Remote game services (OnLive, Gaikai, etc.)

[Chandler55]

uh its the march 25 one, 38 minute mark

also i really recommend this podcast, its very smart, and the chris remo guy knows what hes talking about a lot, and theyre nice people

 

[Rangers]

I just thought of another anti against this service..piracy.

Theoretically this is a subscription service, difficult or impossible to pirate just like WoW.

Well, I'm quite certain many PC Gamers would rather go out and build a 1k rig, knowing theyre then going to pirate and get all their games free, than paying for this service.

 

[Rangers]

A developer on listenup podcast I just heard spoke on this, (actually it was Pete Wanat, who made controversial Wii comments earlier in the week) he seemed quite bullish on the benefits in theory if this worked (personally I dont think it will work). Being:

-Deving is hard to profit on when theres few hardware units in the wild when you're crossing over to a new generation of consoles. Well, Onlive would have ridiculous numbers of hardware units in the wild (theoretically virtually any low end PC).

-No physical production costs, costs to design box art and all that, no lead time either (he mentioned six weeks from gold master to retail). He acted like production costs were very expensive.

-No used sales

-No "fund" necessary for returned games (inventory that doesnt sell) from retail

The latter three of those would be benefits in any digital distribution system though..

 

[obonicus]

I just thought of another anti against this service..piracy.

Theoretically this is a subscription service, difficult or impossible to pirate just like WoW.

Well, I'm quite certain many PC Gamers would rather go out and build a 1k rig, knowing theyre then going to pirate and get all their games free, then paying for this service.

Well, that plus removing the used market plus removing all the costs associated with distributing media are why publishers like this.

 

[Richard]

I think you guys are overreacting on the compression quality... it's free AA people!

Seriously, has anyone thought about the business model here? To be profitable they'll have to go with a subscription model. I.e. they must have at least one high end PC for each x number of users - like ISP contention rates - which means every handful of new users will require more investment. The only way to have a return on this is through a pay-to-play subscription where they hedge their bets on low concurrent service utilization - once again like ISPs.

Let's side-step the ugly debate about pay-to-own versus pay-to-play and get right down to it: this forces them to license the games from the devs directly otherwise they could be sued for what's effectively renting out a game (this doesn't apply to all markets of course - it does here). If you have devs unwilling to license this you're breaking down the market base into two: remote play and local play. If there are games that are only available in local play, people will still need to buy a console/gaming pc which begs the question: how worth it is this? Even if all devs are okay with this there maybe games where the compression artefacts are so noticeable everyone would complain (racing games come to mind) so you'd still have the split.

Another question related to my "end-of-cheating" remark about this business model is: where do mods fit in? Because of the update-cycle, it's in the PC space where this service is most useful, however, the PC is probably the platform that has more corner cases. Here's another one: the various monitor resolutions and especially the current standard today being >720p.

Here's where I think this could work: high-piracy, low-income markets like asia, south america, etc. where people could buy a cheap set-up box and pay a small subscription and yet have good graphics, with the major problem being bandwidth. Even with the current economic climate a PS3 is _cheap_ when you consider the 5-7 year console cycle. Incidentally, the longer you delay your purchase to take advantage of a potential price-drop the less worthy it will be in the end. However, in these markets the upfront cost of consoles/gaming PCs are simply out of reach so OnLive could provide a definite alternative --- if not for the bandwidth. How about an OnLive service for Wii games instead?

 

[Scott_Arm]

Well, I would think that they wouldn't have to license the games because it'll be a platform like Steam. The devs will probably sell their games or rent their games through the service and Onlive will take a cut.

 

[Shifty Geezer]

Millions watch Youtube Music videos, and listen to shitty sound quality everyday, they use it for playlists at work, and parties were youtube is running is common place today (helps with the shitty quality).

