I haven't read the whole thread yet, but the biggest problems I have with this are the fact that there is going to be input lag on the system. Then the second is the fact that this is going to eat bandwidth like candy, and when more and more ISP's are capping bandwidth, this can only be a bad thing.
I have a 60GB soft cap on my internet. if I pass 60GB, they dont' cut me off, and doing it once in a while is fine. But if I pass 60GB say, 3 months in a row, I start getting nasty phone calls, risk overage charges or getting my account turned off. if you take the the fact that they say they require a 5Mbps connection for this service, that means that I can game for just under an hour day in a 30 day month, and I'll hit 60GB. This is also assuming that I don't use my connection for anything else.
Now, they have said that it wont' be using the full 5Mbps constantly, but lets say it even averages at 2.5, that's still only about 1.9 hours of gaming a day to use my entire bandwidth allocation for the month, and that's assuming no one uses the internet for anything in a month.
I'm sorry, but for that reason alone, this will never work.
The negative impact I'm concerned about is, if this service takes off and becomes a huge deal, it's going to severely hurt the modding community. Why's a modder going to make large total conversion mods if they know that 95% of the people who own that game CAN'T use it? Also, if it becomes the primary way of gaming, and if games become exclusive on the service it's going to just relaunch the console wars on the PC, when other companies launch similar services. Have fun going on your PC to play a game only to find out it's the exclusive of a service you're not on. Not to mention the fact that if games start becoming exclusive on this service it's going to really really hurt the market for high-end PCs, this might sound like a good thing, but without people having a demand for their cards, expect companies like sapphire, BFG and others to close up shop, and expect graphic card tech to slow down as the demand's not there any more.
My big concern with this thing, is if it becomes the primary way of gaming on the PC we're going to be locked into a pay by month service. If they raise their monthly fees you have no choice but to pay them if you don't want to lose all the games you've already paid for. Unless, of course, you don't actually have to pay an extra fee for individual games, and you just get access to an entire library on demand. Which sounds like a good thing, but at the same time it probably means the monthly fees are going to be pretty high to accommodate the cost of hardware and software development, which might make it cost prohibitive to those who have a limited amount of play time a month, but if this does become the main method of game playing, you might not have a choice.
My primary concern is essentially the same one I've got with Steam. Regardless of subscription charges (which really is a huge negative for me when I'm already paying to "buy" the game as well), the biggest problem is that your games are fundamentally in someone elses control.
We had a thread about a week ago that illustrated this quite clearly when a guy accidentally got his steam account locked and Steam support refused to believe he wasn't a scammer, and simply wouldn't even discuss the issue at all (whenever he asked why his account was locked they literally told him they weren't going to tell him, it was just locked). It only got resolved after he went past Steam support and sent an e-mail in to Gabe Newell and co. to explain the situation.
The issue for me is losing access to everything I've already purchased because I don't actually own any of it. The games, even the singleplayer games, aren't products anymore they're services. It's pretty much the reason why I don't like online verification. Company goes bust, sells out, quite simply de-activates the authentication servers, and what you're left with is a coaster, or not even that if you bought a DD copy.
I mean, I actually do like Steam, but when buying games I avoid it as much as possible because I know that it's fundamentally ceding authority over whether I can use my purchased products to someone else. Throw in a subscription charge in order to access the games you've purchased, as looks to be the case here, and it makes me a whole lot more sceptical.
At the same time, I think it's a shift that the entire industry is heading towards, not just PC gaming. The two primary complaints of any content industry are always the same: Piracy, and the second hand market. And whilst movie and music manufacturers haven't been as vocal about the second hand market, the games industry has been complaining more and more about retailers re-selling games whilst they get none of the profit from that transaction. My personal opinion on that is basically "cry some more", but that's another topic.
In any case, the industry wants to shift away from a product based architecture to one where they know that every transaction is a profit for them, which is why it's becoming more like a service industry. I can very easily see consoles in a generation or two being primarily based around DD, not retail sales, for their profits. Personally I don't really think this is a good thing, at least not for the consumer, but I also don't think there's much that can be done to stop it.
While I understand the hesitation towards gaming with this type of service, I am pretty optimistic about it.
Yes, it will use a lot of bandwidth, yes, you don't own the games, yes, it could have problems, but there are so many good sides to this.
Depending on how pricing ends up:
Renting games could be the new way to go. Rent a game, play it till you beat it, and then move on. Rather than spending $60 for a game, perhaps you are spending $10 for two weeks and you've beaten it. That's cheaper than turning it back in to GameStop for $35. If the service fails, you haven't lost anything because it was just a rental. Perhaps later you want to pick it up again? Rent it again. Don't like it? You're out a few bucks but at least you can rent PC games to try them out. Where else can you do that?
Everything is moving online. Perhaps your connection or bandwidth limitations won't allow it now, but as more products require bandwidth, ISPs will be pressured to remove restrictions (or convinced with money hats) so companies are able to continue to provide services (and earn money) through the internet. Currently, some business models aren't that great, sure, but things are getting better and better. When we first starting buying music online, it cost $1 for a low quality, DRM song that you had to burn to a CD yourself if you wanted a backup. Terrible, isn't it? yet people bought them by the billions. Now, you get higher quality, no DRM, but you still can't redownload them. Perhaps in the future, you will be able to. But now you can even pay $15 a month and get all the music you want, and keep 10 songs permanently! Progress!
