In part one of this interview, David Perry talked about his latest project, Gaikai, and how the streaming games service was planned to scale.
Here he sheds more light on just how the server infrastructure will work, how quickly the company will move to satisfy demand, and why he gets involved in so many projects...
Then our price goes up. I've already been in discussions with companies that can build our hardware for us and scale almost immediately. They'll have servers ready to go. We have two choices - either they build everything, or we build everything, and I kinda like that they build everything
We're just trying to work out what that cost will be, but because it's a service to me I can then say I need another thousand servers and they can take care of that problem.
But put it this way - it's like a wet dream for investors. To come to them with a problem like that - we can't scale fast enough - trust me, I can line them up. If I go to an investor with that problem, they'll help me solve it.
Well, it varies. I can go from my house to Las Vegas in 17ms, but doing that distance again is 93ms. That means that there's no connection between Las Vegas and the point further on, so it must go around in a circle - it's the only way you could cause such a huge spike.
So that's the real internet - it's a messy beast. What we did was look at a map of the internet and saw all of the hotspots, where the numbers were getting too big, and just filled those in. In the UK we'll effectively reverse engineer.
The second part is that we don't actually know which server will be the best for you - if we have a server in Brighton and you're on the other side of Brighton, that might not be the best server for you because it might be going in a loop to get to you.
So we take, geographically, all the servers around you that we have and you ping them all. That's what the load time is, if you saw the Gaikai demonstration, the text coming up in the corner is it pinging all the servers around to see what the best one is. It then logs on to that one, creates an instance, and the game starts to play. That means that you're guaranteed, whatever your situation is, to get the best server.
Secondly, we choose the compression based on how fast your processor is and how much bandwidth you have.
Well if it was no good people are just going to keep bailing on it, and we wouldn't let that happen. If we saw people bailing we'd ask what was going on, and see that they're getting a 250ms ping time and stop it.
GameInvestors - the concept was to try to make it so that you can get your team out there, and for the rest of the world to be able to find you and know that's what you do. Let's take iPhone developers - how do I find the available ones? There are lots of them, and if I start calling them all there'll be a lot of wasted calls.
I'd rather have a list of all the ones that are available right now that aren't under contract that are looking for work, be able to find out what they've done, and not have to talk to anybody. I want to review all the games they've done, read their bios, I want to contact them to say I want to read more, and they can let me see more because I'm legitimate - so they'll unlock a whole bunch more content, history, and so on.
And then I finally decide I want to talk to those guys - it'll same them time, it saves us time, I found them. I can choose from everyone. Maybe it's a fantasy of mine, but that fantasy is that I don't have to spend the next month trying to find these people - they've just there.
And when a new team surfaces, perhaps a team that spins off from another company, they're available right away to me - I know about it the next morning. They've just added themselves to our database.
It's a bit of a game-changer in the way that we access talent, and it will empower publishers to reach out to talent. And investors that are interested in this space will be able to get in much easier, see all the opportunities that are available right now - I'm not telling you they're going to want to fund them, but they'll certainly be able to reach out.
There might be an amazing team in Brighton right now, and we just don't know.