That's not a straight comparison though. The YouTube vids are a choice between paying for the vid or not. I've watched low grade YouTube vids because I wouldn't pay to watch those particular vids although they are entertaining enough to watch in grotto quality. If people were charged a subscription for YouTube comparable to the cost of buying/renting a DVD, would it be as popular? Of course, YT has the advantage of having ridiculous amounts of content nigh impossible to find anywhere else. It's really not a comparable service to streaming games replacing bought games+hardware.

As for music, the difference between compressed audio qualitywise and these compressed vids is again an order of magnitude or three apart. As long as the music sounds about the same to the listener, it's good. The details don't make or break it. Whereas visually we're very attuned to detail. No-one is happy looking at blurred text like they are happy to listen to 'blurred' music. It's tolerated on SDTV when there's no alternative, but the increase in HDTV shows some people like higher resolution content. And the increase in PC hardware shows that too, with people choosing to render at crazy resolutions instaad of lower resolutions (1024x768) with smoother framerates and AA etc.

I think we really need to appreciate there are different markets being talked about here. One is the existing gamer used to high quality stuff, especially high-end PC rig gamers, wanting to switch to OnLive as an alternative content distribution system, perhaps because all the publishers prefer it. I can't see them ever being happy. I'd much rather take PS3 at it's output than any top-end PC rig splodged down to a 5mb connection. The reason I bought an HDTV was so I didn't have to view graphics like that! Then there's those people who want to play games and have no prior reference point, so that 'Wii looks really good' blown up on their 46" LCD, and perhaps the conveneince of games on demand will appeal to them. And of course, there are those who want to play highend games and can't afford to, for who a low cost of entry is an acceptible sacrifice. I expect some of them will be willing to take the hit in IQ (assuming no revolutionary compression model being used here, which I don't rule out completely but if true, must be thinking so far outside the box the original box isn't even visible!)

But whether these markets are large enough and work well enough with existing infrastructures (not in the UK I bet!) to be relevant now, I seriously doubt. This 2007 report lists most countries at under 5mb average attained. OnLive can only exist IMO as a cable-service paralleled with TV broadcasts. I can't see people buying a thin-client console, or any such system, that works over a standard broadband connection, and surely those with high-performance connections already have invested in high-end gear on average?

 

[AzBat]

Here's a video from PC World. Looks slick to me...

http://www.pcworld.com/article/1621...action_and_xbox_games_on_your_cell_phone.html

Tommy McClain

 

[obonicus]

Possibly, but it's a tiny video of a tv playing the game. Those of you in the US should sign up for the beta. I'm extremely skeptical, but even if it's the second-coming, I probably won't get to experience it.

Off-topic, but did he really suggest that those 3d glasses need a game to render at 120FPS?

 

[Richard]

Well, I would think that they wouldn't have to license the games because it'll be a platform like Steam. The devs will probably sell their games or rent their games through the service and Onlive will take a cut.

Except that unlike Steam, OnLive won't be selling them if they are forced to go pay-to-play. Some devs may want to jump at the opportunity, WoW players are used to the experience, but I'm guessing there'd be quite a few bumps in the road for the business model to gain traction.

Off-topic, but did he really suggest that those 3d glasses need a game to render at 120FPS?

If this is similar to existing techs you need 120fps to get 60fps stereoscopic vision (60fps for each eye).

 

[obonicus]

If this is similar to existing techs you need 120fps to get 60fps stereoscopic vision (60fps for each eye).

Oh, I know. I'm just wondering if he's really expecting games to render at 120FPS on console (since it seems to be a Sony thing). Most struggle to reach 30.

 

[Jedi2016]

I just don't see how something like this can be cost-effective. I'm not familiar with the "cloud" computing thing, but it seems like they would have to have an enormous stock of spectacular hardware server-side to play these games. I've heard the detractors say that OnLive would basically have to have a dedicated "PC" (physical or virtual) for each and every subscriber.. I mean, you can't have a single server rack actively playing a thousand instances of Crysis simultaneously.