So perhaps this service won't be perfect to start with, but if it starts out somewhat successful, the wonderful thing about internet services is how easily they can change. Perhaps pricing will change, perhaps they'll over the ability to request discs for games that you've purchased outright for offline gaming, or perhaps the incentive will be cheaper costs to purchase games through this service than in a store.
Rights management is becoming such a big issue with "non-tangible" goods where you aren't buying a physical product, but rather the rights to use something like code. It's still a developing issue that is attempting to find the right place and type of use that both benefits the user, and makes money for the person selling it. Give it time, and I think things will work out pretty well.
Once the beta starts, and we hear more impressions on how it all works from people like us, I think we can then decide if this is going to work or not. Until then, I'm going to be optimistic because I have NWN2 here that I've never been able to play because I'm too cheap to replace my broken video card since I should probably upgrade my whole computer anyway. If I can play the latest and greatest on a $400 machine, I would love it. I want to play these games, but I can't afford a new computer and the games I want to play, but this service may be able to do that.
Yes, it's not for everyone, but think about how much the PC side of the games industry has declined in the past few years as games required more and better hardware that most people weren't buying? How many games would people have bought if they could be run on any computer?
As far as crashes, etc. They aren't taking a retail game and installing it on these servers. It sounds like the companies are writing specialized installs for them that would take advantage of the setup they have. Bugs would be much easier to iron out as they can write for a specific piece of hardware (similar to consoles) rather than any PC out there with all sorts of different drivers and configurations. It could really allow for a lot less patching for PC games, and invisible patching for us as we don't have to wait for the 2GB patch to download during the time we wanted to play. They patch the server, and anyone connected to it benefits from the patch immediately.
Anyway, that's a huge wall of text, but I'm pretty excited about the prospect, and I am going to look forward to hearing more as this gets closer to release.
I think some people in this thread are way to hyped up over a service we haven't even seen in action brought to you by the same guy who made WebTV and that $800 Moxi DVR box.
My guess is like with he's created something that's ready to sell to another company, but not ready for market. There are just too many factors that will make this not work, and all I've seen from them is slick marketing talk.
As far as crashes, etc. They aren't taking a retail game and installing it on these servers. It sounds like the companies are writing specialized installs for them that would take advantage of the setup they have. Bugs would be much easier to iron out as they can write for a specific piece of hardware (similar to consoles) rather than any PC out there with all sorts of different drivers and configurations. It could really allow for a lot less patching for PC games, and invisible patching for us as we don't have to wait for the 2GB patch to download during the time we wanted to play. They patch the server, and anyone connected to it benefits from the patch immediately.
It would also mean that the service would probably have to wait, possibly months or more for companies to basically port their games over to the service. Either that or the rest of us who don't have the service could end up having to wait months for it to make its way over to us.
I do like the idea of structuring it as a rental system. So you pay like $10 for the game and you get to play it as much as you want for a week, no other service or subscription charges. That could be worth it for some people. Not for me, personally, simply because I like to own among other reasons.
True, we haven't seen it and there's not much to get hyped up about other than the idea, but we're talking on a board that gets hyped up for any game idea that comes along that sounds cool, without even seeing it.
Plus, with the companies that are already on board, it seems that much more tangible.
I can't imagine that porting to this service is too much different from just doing a regular PC game, so I don't think it would really delay anything by months. But I know nothing of game development so who knows..
Anyway, right now, it's a cool idea, it's got some hurdles that they think they've managed to, well, hurdle, and until the Beta players start talking and then eventually it's released, we won't know for sure if it'll work or not.
We also don't know how "on board" the publishers they have listed are either. They may have only commited to one or two games each. And depending on how it works or doesn't work they may not support it beyond that.
Drool on
0
citizen059hello my name is citizenI'm from the InternetRegistered Userregular
edited March 2009
Well, I for one support the idea and hope they succeed.
Right now PC gaming is for PC savvy people and that's about it. Sure, I'm smart enough to work my way around issues that arise, but you can make a lot more money selling your games to the much larger audience of people who aren't that technical about it.
Look at Nintendo - with the DS and especially the Wii and the simple yet entertaining games available, they've made $$$ by appealing to the casual market first. My wife and I have friends picking up the Wii for stuff like Wii Fit, and these are people who otherwise had zero interest in games of any kind outside of maybe solitaire or Pogo.com.
If you can get people to play PC games without having to purchase a crazy expensive machine, and without having to spend an hour or two browsing random internet forums to find an obscure method of making your game work, then why the hell not try it?
Sure, right now we should all be in the "believe it when I see it" stage to some degree, but damn it sure would be nice to see this get off the ground and work the way they say it does.
From an IT perspective I can see more potential for this outside of just gaming. I mean, these computers that OnLive clients connect to are essentially headless Windows boxes, with the ability to stream out full multimedia information with very very little latency onto a low powered thin client.
Other commercially utilized remote clients such as Citrix, X11, RDP, or VNC don't perform nearly this well. If anything, Videoconferencing companies should be showering them with money to license their technology.
From a gaming side, what would be more amazing is when companies don't have to develop games that skew towards the lowest common denominator, making games for the highest end hardware without compromise. I can forsee games that model each and every bit of environmental change (ala Molyneux's original promise with Fable) and MMOs that are more action oriented without concern for ping times between client and server.
