didn't have to build it on something else. This was our consistent experience we were building around. And we were at Microsoft, which meant we had access to ... around 50,000 of the smartest software and service engineers on the planet."
As the team at Millennium E was devising how the system would work, the effort was also underway to devise how games would work with it. A dedicated online console service was new to not only the people building it, but also the people building games for it — and the birth was not entirely painless.
"It was pretty tough to get a consistent experience around the friends list," says Whitten. "One of the best examples was downloadable content, where we started with a bunch of APIs and it was really tough to get it done. A lot of people were just saying, okay, I'm gonna drop this, because I can't get this done in a way that's going to work for our game and for users.
"We ended up having to go back and write our own app to do it separate from the games ... we actually built an app called Downloader.XBE that they could drop into their disc, and put some little mild skinning on it to make it look like their game. If you went to go click on 'Downloadable Content' in the menus of those very early games … it would actually reboot the box into this downloader app. You'd download your content and reboot back into the game, which allowed us to get through it. In fact, that was the genesis of a lot of our thinking in how we did [Xbox 360]."
Two minutes into the game, backup quarterback Matt Hasselbeck drives downfield and scores for an early lead against the squad of Peyton Manning. History is made. Just not football history. This is not week seven of the 2013 NFL season. This is 2002. This is a game being played on Xbox Live.
Manning and Hasselbeck, separated by a continent and each surrounded by film crews, are playing Microsoft's NFL Fever 2002. They are online, playing against each other on a game console over broadband internet. And they are trash talking.
"He had to sub himself in!" cries Manning, then head quarterback of the Colts.
And it's true. Hasselbeck, as he would be again in 2013, was a relief QB. In 2002 he played for the Seattle Seahawks, and it is from his Seattle home that he is beating star quarterback Manning fair and square in a video game played live and filmed for posterity by Microsoft.
"From that immediate point, I knew there was absolute magic here," says Xbox Account Manager James Miller. "These guys had completely forgotten that they were in a press thing, that they were playing thousands of miles apart, and instead they were absolutely trash talking exactly like that vision had always promised. That, to me, was the first time we'd really seen it work cross-country. It was such a powerful way that it happened."
By the spring of 2002 Live was ready to road test. The box had shipped the previous holiday season, and the Xbox team had begun sending out Live beta kits, at first to Microsoft employees, then wider. Around a thousand of the kits went out in the first batch, some to far corners of the globe. Everywhere daylight shone on the Microsoft flag. Each kit came with a headset and a game — a version of Midway's RC racing game Revolt, re-engineered to function with Live.
"[Revolt was] the perfect game to prove out the service," says Henson. "Eight players, multiplayer, voice."
The team wanted to ship a simple game with Live so that players new to online experiences wouldn't also have to learn how to play the game. With a game like Revolt, it would be as simple as picking up the controller.
"There's a night I so distinctly remember in April of 2002," says Henson. "We had the team from Japan and the team from Europe and the team from the United States all on at the same time. We're racing around, talking to each other, choosing cars, laughing, and we did this for hours and hours.
"We were done and we were so excited because we knew, at that moment ... No one in the public had seen this. We hadn't shown it. E3 hadn't come around. But we knew ... Everything we talked about, it was going to happen. This was going to be an amazing thing. We didn't know how big it was going to be, but this experience we'd just had, it was going to set the world on fire."
"I asked her where she was and she said 'Del Mar, California. Isn't this thing amazing? We've been playing for 5 hours!'"
By September the kits were going out to a wider audience. Limited beta testers. Friends of friends. The number of beta kits eventually reached 15,000.
Allard's HiroProtagonist gamertag was a regular sight during the beta tests, usually playing against Xbox team members. One night, playing Revolt, he heard a female voice, laughing. It was a voice he didn't recognize. Allard asked who he was playing against and the woman replied that she wasn't playing, just chatting with the other drivers while her son played the game.
"She laughed and I realized ... holy shit, we're live!" Allard says. "I asked her where she was and she said 'Del Mar, California. Isn't this thing amazing? We've been playing for 5 hours!'"
Allard was awestruck. His vision of a connected device bringing a family together over a game had come to fruition. In just two years he'd built the box of his dreams. And to outsiders, it was everything he'd wanted it to be.
"It was magic," Allard says. "It was flawless. Here I was talking to a 40 year old woman that doesn't play games sitting on the couch with her 15 year old kid and talking to people around the world all day as the kid tried to work his way up the leaderboards. It was that Atari ad. We did it. Fucking awesome."