I could see something like this working on a small scale.. remote computing from a device in your living room to your computer in another room, but that's about it.

And at what point does it stop being cost-effective for the gamer? The OnLive device itself will have a cost, of course.. it might be cheap, a hundred bucks or so, or it may cost the same as a PS3. And what about the subscription fees? And the price of the full game will probably be comparable to today's game prices.. $40-60 each. At first it might be cheaper than owning your own console or gaming PC, but for how long? Eventually, it'll end up being cheaper to just buy an Xbox and have physical copies of the games. And, of course, the question of what happens if it doesn't take off. Sure, it might work out well at first, but what if they don't get the numbers they want? The service goes down, the box becomes useless, and all of your games disappear in a puff of money.

And that's not even going into the latency issues and visual/audio quality. All of the tests we've seen so far are running, for all intents and purposes, locally. That GDC conference was probably running on a dedicated T1 line with a direct connection back to their studios, and not over a standard ISP-provided internet connection like real people have.

Of course, they could try marketing this to the more casual gamer.. the one who isn't going to go buy their own console or gaming rig. But on the flip side, a casual gamer is far less likely to have the necessary internet connection.

Count me out. I'll stick with my own PC, where I can play whatever I want at full resolution with zero compression and zero latency.

 

[Scott_Arm]

Except that unlike Steam, OnLive won't be selling them if they are forced to go pay-to-play. Some devs may want to jump at the opportunity, WoW players are used to the experience, but I'm guessing there'd be quite a few bumps in the road for the business model to gain traction.

In the demo is showed that there were options to purchase or rent the games. They said there will be a variety of payment schemes available. I think they'll stay away from licensing content.

 

[Rangers]

I just don't see how something like this can be cost-effective. I'm not familiar with the "cloud" computing thing, but it seems like they would have to have an enormous stock of spectacular hardware server-side to play these games. I've heard the detractors say that OnLive would basically have to have a dedicated "PC" (physical or virtual) for each and every subscriber.. I mean, you can't have a single server rack actively playing a thousand instances of Crysis simultaneously.

I could see something like this working on a small scale.. remote computing from a device in your living room to your computer in another room, but that's about it.

And at what point does it stop being cost-effective for the gamer? The OnLive device itself will have a cost, of course.. it might be cheap, a hundred bucks or so, or it may cost the same as a PS3. And what about the subscription fees? And the price of the full game will probably be comparable to today's game prices.. $40-60 each. At first it might be cheaper than owning your own console or gaming PC, but for how long? Eventually, it'll end up being cheaper to just buy an Xbox and have physical copies of the games. And, of course, the question of what happens if it doesn't take off. Sure, it might work out well at first, but what if they don't get the numbers they want? The service goes down, the box becomes useless, and all of your games disappear in a puff of money.

And that's not even going into the latency issues and visual/audio quality. All of the tests we've seen so far are running, for all intents and purposes, locally. That GDC conference was probably running on a dedicated T1 line with a direct connection back to their studios, and not over a standard ISP-provided internet connection like real people have.

Of course, they could try marketing this to the more casual gamer.. the one who isn't going to go buy their own console or gaming rig. But on the flip side, a casual gamer is far less likely to have the necessary internet connection.

Count me out. I'll stick with my own PC, where I can play whatever I want at full resolution with zero compression and zero latency.

Yeah hardware cost will be one of the biggest sticking points with this. How are you going to allow for a ton of gamers coming on at "prime time" (say 7-11PM) while other times (early morning hours etc) will be extremely slow? You will have to overstock hardware. Not even to mention as has been mentioned when a big new game everybody wants to play comes out.

I almost get the idea this would be better off as a console service. Serving $199 Xbox 360 Arcades are a whole lot more feasible than high end PC's. The problem then is, it's just too easy/cheap to buy the 199 Arcade for yourself, which would quickly become more cost effective depending on the monthly fee, and why bother with the "service" at all? So it's a catch 22.