With some current and future mobile phones able to process 720p video, a mobile client would effectively circumvent the limitations (other than bandwidth caps and whatnot) of the platform. While twitch games may not be suitable (lack of buttons, spotty connection, etc), imagine playing something like Pixeljunk Monsters or Dawn of War II on your iPhone/WinMo/Android phone, with multiplayer options.
Edit: But I don't want to put the cart before the horse. Let's all hope the beta works as advertised.
From an IT perspective I can see more potential for this outside of just gaming. I mean, these computers that OnLive clients connect to are essentially headless Windows boxes, with the ability to stream out full multimedia information with very very little latency onto a low powered thin client.
Other commercially utilized remote clients such as Citrix, X11, RDP, or VNC don't perform nearly this well. If anything, Videoconferencing companies should be showering them with money to license their technology.
From a gaming side, what would be more amazing is when companies don't have to develop games that skew towards the lowest common denominator, making games for the highest end hardware without compromise. I can forsee games that model each and every bit of environmental change (ala Molyneux's original promise with Fable) and MMOs that are more action oriented without concern for ping times between client and server.
With some current and future mobile phones able to process 720p video, a mobile client would effectively circumvent the limitations (other than bandwidth caps and whatnot) of the platform. While twitch games may not be suitable (lack of buttons, spotty connection, etc), imagine playing something like Pixeljunk Monsters or Dawn of War II on your iPhone/WinMo/Android phone, with multiplayer options.
A few points, first, you are right, this kind of capability would have a lot of uses. Which brings up a good question: why would you start with such a hard case, i.e. action oriented video games?
Second, lag will be worse as far as ping times go, but MMOs could be helped by removing client lag in crowded areas.
Third, console games, especially HD ones are closer to being budget limited than they are to being limited by the hardware. If this OnLive led to more revenue for developers that would do more to make room for more in depth games than the hardware improvements would.
lowlylowlycook on
(Please do not gift. My game bank is already full.)
With some current and future mobile phones able to process 720p video, a mobile client would effectively circumvent the limitations (other than bandwidth caps and whatnot) of the platform. While twitch games may not be suitable (lack of buttons, spotty connection, etc), imagine playing something like Pixeljunk Monsters or Dawn of War II on your iPhone/WinMo/Android phone, with multiplayer options.
I don't know about you, but I'm making split-second decision all the time in Dawn of War 2, being a fraction of a second too slow leads to losing entire squads, or not getting to victory points fast enough.
It would also make the game hideously irritating since it has to pause until all players are synched up again, extra lag would NOT be a nice thing.
From an IT perspective I can see more potential for this outside of just gaming. I mean, these computers that OnLive clients connect to are essentially headless Windows boxes, with the ability to stream out full multimedia information with very very little latency onto a low powered thin client.
Other commercially utilized remote clients such as Citrix, X11, RDP, or VNC don't perform nearly this well. If anything, Videoconferencing companies should be showering them with money to license their technology.
From a gaming side, what would be more amazing is when companies don't have to develop games that skew towards the lowest common denominator, making games for the highest end hardware without compromise. I can forsee games that model each and every bit of environmental change (ala Molyneux's original promise with Fable) and MMOs that are more action oriented without concern for ping times between client and server.
With some current and future mobile phones able to process 720p video, a mobile client would effectively circumvent the limitations (other than bandwidth caps and whatnot) of the platform. While twitch games may not be suitable (lack of buttons, spotty connection, etc), imagine playing something like Pixeljunk Monsters or Dawn of War II on your iPhone/WinMo/Android phone, with multiplayer options.
A few points, first, you are right, this kind of capability would have a lot of uses. Which brings up a good question: why would you start with such a hard case, i.e. action oriented video games?
Second, lag will be worse as far as ping times go, but they could be helped by removing client lag in crowded areas.
Third, console games, especially HD ones are closer to being budget limited than they are to being limited by the hardware. If this OnLive led to more revenue for developers that would do more to make room for more in depth games than the hardware improvements would.
I would assume that their choice of market is because: 1. Enterprise software vs Commercial software development is tough and yields little reward since the audience is small 2. It's just more fun.
What I meant in terms of client/server connections in MMOs is not the connection between the user and the Onlive servers, but rather, the client and server inside the NOC itself. What if an MMO developer were to put their servers next to Onlive's servers, allowing the client computers direct access not over the Internet, but within the LAN?
I agree that the cost of producing a high definition game is a limiting factor, that's why I didn't mention increasing graphical improvements in future games. Making games more in-depth (say, in terms of advanced AI interaction or variables in choices) as you and I agree will require more that just GPUs, but it does require more in processing power and in disk space, something that a netbook can't provide locally.
I don't know about you, but I'm making split-second decision all the time in Dawn of War 2, being a fraction of a second too slow leads to losing entire squads, or not getting to victory points fast enough.
It would also make the game hideously irritating since it has to pause until all players are synched up again, extra lag would NOT be a nice thing.
Agreed, that was a bad example. I was just thinking of a RTS and since that's what I was playing recently, it immediately came to mind. A better example would be a Strategy game, such as Disguea or a turn based RPG. Even something like Valkyria Chronicles would be nice to play on a phone, provided something could be done about readability of text and whatnot.
I don't know about you, but I'm making split-second decision all the time in Dawn of War 2, being a fraction of a second too slow leads to losing entire squads, or not getting to victory points fast enough.