The official launch came two months later.
On the night of Nov. 14th, 2002, Microsoft held an official launch party in Hollywood. Samuel L. Jackson was there. Freddie Prinze Jr. was there. They were playing NBA 2K, online. For the Microsoft executives and Xbox team members it was a brush with fame they didn't typically see. And Microsoft products didn't typically get launch parties at The Sunset Room.
The next day in Redmond there was a celebration at the Millennium E cafeteria. Multerer was called back to the Op Center. There was a problem with the numbers coming in from the service. Multerer investigated. It's actually a bug in the reporting software. The numbers were fine. More than fine. They were breaking every projected target, and then some.
"I remember being there when we turned the service on," says Whitten. "Just when you're sitting there in the op center watching the counters of users joining the service. There's nothing more magic than that."
To mark the occasion, Mott, Ferroni and Allard each grabbed a controller in the cafeteria, fired up Xbox Live launch title MechAssault and waited for one of Live's earliest subscribers to join. When he arrived the fathers of Xbox Live were there to greet him.
"We said, 'Hey, you're on with the Xbox Live team and we're playing one of the first official multiplayer games online,'" says Mott. "That was really cool. That guy was like, 'Oh, this is awesome!'"
Meanwhile Henson and other team members in LA were getting reports from Redmond.
"Every number we ever thought we were going to have was completely shattered on the first day," says Henson. "It's like that commercial where they're sitting there watching the counter. Uh-oh. We've got a new problem. We've already exceeded all the goals we thought we were gonna have in a day."
"We all hung out [in Millennium E] for the next couple of days and watched all the data we could," says Multerer. "If something didn't work right, we were fixing it on the fly. We were running the service."
Live was up. Live was working. And people were connecting. They had done it.
Now all they needed was two things: a game that would drive consumers to the service, and a partnership that would all but guarantee its success.
One would be easier than the other.
The EA curse
The world is full of mystery. Historically, it is not a new thing that we humans, observing these mysteries, will stare into the abyss and make up our own answers. Why does the sun chase the moon? Why does fire lash down from the sky? Why do things inevitably fall apart?
Take, for example, the Madden curse. The Madden football franchise defines the sport on video game consoles, and for over a decade, each new entry in the series has arrived in stores emblazoned with the likeness of a popular player. There was a time when this was considered an honor. That was before the curse.
Since 1999, nearly every player asked to be on the Madden cover has suffered injury. Some of these injuries have been minor, some career-endangering, but almost without exception, every player on the cover of Madden has suffered one.
Is it scientifically provable that by simply appearing on the cover of a game, a player will fall victim to injury? Doubtful. More likely is the explanation that football is a dangerous sport and that these things tend to happen, video game covers or not. But humans have evolved to find patterns in the world around us. Take, for example, the other EA curse.
On Sept. 9, 1999, Sega launched the Dreamcast console. Expected to be a technological breakthrough, it instead crumbled, taking the company's entire hardware business with it.
On Oct. 4, 1993, a company founded by EA co-founder Trip Hawkins released the 3DO game console. Like the Dreamcast, it was packed with cutting edge hardware and lauded as a brilliant entry into a competitive market. And yet it, too, crumbled.
What do these two game industry failure stories have in common? Quite a lot, actually. Both consoles launched at volatile times for the games industry amidst stiff competition from nimble rivals. Dreamcast was outperformed by Sony's PlayStation 2, Nintendo's GameCube and the Xbox from Microsoft. 3DO was inevitably outshone by Sony's Playstation, Nintendo's Nintendo 64 and Sega's own Saturn.
Market forces, relationships with game developers and publishers and some factors beyond anyone's control all combined in a wretched stew to foil the ambitions of the companies behind these products. Yet, ask anyone and they'll tell you different. They'll blame EA.
It should surprise no one that EA, as the historically largest and most influential game publisher, is an attractant of curses. Its fortunes rise and fall in tandem with those of the industry it dominates. "There must be a reason," we cry. And there usually is. But it can't be so simple as good leadership, canny business prowess and luck. It can't be, we reason. Because that's too simple. And, perhaps, too boring.
Yet whatever the cause, the curse had its teeth into the Xbox Live team. EA had to be on Live, or Live would fail.