 

[Nano]

I've heard the detractors say that OnLive would basically have to have a dedicated "PC" (physical or virtual) for each and every subscriber.. I mean, you can't have a single server rack actively playing a thousand instances of Crysis simultaneously.

Could be a concept server set up perhaps, a new approach to putting memory, GPUs, CPUs onto some form of board, cutting out unnecessary components and allowing a phenomenal number of units. You won't need a compact or distinct physical product seeing as its not going into a home or office. It will be very interesting to see what they have in terms of h/w in the background and what it looks like. I'm just starting to imagine the scope of such a big project and how clever they'll have to be with this technically...

Yeah hardware cost will be one of the biggest sticking points with this. How are you going to allow for a ton of gamers coming on at "prime time" (say 7-11PM) while other times (early morning hours etc) will be extremely slow? You will have to overstock hardware. Not even to mention as has been mentioned when a big new game everybody wants to play comes out.

I dont know to what extent, but I'd presume it will be a big help cost wise that they plan to have much less powerful servers for less sophisticated games. Cost saving for playing a title like Lego Batman, compared to Crysis would be big, then again how do they decide the sheer capacity they need of lower end vs higher end units?

I also wonder if they're planning to outsource a lot of the capacity for this, brings in so many questions. What is the sheer scale of the project, who are the big stakeholders, what is the cost of such an amibition...

 

[Brad Grenz]

The business model will assume that at any given time only a small fraction of your paying customers will actually be playing. They may also be able to sell spare computing cycles as a distributed computing service for off-peak times. They are helped in the upgrade cycle in that once you have a system fast enough to play any given game it will always be fast enough to play that game. Your CoD4 servers can continue to serve that game for years. You only need faster systems when a next gen game comes along that your current builds are inadequate for, but the old systems will always be useful for older games.

 

[hoho]

Could be a concept server set up perhaps, a new approach to putting memory, GPUs, CPUs onto some form of board, cutting out unnecessary components and allowing a phenomenal number of units.

Problem with that is that machine will still have bottlenecks that you can't overcome without heavily modifying OS and/or games. I also have hard time believing creating such custom HW would be much cheaper than just off-the-shelf boxes.

 

[grandmaster]

I think the bottom line is that you have to look at this whole 'cloud' thing as something far more competitive than a console race, and the stakes are unimaginably high.

That means it boils down to three things:

- getting your 'micro console' or client-side code into as many places as possible (eg Flash vs Sliverlight)
- cramming as much power into your servers as possible, more than the other guys are doing
- getting as much content as possible from all fields - applications, games, movies etc

Today's game servers will be tomorrow's "hire Expensive App X for a couple of hours" server, so I think you need to take a long term look at the cost structure.

You reckon Sony and Microsoft lost a lot of money on PS360? You ain't seen nothing yet.

Although there's still tons I don't buy about OnLive - and we appear to be getting salesmen talking to us rather than engineers - I'm buying into the Space Race concept of the ability to solve any problem by throwing enough money at it.

 

[Richard]

I think the bottom line is that you have to look at this whole 'cloud' thing as something far more competitive than a console race, and the stakes are unimaginably high.

<snip>

- getting as much content as possible from all fields - applications, games, movies etc

<snip>

I'm buying into the Space Race concept of the ability to solve any problem by throwing enough money at it.

This is an excellent point. If they setup huge server farms and gaming doesn't work out they can switch to rendering farms, or Server VM farms, or... It definitely isn't like launching a new console and seeing fall flat (Jaguar, 3DO, CD-I, CD32, Saturn, etc.) and you just dumped a whole lot of money into the bin. Maybe that's the basis for all the hand-waving we're seeing for now.

 

[grandmaster]

Interesting interview with Crytek:

http://www.gamesindustry.biz/articles/crytek-streaming-games-service-viable-in-2013