It would also make the game hideously irritating since it has to pause until all players are synched up again, extra lag would NOT be a nice thing.
Agreed, that was a bad example. I was just thinking of a RTS and since that's what I was playing recently, it immediately came to mind. A better example would be a Strategy game, such as Disguea or a turn based RPG. Even something like Valkyria Chronicles would be nice to play on a phone, provided something could be done about readability of text and whatnot.
The one thing I haven't even seen mentioned yet is that there is at least 80 ms of latency just to start off. Then you have the round trip latency from your onlive box to the server located somewhere in the US. For me on dsl to ping google is about 70 ms, so lets estimate 100ms. Now you have 180 ms of latency in any game. That's 11 full frames in a game running at 60fps. You also have the regular control latency on games which I think averages around 2 frames these days, which normally doesn't matter, but in this case it's added to a higher number. Now we've got around 13 frames of latency from when we see the result of our action on screen.
With that kind of delay this system really isn't going to work well for anything that requires reactions even close to real time. The best case scenario for this company is probably that Time Warner will move from being an investor and just out-right acquire them. You will then be able to play checkers, chess, and settlers of catan on your cable box.
Even complicated turn based games are out given the 1:1 scaling ratio required on the backend because of the technical requirements of the games, compression, and streaming.
Cronus on
"Read twice, post once. It's almost like 'measure twice, cut once' only with reading." - MetaverseNomad
My ISP has a 60GB limit per month as well (fucking Rogers). For every gig I go over, its almost $2 per gig that is charged. I'm sure more ISP's are going to implement these caps in the future. I don't see this service lasting long. Do you guys honestly believe ISP's are going to change their policy on caps just because a bunch of people want to play games? I mean, fuck, these are the same guys that want to abolish net neutrality (my ISP is one of many). ISP's want your fucking money, and will try any excuse to get it from you "wah wah bandwidth is too pricey, we have to cap it".
If anything, if this service does take off and lots of people are playing it, this might make the ISP's more pissed off because now alot more bandwidth is being used, and you know how lazy these shits are in terms of upgrading their infrastructure, and instead will enforce harder caps.
My ISP has a 60GB limit per month as well (fucking Rogers). For every gig I go over, its almost $2 per gig that is charged. I'm sure more ISP's are going to implement these caps in the future. I don't see this service lasting long. Do you guys honestly believe ISP's are going to change their policy on caps just because a bunch of people want to play games? I mean, fuck, these are the same guys that want to abolish net neutrality (my ISP is one of many). ISP's want your fucking money, and will try any excuse to get it from you "wah wah bandwidth is too pricey, we have to cap it".
If anything, if this service does take off and lots of people are playing it, this might make the ISP's more pissed off because now alot more bandwidth is being used, and you know how lazy these shits are in terms of upgrading their infrastructure, and instead will enforce harder caps.
I mentioned above that I think this service will only see profitability if bought by Time Warner, or another cable company. In that case it doesn't matter as it will be done over TV cable.
However if they somehow do come to market and people actually use it to buy/rent modern games from it, then yes the caps will be a problem. Comcast, Qwest, and Charter all have caps. I'm sure that many other ISP's do as well. Playing these games over them will really speed up people hitting their bandwith limits.
Also something I didn't even take into account in my previous post is traffic shaping. If done to the packets for this service, it will add even more latency to the game. In the US this isn't too much of a problem yet, however Canadian and British gamers probably would find the service unusable if it found the ISP's ire.
Cronus on
"Read twice, post once. It's almost like 'measure twice, cut once' only with reading." - MetaverseNomad
Difference between Citrix/RDP/etc and this is that you'll accept a little blurring of sharp edges in a game - hell, it's free anti-aliasing, amirite? :P
If I had to stare at a screenful of blurry text all day, ow my head
PeregrineFalcon on
Looking for a DX:HR OnLive code for my kid brother.
Can trade TF2 items or whatever else you're interested in. PM me.
There is no way to remove the ping time between your computer and theirs. The controls will ALWAYS lag, and it will often be a noticeable amount. This isn't like networked gaming where lag is taken in to account by the server and all this prediction stuff comes in to play, this is introducing a timeout between manipulating your interface and what you see. Can you imagine playing a twitch shooter where every movement you make is a tenth of a second behind what's on the screen? You're dead before you even see what killed you.
I'm pretty sure this is going to end up just like the Phantom...impossible to pull off in the real world.
EDIT: ALSO...
Cloud computing works on the idea that not everyone uses peak resources, so you shift the load around. Games always use the full resources of the system. You can't have two people using the same computer to play crysis, nor can you spread it out over time, as gaming happens at very specific times during the day. It will be very hard for them to economically have an affordable subscription that is actually usable and won't find you waiting behind a huge queue when you want to play.
jonxp on
Every time you write parallel fifths, Bach kills a kitten.
3DS Friend Code: 2707-1614-5576 PAX Prime 2014 Buttoneering!
Thought came to mind. If this is actually deliverable, and gets any kind of penetration due to the gaming aspect, IPTV distributors will be salivating at the prospect of using OnLive as a rider for their shows and content. That's just streaming video, not even requiring the horsepower that a gaming system would. Maybe a minimum of interactivity.