The problem: EA had its own servers. You could already play Madden online — for free. The big question for EA was, why would it want its games on a service where its players would be bombarded with information about other games? Why share?
"The idea of a community and finding your friends and playing games with your friends requires that I have a single identity and I have a single friendship and I can move from game to game," says Multerer. "The game developers, at the time, because social was not a well-known concept, they reacted by saying, 'Wait a minute! Are you telling me I don't actually own my customer? Are you saying you own the customer and I don't own my customer? I don't think so!'"
For the Xbox Live team, this was not a problem localized to EA. In forging a new product inside a massive company, the Xbox team found it had bumped against one of Microsoft's boldest trends: hubris. In the PC software market, if Microsoft didn't own the product, the company would buy the product and then own it. On the PC, Microsoft was king and it expected everyone to fall in line. That didn't work in games.
"It was painful because there was so much talk about 'owning the customer' from their side. I would ask, 'How about we delight the customer?' And suggest that maybe the customer really owned us."
"Being a hit-driven entertainment business, the games industry works completely differently," says Henshaw. "There are business deals around exclusive content. There are business deals around just bringing franchises to a platform. There are deep trust relationships that get built between platform builders, the console manufacturers, and the best game development shops in the world.
"We very early on realized that this was not a business game we had been playing before. We needed help."
Xbox's solution: the Advanced Technology Group led by Laura Fryer. Part salespeople, part engineers, ATG's technical game managers would work closely with game developers and publishers to find the right solutions for the right games. Deals would be struck, partnerships formed and then the software development wizards of ATG would roll in after to ensure that the games developed for Xbox were running as well and beautifully as they could.
"This ATG team, I swear, it really was an all star programming team, in so many ways," says Pobst. "Even the rest of the Xbox organization, there was passion there, but most of the Xbox organization were already Microsoft employees who transferred. ATG had probably the highest percentage of people who came to Microsoft as their first gig after having made games for a living. That gave us ... a better voice out to the customer, to our developers that we were trying to work with."
While Mott and Henson evangelized Live to broadband internet providers, technical game managers like Sam Charchian performed the same pitch for third-party game development studios. But reception among game makers was often less enthusiastic.
"Epic, in particular, I remember going to their office and them just basically ridiculing us," says Charchian.
Epic Games had scored a hit on PCs with its online arena game Unreal Tournament. UT had near total penetration on the PC platform, and everyone playing it was using dial-up. Why would Epic ask its customers, who were already playing UT for free over dial-up modems, to pay for Xbox Live and broadband service just to play it again on Xbox?
"I remember phrases like, 'That's the stupidest thing I've ever heard!' being tossed around the room," Charchian says.
It's a testament to the ability of Xbox's account managers that Epic and other companies eventually saw the light. UT would even be an Xbox Live launch title in 2002. But EA was another matter. For EA it wasn't about just one game; it was about an ecosystem. It was about control.
For years, Xbox's account managers and executives hammered on the problem, getting nowhere. The EA problem seemed irreconcilable, and some at Microsoft began to wonder if the gamble on Live could survive the dreaded EA curse.
"It was painful because there was so much talk about 'owning the customer' from their side," says Allard. "I would ask, 'How about we delight the customer?' And suggest that maybe the customer really owned us."
Microsoft executives chose sides, many coming down in favor of compromising to land EA and stave off the curse. Bach and Allard refused to budge. Live was to be one ecosystem, and EA could be on it or it could go to hell.
"You know, it's a classic two big company conversation, right?" says Mott. "[EA was] a big company that knew their content moves platforms, and they wanted value back for that. We just went back and forth."
The meetings wound on for months, well past the Xbox Live launch. Finally, over the winter of 2002-03, the ice began to break. Xbox Live was growing and thriving. The Wall St. Journal wrote about it. It was working. But it didn't have EA. Meanwhile, EA was struggling to find traction with its own online service on the PlayStation 2, which didn't have the seamless architecture of Xbox Live. And didn't have broadband internet, which, bolstered by Xbox Live, had grown. Just as Allard predicted.
On the other hand, Microsoft's own fledgling "Fever" series of sports titles was struggling online. NFL Fever was a two-million selling game that had roughly 7,000 players online. Sports gamers were still playing on couches instead of virtual couches — because that's where Madden lived. The light was dawning on both sides of the divide that if either company was going to succeed, they needed to work together.
