Thoughts OnLive, or the One Where I Try to be Topical
When I first heard of OnLive last night, I did not think much of it. As I started going through the news stories and getting the impressions from industry heads (I could aim a low-blow to Denis Dyack here, but I will not), I admit the Kool-Aid started looking a bit tastier. Steve Perlman has done a damn good job selling his product.
After some time to mull over this new tech, my thoughts are this: if everything went right for OnLive, it would mean a true paradigm shift in this industry. The reality, however, is that this sweet tune is probably going to go a bit off-key when it reaches consumers.
Let us start with what it is. OnLive is a digital-distribution service ready to go into Beta this summer, with a launch date set in Winter of this year. It was announced yesterday at the Game Developer’s Conference (GDC), and the company describes the service as using “Cloud Computing” to allow gamers to play the most demanding games on relatively low-end machines, or even television sets, through the use of their “mini-console”.
The mechanics of this service are relatively simple. The service needs at least a 2Mbps (Megabit-per-second) internet connection — meaning that most residential connections will qualify — and works by taking the user’s input, either via keyboard and mouse, or gamepad, and returning a real-time video feed of your gameplay.
Perlman has said that this project has been in a think-tank for seven years, and I imagine they must have made some pretty substantial breakthroughs, because this model is extremely ambitious. The term “cloud computing” is not a new one, unless Perlman is working on some obscure interpretation of it. The concept of distributed processes has been around for years in computer science, and has been put into practice a number of times, such as with Folding@home.
Essentially, it involves breaking large jobs with a high number of computations — computer game graphics certainly qualify — into smaller chunks, and assigning the work to a number of computers on a network. Here, OnLive plans to do all of its computations off-site, only returning audio and video to the client.
Think about it from a logistical standpoint. If, say, the same number of gamers are playing Left4Dead during peak hours as before, except they are all using OnLive’s service. Now, instead of millions of gamers each with their own mid- to high-performance PC, now we are looking at either a few less million top-shelf gaming rigs running in server farms, or a few more million mid-range PC’s running distributed workloads.
Either way, OnLive is looking at substantial start-up costs, not to mention big money every time the PR teams at nVidia and ATI so much as cough. Even if OnLive charges the same amount of money for their games as the retailers, subscribers are still looking at a hefty annual fee, likely much more so than Xbox Live’s service.
And the pricetag is an important factor here. OnLive is marketing towards gamers who cannot afford top-of-the-line gaming rigs, but can still pay for the games. Even on the beta machines demoed at GDC, with nearly no traffic on OnLive’s servers, journalists noticed blurring and other signs of the video compression taking place. Therefore, the more OnLive charges, the larger the portion of their potential customer base that will slip into the category of consumers willing to pay the difference for a perfect picture.
The good news about OnLive is that a MacBook user can run Hertz-hogs like Crysis at near-max graphical settings. The bad news is that your game, with all its proprietary code and its beautiful art objects is on someone else’s machine.
Let me go ahead and tackle the biggest issue right now. SecuRom cost EA a few dumptruck-loads of money. Why? Because it gave gamers the sense that they were not really buying copies of Spore and Mass Effect. Oh, no. It was more like they were just letting gamers borrow their new toy. As long as they played with it in the corner. Where EA could see it. And god forbid if the gamers let three others play with it.
What companies like Steam and GameTap do not understand (or just ignore) is that a large portion of gamers still prefer hard copies to electronic. SecuRom shook their concept of a “hard” copy, and OnLive would deny the option entirely. Sure, I will pay $10 for Rainbow Six Vegas 2 on Steam when it is still $30 in the stores, but if the price was the same, or even within five bucks, I would drive the fifteen minutes to the nearest Best Buy or Gamestop instead of waiting on a download every time. Yes, it is mostly psychological, but it is a powerful deciding factor that retailers have been (rightly) counting on for years.
Consider also, that while console gamers are used to a product that fits on a disc, and, in their minds, is wholly contained on that disc, PC gamers know that the products they buy are not just neat contained games, but a finished product, and most of the tools used to make it. If Half-Life had been a console-specific title, then would Counter-Strike have ever existed?
Even if OnLive develops a way to distribute major Mods for big PC titles, the best, strongest, or strangest ones will never make it, because OnLive cannot or will not recognize them. So what if Star Wars has nothing to do with Morrowind. Some modder put Darth Maul in Vivec, and a bunch of people enjoyed it. By outsourcing your game, you give up all rights to play it your way.
I have not mentioned the issue that most consider first when they hear about OnLive, but OnLive’s ability to handle massive amounts of internet traffic will certainly be tested from day one. Anyone who has played an online game from a regular residence (not a university or business) knows that internet connections rarely operate at their specified top speed, and that is not even the fault of the server host. Every online gamer has learned to play with lag, but how does that work when the game itself is remote?
If data sent from the clients (OnLive subscribers) gets lost or delayed enroute to OnLive’s servers, then most games would probably interpret it as a lack of input. Basically, your character will stand still, while four different snipers draw a bead. On the other side, if clients cannot pull down the image data fast enough, the picture will start to pixellate, costing gamers precious detail.
Again, online gamers may learn to live with that, but what about single-player games? Imagine playing Resident Evil 5 and trying to keep up with the quick-time events when half of the response window is lost to lag. Rather, imagine cursing your ISP while you watch Sheva take a motorcycle tire to the face for the sixth time.
One of the most enticing bits from the Perlman interview I saw linked on Kotaku (URL below) dealt with how OnLive would be, essentially, a TV-streaming service, where each user had their own interactive channel. With the signal in this state, it would be nothing to send this signal to other users, as well. You could bring up a list and see what your friends were playing at any given time, and heckle them at every wrong turn. You could watch the player at the top of a leaderboard, and learn from his moves. You could watch someone play a game you do not even own, because, hey, the code is not on his machine, either (yay!). Hell, you could even make your friend play through all the boring “gameplay” in a Kojima game while you pop popcorn for the next mov- err, cutscene.
Again, stewing on this after doing the research, I realized that this is not really above current systems, though. Games like Command & Conquer 3 already have spectator modes. Why not just take the feeds and move them out of the game? It would take some effort on developers’ parts to turn the output into video feeds instead of game data, but if the demand was there, they would probably bite the bullet.
OnLive smacks of potential. If the service could keep costs down, it could revitalize the PC gaming industry and put it leagues ahead of the console market, like it used to be. It would be a developer’s dream come true. No more piracy, and the guarantee that every gamer in the audience would be running the latest hardware. But for us gamers, few things are ever as good as the bullet points on the box, and reading between the lines is vital. With regards to OnLive, I read high costs and maddening lag.
Leave a Reply