There is no way to remove the ping time between your computer and theirs. The controls will ALWAYS lag, and it will often be a noticeable amount. This isn't like networked gaming where lag is taken in to account by the server and all this prediction stuff comes in to play, this is introducing a timeout between manipulating your interface and what you see. Can you imagine playing a twitch shooter where every movement you make is a tenth of a second behind what's on the screen? You're dead before you even see what killed you.
I'm pretty sure this is going to end up just like the Phantom...impossible to pull off in the real world.
EDIT: ALSO...
Cloud computing works on the idea that not everyone uses peak resources, so you shift the load around. Games always use the full resources of the system. You can't have two people using the same computer to play crysis, nor can you spread it out over time, as gaming happens at very specific times during the day. It will be very hard for them to economically have an affordable subscription that is actually usable and won't find you waiting behind a huge queue when you want to play.
This is good. A more in-depth and descriptive version of the scaling problems I mentioned.
This is much different than the Phantom though. The Phantom was essentially a computer that ran nothing but Steam that you hook up to your TV. It's not a terrible idea. It was a little ahead of it's time, but the main problem was that the company was run by very corrupt people who just wanted to steal VC money.
Cronus on
"Read twice, post once. It's almost like 'measure twice, cut once' only with reading." - MetaverseNomad
jon - more than one person cannot play Crysis at a time on a single normal computer. They are claiming to have custom servers with massive computational power.
I'm not saying it resolves the issue, but it does seem to answer that one point in some small way.
NotASenator on
0
Olivawgood name, isn't it?the foot of mt fujiRegistered Userregular
edited March 2009
Lots of people seem to be making wildly different guesses as to exactly how many milliseconds of lag there is going to be
I think we will have to wait and see before we know exactly how hard it's going to be to control
Of course, caution and skepticism is fine, even warranted in this situation, but you have to admit, the prospect of this technology is exciting, and if it works, it'll have effects beyond the game industry
Also it seems like a lot of people didn't catch that presentation they did. It should be up on youtube or whatever by now, so everyone should go watch it, it answers a few questions at the very least
jon - more than one person cannot play Crysis at a time on a single normal computer. They are claiming to have custom servers with massive computational power.
I'm not saying it resolves the issue, but it does seem to answer that one point in some small way.
There is no such magical server hardware that allows one to have the processing powers of multiple gaming rigs for less than the price of multiple gaming rigs.
Especially considering that games are designed for consumer hardware, not server hardware.
EDIT:
"Speculation on how many milliseconds of lag" is grounded in reality. There will be a minimum network latency no matter what. This is just how the internet works. Data must travel between many points before it arrives at its destination, and each point takes time to process it. You can remove all the latency on the server side, and client side, but the network latency can only be removed by placing the servers closer to you. Which is fine if you live near a peering point like Atlanta, Dallas, or LA. It will SUCK if you live in Minnesota and want to play games.
jonxp on
Every time you write parallel fifths, Bach kills a kitten.
3DS Friend Code: 2707-1614-5576 PAX Prime 2014 Buttoneering!
0
Olivawgood name, isn't it?the foot of mt fujiRegistered Userregular
jon - more than one person cannot play Crysis at a time on a single normal computer. They are claiming to have custom servers with massive computational power.
I'm not saying it resolves the issue, but it does seem to answer that one point in some small way.
There is no such magical server hardware that allows one to have the processing powers of multiple gaming rigs for less than the price of multiple gaming rigs.
Especially considering that games are designed for consumer hardware, not server hardware.
That's what they were saying though, if I understood correctly
Using these weird custom built servers using powerful GPUs and shit, one server can run multiple games at once, since running a game doesn't require all the processing power of a single computer or GPU or whatever
Look, I don't know science, but I got the distinct feeling that they have some idea of what they're doing
But again, none of us knows what the fuck
edit: yes I know jon but some people are saying 22ms and some people are saying 180ms
I don't know but there seems to be a bit of a discrepancy there
OnLive is pretty much the belle of the ball at GDC thus far, and its promise of rapture must be pure exhilaration to anyone who has never worked IT. I wouldn't describe it as a fight, per se - the knives remained sheathed - but I would say that Gabriel and I entered into what you might call a scale model of a fight. There was fury, but it was a very compact fury.
I received an SMS from Robert late last night, no doubt typed at some Game Developer's Conference bacchanal, and something in the way it was formatted managed to communicate a breathless enthusiasm. I reminded him that years ago - during E3 That Was - we saw a demo of the Phantom that was impressive enough to elicit a purchase. The demo was that good. So good that I never wrote about it - it was, in fact, too good. The machines that served the content were situated in a room not ten feet away, connected via gigabit ethernet. Every time I pore over some unchecked torrent from an ecstatic new convert, I start to become very curious about topology.
I mentioned IT before because people who have done time in the field have already been through several cycles of this kind of thinking. I'm not making a value judgment of any kind - I'm just saying that terminal computing is perpetually making a case for itself, because the seat processing power is always shifting. Back when I worked for the school district, the basement of the administration building was given over to an ancient mainframe serviced by equally ancient people who only knew how to tend to their beloved, humming sarcophagus.
What I'm saying is that server-side solutions invariably lead to sinister necromantic cabals.
I'm not saying that exactly. Even sight unseen, there are genres I prize greatly that would work well on a mechanism like this. Virtually any raw tactics experience would excel under these conditions. Many role-playing titles and even some massive games would perform adequately, even in the absence of any unique technology. The new Prince is the sort of platformer that might function, and it features prominently in their materials. When I see a racer or a shooter, though, something in my mind rebels.