Bach and EA's John Riccitiello met to hammer out an agreement on the business terms. Afterward, Bach returned to Millennium E a conquering hero. EA would bring all of its titles to Xbox Live, on Xbox's terms. Microsoft's concession: shuttering "Fever." EA didn't want to compete against its new platform partner. Microsoft was more than willing to comply. Fever had been designed, after all, merely to fill an EA-shaped gap. And by trading Fever, Bach had gained a partnership with the largest game publisher on the planet, using Xbox servers, Xbox gamertags and Xbox architecture. Bach and Allard had held the line, and it had worked.
The next year, on stage at E3, Bach would shake hands with EA's president of Worldwide Studios, Don Mattrick. Flanked by professional sports superheroes and introduced by Muhammed Ali, the two companies tied up their new partnership in a bow in front of the world. It had been accomplished. EA and Xbox Live were now BFFs.
Xbox engineers began holding summit meetings with EA's own product teams to begin the long, tedious job of combining Xbox Live with what EA had engineered for its PC service. These teams, over the course of months, would sort out how voice traffic was fielded and by what servers, how users logged on to EA games via Xbox Live and how security worked; Xbox had solved the same problems as EA but from different directions. It would be a gargantuan project.
"You can say it started tense, but really everyone there was motivated to find solutions," says Miller. "It came down to ... education, on both sides of the table, in terms of, 'Here's how we built our systems. Here's how our UI draws friends and things like that. Here's how our servers talk to the consoles.' ... I think they were very healthy working meetings. They had a lot of moments of, "We don't know how this is gonna work" when they started, but by the end they had everyone shaking hands and going out to dinner and agreeing that there was an active plan we could go and do."
In July of 2004, EA released the first title enabled for Xbox Live, NCAA Football 2005. Madden and other titles would soon follow. EA was officially on Live.
J Allard would ultimately call this chapter in Xbox Live's history "a speed bump."
"Now," says Mott, "there was nothing stopping us."
Nothing, perhaps, except Halo 2.
When the Xbox went to store shelves in the holiday season of 2001, it rested alongside what would become its "killer app": Halo: Combat Evolved, developed by newly acquired Bungie.
Halo marked a starting point for the Xbox's legacy in many ways. It represented Microsoft Game Studios' entry into the blockbuster AAA genre, the Xbox's first and most influential "launch title" and the first installment in what is now a Halo mega brand.
Like Allard, Bungie's founders wanted to push the barriers of online gaming. They wanted to create the best online experience the world had ever seen. And they saw the problem with the clarity and focus of veteran game developers who'd been hunting that dream for years. Microsoft's ambition for the Xbox console and Xbox Live played a role in convincing the fiercely independent Chicago studio to sell to Microsoft, and once in Redmond, the studio drove Microsoft harder than any other creative team to innovate with Xbox Live.
Halo: Combat Evolved, from the ground up, was designed for multiplayer. It was inspired by multiplayer. It was a product of founder Jason Jones and the original Chicago crew's obsession with Tribes and Quake. Yet through attrition and the chaos of game development, that multiplayer thread lost its way over time, and the Halo team began to focus more and more on the single-player experience, the story of the "Master Chief" and the now legendary space opera fiction that has spawned a dynasty of games and entertainment properties.
Still, for Bungie, the multiplayer experience kept crawling back into minds, and it saw Halo as its chance to get that grand vision of the future off the ground.
"It really wasn't until the very end of the project that we had multiplayer functional on the system link," says Bungie's Chris Butcher, referring to Xbox's pre-Live local multiplayer. "We only had a couple weeks of that working before we ended up shipping it. We kind of knew that we had our hands on something that really captured a lot of the essence of people sitting around with their friends and playing together. A lot of that social community that springs up around multiplayer. That was something that was really fascinating."
Halo proved to be the system seller for Microsoft. It defined the console, both to gamers and the world at large. But it shipped a year before Live would be ready, and by the time Live went out the door in November of 2002, Bungie had already moved on to its next project, Halo 2. With Live subscriptions picking up and Xboxes shipping out the door in droves, Microsoft was faced with the prospect of its premiere online service going at least a holiday season, maybe two, without its premiere online experience.
"The first year of Xbox Live, there was no Halo," says Charchian. He was Bungie's account manager for ATG and Xbox, working closely with the new residents of the Millennium Campus on their upcoming addition to the Xbox Live lineup. "There was a lot of controversy around, 'Why didn't we take the Halo code base, give it to some other team if Bungie doesn't want to do it, and let them make it into a Live product that's ready for the Live launch?'"
