Massively

Massively: So tell us a bit about your background and how you got into this line of work.

Dr. Ed Ikeguchi: I spent a lot of time in business, and after a while had actually done a lot of software development in creating large enterprise-wide solutions for drug companies doing drug research. After about ten years doing that, the company grew to about 800 people in size and we sent it off on its way, we had a happy IPO. I decided to take a few months off and get to know my kids again, and lo and behold here I am completely immersed in this gaming world because that's what they love to do. So I found myself kind of really addicted to it. I have to say, one of the things that really caught my attention just from a pure interest perspective was this whole social interactivity. That was really appealing to me, in MMOs nowadays. I started with World of Warcraft, and did a lot of that with my kids, I was in a guild, [and am] exploring more of Aion now.

But really, one of the things that was really interesting to me was this whole phenomenon around Vent. Because it's kind of what holds everybody together; it's the glue. Everybody logs into these Vent servers, and as a business guy looking at this it was interesting because it was so fractionated. You don't buy Vent from Ventrilo, you buy it from any of number of tons and tons of these hosting companies all over the place, and it just seemed crazy that it was so fragmented like that. Because if you have a set of user requirements that you want to build in for a new MMO coming out, or new ways to set up a team or an arena, or if you want to talk in different ways, or want to use your iPhone, all these kinds of things.

"I started thinking it'd be a lot easier to constantly churn out new features and distribute them, because it's all web based."

It's just my experience from software development. It's just really hard to create that kind of constant churning in terms of software development, making things improved for a specific user group when you have a business model that distributes through all of these hosting companies. You don't actually license directly from the vendor, and worse yet, if you have upgrades, you have to go to the website and download something. You have to look at numerous compatibilities. I started thinking maybe this is a market where it would be an ideal situation to try and port some of this functionality onto the web, because after all, we're still just using internet to connect, it's just through this heavy client product. But what it does, is that if we can get this thing onto the web, I started thinking it'd be a lot easier to constantly churn out new features and distribute them, because it's all web-based.

So that's kind of where we began, just shy of a year ago. I ran into someone who's become a good friend -- his name is Dean Elwood -- a guy who's really experienced in the voice-over IP field -- he's worked with Skype, everything from handheld radios on Motorolas to Facebook apps on VoIP -- and we started talking about what gamers want on systems. [We] hit upon things like very, very good sound quality, but we also talked about things like eliminating lag (vent lag is torture), why people don't actually use in-game systems for voice. You know World of Warcraft has voice, but nobody uses it because the sound quality isn't that good, and you can only talk amongst people that you're partied up with, so how do you deal with a guild that's not going to be partied up all the time just to talk to each other? These are the reasons that people are driven to things like Ventrilo, but at the same time Ventrilo has its issues around, you know, like vent leeches. Last night's raid we had to move to another sever because some guy was constantly heckling us, and we had vent lag.

These are the kinds of things we set out to improve upon but do so in a way that was web based because it would ultimately allow us two things. One, it would allow us to constantly create an improving system -- something we could churn out pretty rapidly and distribute to our entire userbase the moment it was uploaded. The second thing we wanted to do was from a support perspective. We wanted to make the thing as lean to sustain and support as possible, because our initial vision in terms of how we'd distribute this is that it would be free for users to talk. Whereas systems like Ventrilo, they kind of peak out at 250-300 users; the server just sort of fizzles. We literally couldn't support millions of users for free if we were having to provision a new server for every 250 users. It would not be economical, so we had to create something that was massively scalable. That's what we aimed to do and that's what we achieved, with the help of some real pros -- somebody who really knows VoIP extremely well.

Right now the Blabbelon system we created is web-based. We worked very hard with the people over at Skype to see if we could incorporate their voice codec. They were actually really impressed that we were able to get their thing on the web! We're talking over Skype now, but it's a heavy client that we've downloaded sometime previously, so they were really impressed that we were able to get their codec onto the web. The SILK codec, in my opinion, it's an engineering wonder because it provides better sound quality, a deeper depth of voice quality while actually using less internet. So if you're on a really robust system, it's probably going to be negligible in terms of what you see in game, but every little bit counts. If you want to avoid lag, either on the voice side or definitely on the game side, you're basically achieving with SILK the highest quality of voice while sacrificing the least amount of internet bandwidth. I think that was, from an engineering perspective, something that we're really proud of.

The other thing that we have achieved is this nice scalability. We've tested the system on one server and it'll run comfortably 7,000 simultaneous users. It'll probably go higher than that, but you're never going to have 10,000 guildies on at the same time. What that means is that from our perspective, with basically just a handful of machines, we can support a whole lot of people and do it at a cost basis that makes sense to us to offer it for free.

So that's kind of where we are with the product. We're launching next week and we've got some features in there that I think are already at least on a par with Vent. Because the sound quality is there, the ability to create rooms is there, the ability to create nested rooms or sub rooms is there.

All those basic things that you'd want to have that are core for a game voice system are there, and then we've added to that some neat features where if you are in a group or in a raid, and you have somebody who's a raid leader and -- it's just irritating when you have the whole peanut gallery saying stuff at the same time -- we've created a ranking system within the channel so that people who have higher level ranks, when they hit the push to talk, it overrides other people. These kinds of things are in the system now and we're constantly looking for new features. Somebody came up to me the other day and said "Hey, you know it'd be really great if inside the channel you could just allow the officers to talk among themselves by hitting a different keystroke." These are the kinds of things that we can build very quickly and implement and upload them because it's totally web based, and literally at that moment we've distributed it to our whole user base. We're really pretty excited about it, and have run a couple raids on it, actually gotten our guild on it, and that's kind of where we are.