They have some incredibly forward thinking community solutions, and as a rental scenario what OnLive delivers is near optimal. I resonate in a harmonious way with the idea that discrete gaming "platforms" are a divisive, untenable regime that often obscures the medium. This is precisely why I am careful about things of this stripe: it is the will of my heart to believe it. I believe Matthew 24, verses 23 through 26 might be appropriate here. Due diligence.
If a man claims to be Jesus Christ, you can bet I'll check the wrists.
Personally this whole thing seems comparable to being told that someone has used phychophysics to unlock cold fusion and is starting up a company to make lawn mowers that don't ever need to be refuled, and would you be interested in investing in said company.
lowlylowlycook on
(Please do not gift. My game bank is already full.)
jon - more than one person cannot play Crysis at a time on a single normal computer. They are claiming to have custom servers with massive computational power.
I'm not saying it resolves the issue, but it does seem to answer that one point in some small way.
There is no such magical server hardware that allows one to have the processing powers of multiple gaming rigs for less than the price of multiple gaming rigs.
Especially considering that games are designed for consumer hardware, not server hardware.
That's what they were saying though, if I understood correctly
Using these weird custom built servers using powerful GPUs and shit, one server can run multiple games at once, since running a game doesn't require all the processing power of a single computer or GPU or whatever
Look, I don't know science, but I got the distinct feeling that they have some idea of what they're doing
But again, none of us knows what the fuck
edit: yes I know jon but some people are saying 22ms and some people are saying 180ms
I don't know but there seems to be a bit of a discrepancy there
Reading and listening to their stuff, it seems that games that use fewer resources, they'll run multiple instances of the game using VMs and such, but they're still running regular hardware. For "big" games, they'll have to using a one-to-one server thing. I still think they're going to run into bottlenecks in the timesharing based on the narrow timing of people playing games in the afternoons.
For lag: As an example, my home internet happens to have the same "home" as my TF2 server. I get an average of 10ms latency for bare pings between here and there. It's less than 15 miles away, and only touches four routers. It just gets worse from there.
jonxp on
Every time you write parallel fifths, Bach kills a kitten.
3DS Friend Code: 2707-1614-5576 PAX Prime 2014 Buttoneering!
OnLive is pretty much the belle of the ball at GDC thus far, and its promise of rapture must be pure exhilaration to anyone who has never worked IT. I wouldn't describe it as a fight, per se - the knives remained sheathed - but I would say that Gabriel and I entered into what you might call a scale model of a fight. There was fury, but it was a very compact fury.
I received an SMS from Robert late last night, no doubt typed at some Game Developer's Conference bacchanal, and something in the way it was formatted managed to communicate a breathless enthusiasm. I reminded him that years ago - during E3 That Was - we saw a demo of the Phantom that was impressive enough to elicit a purchase. The demo was that good. So good that I never wrote about it - it was, in fact, too good. The machines that served the content were situated in a room not ten feet away, connected via gigabit ethernet. Every time I pore over some unchecked torrent from an ecstatic new convert, I start to become very curious about topology.
I mentioned IT before because people who have done time in the field have already been through several cycles of this kind of thinking. I'm not making a value judgment of any kind - I'm just saying that terminal computing is perpetually making a case for itself, because the seat processing power is always shifting. Back when I worked for the school district, the basement of the administration building was given over to an ancient mainframe serviced by equally ancient people who only knew how to tend to their beloved, humming sarcophagus.
What I'm saying is that server-side solutions invariably lead to sinister necromantic cabals.
I'm not saying that exactly. Even sight unseen, there are genres I prize greatly that would work well on a mechanism like this. Virtually any raw tactics experience would excel under these conditions. Many role-playing titles and even some massive games would perform adequately, even in the absence of any unique technology. The new Prince is the sort of platformer that might function, and it features prominently in their materials. When I see a racer or a shooter, though, something in my mind rebels.
They have some incredibly forward thinking community solutions, and as a rental scenario what OnLive delivers is near optimal. I resonate in a harmonious way with the idea that discrete gaming "platforms" are a divisive, untenable regime that often obscures the medium. This is precisely why I am careful about things of this stripe: it is the will of my heart to believe it. I believe Matthew 24, verses 23 through 26 might be appropriate here. Due diligence.
If a man claims to be Jesus Christ, you can bet I'll check the wrists.
Personally this whole thing seems comparable to being told that someone has used phychophysics to unlock cold fusion and is starting up a company to make lawn mowers that don't ever need to be refuled, and would you be interested in investing in said company.
The company with the most lawnmowers is the company that wins.
Latency is caused by physics. Decades of science fiction notwithstanding, physicists have yet to figure out how to surpass the speed of light. Because of their failure (the fools! I mean really, how hard can this be?), we are unable to send any data faster than 186282 miles per second.
That's still pretty fast, though, right?
I live in Seattle. My colleague Eli used to live in New York. That is 2413 miles away, which is 13 milliseconds at the speed of light.
I used to live in England. From Seattle to England is 4799 miles, which is 26 milliseconds.
In a 60 frames per second game, each frame gets 16 milliseconds. So I already have nearly two frames lag when I play with my friends in England.
But wait! This is not the whole story...
Network data does not travel through a vacuum. The speed of light is usually measured in a vacuum, but when data travels down fiber optic or copper cables, it slows to only 60%.