"If you go and look at that document, it was essentially a superset of all the multiplayer features for Halo 2, Halo 3, Halo Reach and other games yet to be created in the Halo series."
The response from Bungie: No way. Halo was Bungie's baby. The idea of handing it over to someone else to make a quick port was anathema, and diverting talent from the ambitious Halo 2 project just to port Halo likewise just wasn't going to happen. Bungie might have been bought, but it hadn't sold its soul. The studio insisted on its own building, its own culture and its own rules. There would be no Live-enabled Halo, until Halo 2. Microsoft would simply have to deal with it.
Then-fledgling publisher Ubisoft helped pick up the slack, with Tom Clancy's Splinter Cell: Pandora Tomorrow. A game Multerer credits as one of the first major Live success stories.
"Ubisoft was small. They were the hungry little game developer at the time," says Multerer. "So they went for [Live] ... We were waiting for a third party game to really take it and push it. That was Splinter Cell: Pandora Tomorrow. That game came out and ... It wasn't a shooter. It wasn't like super fast action. It was kind of slow. It was perfect for trying to figure out how to do networking when no one really knew how to do it. You weren't as worried about latency. ... It was quiet, it was dark. And it was just perfect. The game was clearly better because Xbox Live was there. Ubisoft made a bet on us, and it paid off for them, and it paid off for us."
Meanwhile, the Halo 2 team struggled to reconcile its ambition with Xbox Live's capabilities. Bungie co-founder Jason Jones, Butcher and the multiplayer design lead, Max Hoberman, sat down to plan out where they wanted Halo multiplayer to go, from Halo 2 and beyond. Their ideas centered on the vision they shared with Allard of the "virtual couch," where players could find, interact with and play against their friends seamlessly, safely and consistently. Hoberman wrote up the proposal and the three men walked it across the parking lot to Millennium E for the kick-off meeting.
"If you go and look at that document, it was essentially a superset of all the multiplayer features for Halo 2, Halo 3, Halo Reach and other games yet to be created in the Halo series," says Butcher. "Max's pitch to Microsoft was essentially for the ultimate Halo multiplayer experience, with tons more features than could ever possibly fit in one game."
Hoberman remembers Jones telling him and Butcher to "let him do all the talking."
"We went along with this and it was clear from listening to the exchange that we were [Microsoft's] most important customer," says Hoberman, "and that they were open to accommodating our needs, as long as we were clear about those needs."
Those needs included seamless matchmaking, accessible friend lists, clans, in-game messaging, parties, party chat, dispute arbitration, moderation and host migration. Features the Xbox Live engineers hadn't even considered, much less begun. Bungie's designs blew minds at Millennium E and prompted no shortage of consternation. Hoberman's plans for Halo 2 were brilliant, without question, but they would involve taking time away from pet projects the Xbox team had already begun. To make Halo 2 work, they would have to completely rethink the Xbox Live roadmap.
Charchian worked within Microsoft as Bungie's champion. As their ATG representative, it was his job to find the middle ground between what Bungie wanted and what Xbox could deliver. And more often than not, Bungie's needs won out.
J Allard: "Everyone on the project was doing the biggest project/hardest job they'd ever done. The console, accessories, launch lineup were on shelves 19 months and one day from approval. It was pure insanity."
Jeff Henshaw: "The name of the folder that I keep [the original Xbox Live pitch] PowerPoint in is 'Epic Fucking Pitches.' Because this one was an epic fucking pitch. It turned into something big."
James Miller: "I love this place. The smartest people work here. I've been in the same division for more than a decade, which is a rare thing in the tech industry. I'll be here another decade if they let me. It's a magical place to work, and the technology we're working on is harder than anything I've ever seen. The end result is my 12-year-old kid has awesome magical experiences with the stuff I do for a job. There's not a lot of people who get to say that."
Eric "e" Neustadter: "It was a team of people building the thing that they really wanted and no one else was building. I have never worked with a team that is as passionate about the product as this team. Which is why I'm still here, 12 and a half years on Xbox later."
Chris Butcher: "There was one guy who was a writer in Hollywood. He said that he first knew about Bungie because, apparently, all of the people who were writing Hollywood screenplays would get together and play Halo 2 and chat online and just play online and bullshit with each other about the screenplays they were working on through the medium of playing Halo 2 with voice chat. And I was like, 'Really? That's a subculture that existed?' And he said, 'Yeah. For two years back then, it was the place you would go if you wanted to work on your screenplay.' That's amazing. That kind of connection is really satisfying."