This sounds great, because I often find myself really frustrated with a lot of the voice chat out there. Recently I've been playing a lot of Dungeons and Dragons Online, and they have integrated voice chat, but there's that problem where all of the voices are different volumes. Can you individually adjust volume?

We don't. Actually, what we do on the server side, we normalize everyone's voice. And what we do is get it to a point where it comes out so that if it's really loud for a particular user, everyone's voice is loud, and we allow them on their end to put the volume down on their PC. But in terms of the volume of voice relative to other people, it's something that we try to take the complexity out of things, because you get into this incoming/outgoing, and you do it per user, and it gets really complicated. It might be something that we wind up coming back to, but for the time being what we really wanted to do was try and get that into the technology's hands to solve, and on the server side what we do is normalize everyone's volume

Now you mentioned an elimination of the lag that you find on Vent, and that's another thing that's probably one of my top frustrations with Vent. Without getting too technical, how does that work? Is it because it's web based?

I'm going to get a tiny bit technical here. The first thing is, the codec which delivers to you the quality of voice you want, is what determines how much internet it's going to consume. Obviously, the higher quality of sound you want, the more internet it's going to consume, because it's transferring more voice data over the web, and that's competing with your game. I think I'm coming through to you on Silk right now, because I'm using the latest MAC version of Skype. The amount of internet that my codec is using to transfer my voice to you is a lot lower through Skype than just using Speex on Ventrilo. Speex is the open source codec that's used in Ventrilo right now. That's one thing.

The other thing is this: On the server side, what Ventrilo does, is if I talk and someone else talks, Ventrilo takes both of our voices and on the server side, does a ton of floating point math to integrate our two voices together and then spits it back out for the rest of the users to hear. What happens there is that it's an enormous volume of work on the CPU to get that done, so part of the problem is coming from the amount of internet consumption, but part of the problem is actually coming from the server side where the server is having to do a lot of work to integrate people's voices together. If you look at what we do, we actually don't do that. It's actually something that is pretty unique, that we filed a patent on: we don't do any audio blending on the server side. The server, all it does is it takes your voice in and sends it out to everybody else. We do have people who of course can talk simultaneously, but that blending occurs on the client side, meaning it happens at the level of your computer. So we take all the load off of the server, which is just working hard to take in voices and send them back out again, and we free it of all that necessary work it would have to do, to do blending. That allows us to do this kind of scaling.

It's not just Ventrilo, actually. If you look at almost any of these modern digital conferencing systems, they all peak out -- even Skype -- at about 250-300 users. 300 users if you've got really really smart engineers, because the server comes to a grinding halt. It can't mix that many voices together at the same time and it does so with a very arduous floating point math. So what we've managed to do is take all that load off the server and bring it onto the clients machine, and it gives us the scalability we need to make this feasible from a business perspective. We can literally scale to 7,000 users, we can probably scale higher. I have a feeling we probably could scale it to 10,000, but we've only tested it to 7,000 users.

One thing I'm curious about, you said this was all going to be free. Will there be any kind of payment for better "VIP" options, or anything like that in the future?

Nope, it's free. When I say free, it's free. But yes, of course, we have a business model that we are working through certainly for 2010 and this is our vision. The assumption is that gamers are real people too. We're not all, like, basement dwellers. Somewhere along the line you have a family, somewhere along the line you have a girlfriend or a boyfriend. A lot of people actually use Vent to talk to their real life contacts, and that's fine, and you can do that through Blabbelon as well. But what we're thinking about is eventually we should be able to then take our user base and later on, more value added features leading down to things like how do we take a voice service and add to that the ability to share content.

"You look at the iPhone app store and all that success, and wouldn't it be interesting to put that type of app environment and encase it in a voice system?"

For us, everyone's always sharing YouTube videos on how to kill the next boss. So you should be able to share that within the interface of Blabbelon. And then you think to yourself okay, if you can share a YouTube video, why not share a PowerPoint presentation, or a stack of photos you took over the weekend? Then if you can do that, then why can't you share a Word document to use in business or in a school. and then ultimately it comes to, why wouldn't you be able to share a game? On that last point, that's ultimately where we would love to go with this, which is "yeah we know gamers need to talk", but gamers are very social, a la MMOs, and if they're going to socialize in the game. There are potentially other things like, you look at the iPhone app store and all that success, and wouldn't it be interesting to put that type of app environment and encase it in a voice system? So go to the app store on Blabbelon for purchase and pick up, let's say a casual game -- let's just call it tic-tac- toe -- and amongst the people that you have in your channel, either play it and let other people watch so that they can give you a hard time about the crappy move you just made, or share that game across your screen and their screen so they can compete with you. However, that's pretty far down the line -- 2010 stuff. For now and for the foreseeable future, at least talking, doing what we do with Vent, we really strongly believe that it's for free.

I don't know if you find that it's the same group that you play with, from one MMO to another, but it's the kind of thing where we also think some of these relationships that you start to forge in one game eventually start to bleed out into other things as well. It's the kind of thing where, do you have to get another Vent server because you're going to go play Guild Wars instead of Lord of the Rings Online? What we're trying to do is give ownership of the voice system to the individual and not the guild or not the clan. In other words, it's yours. You set it up, if someone invites you to a channel because that's what they use for the guild, that's fine, but we give you a whole set of tools to create new channels for everything else that you do.

Thanks Ed for your time, and we look forward to seeing what Blabbelon has to offer after launch.