Also, network data does not only travel along cables. There will also be two modems, one at either end. Each modem adds around 10 milliseconds latency. If my friend does not use the same ISP and switchboard as me, there will also be several routers along the way. Each router adds between 5 and 50 milliseconds latency.
How bad can it get?
Xbox games are expected to work with latencies up to 200 milliseconds
Obviously, prediction algorithms and shit effect that stuff, and the concept is that they will be close enough to you to handle stuff. From here at work in Reston, VA, pinging the Speedtest.net server in DC, about 15 miles away over a Verizon Business T1 averages about 250ms roundtrip, but there is a lot of traffic going in and out of here, so I guess I won't be gaming from work.
Using VNC to remote into my home system, I am getting an average of 105ms roundtrip to the same server, which is probably about five miles closer there, and this is the result of the speedtest.net test on their page:
So it's a pretty fast connection.
None of this speaks to the type of latency we'd see with this service, but it does go to show that even just jumping across the river to throw out a ping can take a tenth of a second.
They're using patented time manipulation hardware wherein the computer actually KNOWS what you will be doing 180 ms in the future to compensate for the latency.
I call bullshit. They claim 1ms latency up to 1000 miles away? Or did I get that wrong? Well, anyway, suppose you had a direct fiber link to their box (no switches in between, nothing), then Einstein claims:
(1000 miles)*(1609.3 meters/mile)/(3e10 m/s) gives 5.37ms. And that's just for one-way, with no switches, on a perfect slightly cloudy day.
P.S., I hope I don't offend anybody with my lack of significant figures.
I call bullshit. They claim 1ms latency up to 1000 miles away? Or did I get that wrong? Well, anyway, suppose you had a direct fiber link to their box (no switches in between, nothing), then Einstein claims:
(1000 miles)*(1609.3 meters/mile)/(3e10 m/s) gives 5.37ms. And that's just for one-way, with no switches, on a perfect slightly cloudy day.
P.S., I hope I don't offend anybody with my lack of significant figures.
1ms is just the time to encode the video and push it down some fat pipes.
I call bullshit. They claim 1ms latency up to 1000 miles away? Or did I get that wrong? Well, anyway, suppose you had a direct fiber link to their box (no switches in between, nothing), then Einstein claims:
(1000 miles)*(1609.3 meters/mile)/(3e10 m/s) gives 5.37ms. And that's just for one-way, with no switches, on a perfect slightly cloudy day.
P.S., I hope I don't offend anybody with my lack of significant figures.
The 1ms claim is the time it takes to encode the stream (a pretty significant claim in itself), not the latency between the client and server.
Edit: Since my post is redundant *Shakes fist at agoaj*, I'll address something else:
jon - more than one person cannot play Crysis at a time on a single normal computer. They are claiming to have custom servers with massive computational power.
I'm not saying it resolves the issue, but it does seem to answer that one point in some small way.
There is no such magical server hardware that allows one to have the processing powers of multiple gaming rigs for less than the price of multiple gaming rigs.
Especially considering that games are designed for consumer hardware, not server hardware.
That's what they were saying though, if I understood correctly
Using these weird custom built servers using powerful GPUs and shit, one server can run multiple games at once, since running a game doesn't require all the processing power of a single computer or GPU or whatever
Look, I don't know science, but I got the distinct feeling that they have some idea of what they're doing
But again, none of us knows what the fuck
edit: yes I know jon but some people are saying 22ms and some people are saying 180ms
I don't know but there seems to be a bit of a discrepancy there
As jon stated (and myself in a earlier post), even if they managed to get racks of servers with multiple PCs on a board, you can't timeshare all of them. High end games were built to take up as much resources in a single computer as possible for maximum performance. Thus you run into a scenario of several thousand people attempting to run Crysis and having to scale your network center appropriately to meet demand. You can cut costs to heat and power by deploying more efficient design (aforementioned PC on a board + physical case layout), but it's still a 1:1 relationship for remote computers and each client has a pretty high X associated cost that may be hard to profit from. I don't doubt they can do it (nothing appears shady within their ranks), but the business model, along with the ongoing latency issue does seem suspect.
Actually, based on refraction, the effective speed of the light traveling down the optical fiber is slower than that. So your number is actually generous.
I hope some people in the thread made the same comments as I'm about to make, it's important to realize the repercutions.
Cloud computing means the information is stored on their servers, you don't have anything. It means major privacy issues. Everything that you do in a game, and also in your private life would be at someone else's reach, potentially. This poses a huge risk I'm not willing to take.
Also with the fact you DEPEND on them for even getting anything or playing any game at all, whenever you're bored, it means that they can charge you far higher than if you had bought a high end PC. This is like X-Box Live but worse, pretending to save money but really you're paying $50 a year for the past x years, it adds up. This would be far worse, you'd burn through the same amount as a high end PC. And don't think that games wouldn't be tiered. You want to play a hot game with great graphics on release? That's extra than playing 1 year old games on medium graphics. Of course they'll do it.
This also means that small companies won't be able to get their name out there, a lot of word of mouth is generated by sharing CDs with friends, playing a free demo etc Already you might have heard of the PS3 really crippling small publishers by charging for all bandwidth, a small publisher simply cannot get its demo out on the PS3.
You have to realize that this cloud computing thing means you give all power to someone else. This from a consumer point of view is not very good at all.