Marc Whitten: "[Microsoft] knew that it didn't know how to build game consoles. It was like, 'We're not telling you how to do this. We're putting this team together to go learn how to build this.' We were given a lot of latitude to have a real vision and to have a long runway to go get this right. That was pretty special."
Michael Mott: "The energy was palpable. We would get together as a cross-functional team, everyone from customer support to the development team to the marketing team, business development team, and work through holistically what the experience was going to be, for our partners and for the consumers. And then we built proof points along the way."
Scott Henson: "We had a culture that was unique, but at the same time we weren't so unique that we went and said, 'Yeah, we'll just go reinvent everything.' That would have just been completely crazy. And it wouldn't have allowed us to scale the way we did."
Jeff Pobst: "If there's one thing that Microsoft is totally world class at, it is engineering. That company has the right processes in place to build something well. ... Live was built well. It worked."
Sam Charchian: "Microsoft was full of absolute rock stars who all loved to work on something like Xbox. It was the perfect melding of the right product and the right place and the right time with the right people."
Boyd Multerer: "I would not say that only Microsoft could have made Live, but I will say that it sure made it a lot easier."
"I would go to the Xbox Live team and say, 'Here's what they need. Here's what they want to do,'" says Charchian. "And they'd say, 'We can't do that.'"
The Xbox engineers began making a list of the planned features that would have to be scrapped in favor of Bungie's plan for Halo 2, and it was a big list. Email conversations became debates, which turned slowly into battles. Inevitably Charchian would call in Butcher and Hoberman to have a sit down with Xbox, and the Xbox engineers would be swayed by the scope of Bungie's vision.
"The Halo guys would lay out a really compelling case for the experience they wanted to build and how it would be great on Live and here's what they needed," Charchian says. "We did that kind of meeting probably five or 10 times. Every single time, the Xbox Live team did what Bungie wanted them to do. They always gave in and said, 'We gotta do what the Bungie guys need. They're visionaries on this.'"
Hoberman described Bungie's plans for Halo 2 as not just a benefit to Halo 2, but to every game that would follow. Bungie wasn't just making a game; it was helping to build the platform.
"We even went so far as to reach out to other developers to get their buy-in," says Hoberman, "their confirmation that yes, the proposed service would meet their needs, and yes, if the Live team built it they would use it. We got great traction with this approach."
For Bungie, the ATG connection was more than a lifeline; it was a conduit for a true creative partnership with Microsoft. It was the fuel to launch the Bungie vision of multiplayer into orbit. And it worked both ways. While ATG was helping Bungie make Halo 2, sending problem-solvers to help diagnose the code, Bungie was discovering and diagnosing problems with Xbox Live the Xbox team hadn't noticed, because no one else was testing it as hard as Bungie.
It was a true creative partnership — Bungie helping Xbox build out Live, and Xbox helping Bungie on what both hoped would become their true "killer app." Bungie would frequently work late into the night on Halo 2 and call on ATG to wake someone up to help solve a problem.
"I remember there would be times where we were debugging things at 11:30 at night, and [Charchian] would come over to our offices when we were crunching," says Butcher. "We'd be like, 'Hey, Sam, we're seeing this. I remember that the dev you put us in touch with last time ... do you have [his] home phone number?' ... They'd help us track down these problems that I think couldn't have been tracked down any other way."
By the time Halo 2 finally completed in 2004, Bungie had worked on the game for at least three years and worked overtime, or "crunched" for a third of that.
"It was just insane," says Charchian. "I remember Jason Jones ... I was walking over to visit those guys, and Jason Jones was walking out. He could barely walk. He was staggering like someone who was drunk, running into the walls. He hadn't slept in like two days or something. It was insane levels of crunch to get that game out with all the features that they wanted in it."
Bungie had bitten off more than it could chew and eventually cut two additional games' worth of multiplayer features from Hoberman's original design just to get the game out the door in time. For Butcher, it was equal parts exhilarating and disappointing. What eventually shipped in Halo 2 was more advanced than anything anyone had ever done on consoles, and arguably PCs as well. But for Bungie it was a fraction of what was possible.
"At Bungie, we always try and do way too much," Butcher says. "We were always trying to shoot for the moon and bite off so many things. We get so excited about these experiences and what it's going to be like when these things all come together. ... We'd have this vision for an amazing, integrated Halo multiplayer community experience. And then we'd try and build it. With Halo 2, maybe we got 50 percent of the way there. With Halo 3, maybe we got 80 percent of the way there. Whatever it is. But there would always be these things that were left on the table."
By November of 2004, Halo 2 was as ready as it was going to be. The game was printed and shipped, and everyone at the Millennium Campus held their breath to see what would happen. The Xbox team knew it would be the first major test of the stability of Xbox Live. No other game on the console had the buzz of Halo 2. Demand was outrageous. They planned for an exponential increase in users the very second the game became available.
In the op center at Millennium E, Neustadter manned the controls, ready and waiting should the launch turn catastrophic.
"I was sitting in a conference room in front of all the graphs I could lay my hands on ... monitoring everything, looking for any sign of trouble," Neustadter says.
Once the game was available and players started streaming on, Neustadter monitored the numbers and sent out periodic updates to Bungie and his higher-ups at Microsoft. There were no emergencies. The system worked perfectly. And so everyone turned their attention to numbers. How many players were online? How big was Halo 2 going to be?
Neustadter saw the numbers increasing. And increasing. And increasing.
The previous record for concurrent players on Xbox Live was just over 40,000. In less than 24 hours, Halo 2 had quadrupled it.
"I actually wound up getting a private email from J Allard ... somewhere around 2:30 or 3:00 in the morning," says Neustadter. "It essentially was, 'OK, your mails are really cheery and happy and that's great, but what's really going on?'"
Neustadter assured Allard that his emails were accurate. Halo 2 was crushing Xbox Live's old record highs and the service was stable. They'd done it.
After years of struggle, jealousy and curses, Live's numbers had reached the tipping point. With so many active subscribers, it was inevitable the service would grow. Xbox Live had reached critical mass.
Neustadter sent one final email, then turned out the lights and went home.
From there to here
From the launch of Halo 2, the history of Xbox Live is written in the present. In November of 2013, Microsoft will begin selling its third Xbox console. Live is so much a part of the device that it's hardly ever mentioned. It's merely assumed to be there.
In the years following Live's completion, the idea of an internet-connected console went from outlandish to ordinary. The process of creating it went from incredible to accepted. Nintendo's and Sony's competing consoles have adopted their own versions of Live-like services, and in many respects have closed the gap, and new and more online entertainment services have sprung up elsewhere. As a result, the controversy and wonder of the effort at Millennium E has been woven into the fabric of the everyday so completely we can no longer imagine a console (or TV or Blu-ray player) that is not online. What's extraordinary today is when they are not.
Xbox Live itself has been revamped and improved over the years. New features have rolled out alongside new games that demanded them. Much of the original hardware and software architecture has been replaced times over. What started under a handful of desks managed by a half a dozen engineers has expanded over the years to become one of the preeminent online services in the world.
In 2000, Xbox Live launched with 300 servers. In November of 2013, the modern version of Live will support Xbox One consoles with more than 300,000 servers.
Today, Xbox Live delivers more than games and voice chat; it delivers movies and television on demand, music, cloud data and more. The original plans, drawn up in 1999, contained much of what's still being added to this day. Plans like Project Orange, the 2008 "New Xbox Experience" and more have become technologically possible only since Live's 2002 launch. The Xbox Live groundwork laid by Allard and his team has allowed Microsoft and others to dream even bigger dreams, standing on the shoulders of the geniuses at Millennium E. And the full vision of those early days is still yet to be realized.
In the meantime, Xbox Live has proved so robust it has been co-opted by other divisions within Microsoft. It now serves more than just the Xboxes, Xbox 360s and (soon) Xbox Ones, but also smartphones, Surface tablets, sync-enabled cars and PCs running Microsoft media software. It has gone beyond gaming to enter the public consciousness. It has become the entertainment platform Allard envisioned, and it has united people all over the world in ways even he could not have expected.
The birth of Xbox Live was a near perfect intersection of technology, opportunity and passion. These men and women brought to life a vision from science fiction. They did it against difficult odds and the objections of their contemporaries. They did it by striving to do something that hadn't been done — and that many believed couldn't be done. And many of them are still at Microsoft to this day, working on whatever wonder will come next.
Illustrations: Dylan Lathrop
Editing: Matt Leone
Design / Layout: Warren Schultheis, Matthew Sullivan