What is really the ONLY benefit? Really, the only one : No more need to upgrade computers. Do people really understand how the pricing will rip us off? Do people really care about games on demand? You can already have that with services like Gametap anyway.
Think this cloud computing, cloud gaming thing through, personally I will never ever buy into such a scheme, ever.
As for upgrades, am I mad that I can't play Empire : Total War? Sure. But I'm still loving it playing old games again, or maybe Drakensang and it's lower requirements . Also I don't own a HD TV and probably never will. People spend a lot of money on TVs. Save that TV money and just use it for PC upgrades.
Ah, ok, 1ms for encoding. So, 1ms for encoding, 5.something ms for one-way trip (as NotACrook points out, more than that), 5.something ms for the return path to get the video to you. 11-13 ms just for optics and encoding. There's also generating the content, but that should be negligible.
So,
* what do the switches and ISP's add?
* assuming optical backbones, what is the cost of fiber<->copper conversions? (edit: gotta be negligible, right?)
* how much does the typical household router add?
Really, you can get by playing modern games on a pretty cheap computer. You won't have the highest graphics, but when I've ran UT3 at 30FPS on a geforce 6200GT, 512mb of DDR1 ram, and a AMD 1600+, people claiming that you have to pay out the ass for PC gaming is just lies
Posts
I have a 60GB soft cap on my internet. if I pass 60GB, they dont' cut me off, and doing it once in a while is fine. But if I pass 60GB say, 3 months in a row, I start getting nasty phone calls, risk overage charges or getting my account turned off. if you take the the fact that they say they require a 5Mbps connection for this service, that means that I can game for just under an hour day in a 30 day month, and I'll hit 60GB. This is also assuming that I don't use my connection for anything else.
Now, they have said that it wont' be using the full 5Mbps constantly, but lets say it even averages at 2.5, that's still only about 1.9 hours of gaming a day to use my entire bandwidth allocation for the month, and that's assuming no one uses the internet for anything in a month.
I'm sorry, but for that reason alone, this will never work.
My primary concern is essentially the same one I've got with Steam. Regardless of subscription charges (which really is a huge negative for me when I'm already paying to "buy" the game as well), the biggest problem is that your games are fundamentally in someone elses control.
We had a thread about a week ago that illustrated this quite clearly when a guy accidentally got his steam account locked and Steam support refused to believe he wasn't a scammer, and simply wouldn't even discuss the issue at all (whenever he asked why his account was locked they literally told him they weren't going to tell him, it was just locked). It only got resolved after he went past Steam support and sent an e-mail in to Gabe Newell and co. to explain the situation.
The issue for me is losing access to everything I've already purchased because I don't actually own any of it. The games, even the singleplayer games, aren't products anymore they're services. It's pretty much the reason why I don't like online verification. Company goes bust, sells out, quite simply de-activates the authentication servers, and what you're left with is a coaster, or not even that if you bought a DD copy.
I mean, I actually do like Steam, but when buying games I avoid it as much as possible because I know that it's fundamentally ceding authority over whether I can use my purchased products to someone else. Throw in a subscription charge in order to access the games you've purchased, as looks to be the case here, and it makes me a whole lot more sceptical.
At the same time, I think it's a shift that the entire industry is heading towards, not just PC gaming. The two primary complaints of any content industry are always the same: Piracy, and the second hand market. And whilst movie and music manufacturers haven't been as vocal about the second hand market, the games industry has been complaining more and more about retailers re-selling games whilst they get none of the profit from that transaction. My personal opinion on that is basically "cry some more", but that's another topic.
In any case, the industry wants to shift away from a product based architecture to one where they know that every transaction is a profit for them, which is why it's becoming more like a service industry. I can very easily see consoles in a generation or two being primarily based around DD, not retail sales, for their profits. Personally I don't really think this is a good thing, at least not for the consumer, but I also don't think there's much that can be done to stop it.
I made myself depressed.
Yes, it will use a lot of bandwidth, yes, you don't own the games, yes, it could have problems, but there are so many good sides to this.
Depending on how pricing ends up:
Renting games could be the new way to go. Rent a game, play it till you beat it, and then move on. Rather than spending $60 for a game, perhaps you are spending $10 for two weeks and you've beaten it. That's cheaper than turning it back in to GameStop for $35. If the service fails, you haven't lost anything because it was just a rental. Perhaps later you want to pick it up again? Rent it again. Don't like it? You're out a few bucks but at least you can rent PC games to try them out. Where else can you do that?
Everything is moving online. Perhaps your connection or bandwidth limitations won't allow it now, but as more products require bandwidth, ISPs will be pressured to remove restrictions (or convinced with money hats) so companies are able to continue to provide services (and earn money) through the internet. Currently, some business models aren't that great, sure, but things are getting better and better. When we first starting buying music online, it cost $1 for a low quality, DRM song that you had to burn to a CD yourself if you wanted a backup. Terrible, isn't it? yet people bought them by the billions. Now, you get higher quality, no DRM, but you still can't redownload them. Perhaps in the future, you will be able to. But now you can even pay $15 a month and get all the music you want, and keep 10 songs permanently! Progress!
So perhaps this service won't be perfect to start with, but if it starts out somewhat successful, the wonderful thing about internet services is how easily they can change. Perhaps pricing will change, perhaps they'll over the ability to request discs for games that you've purchased outright for offline gaming, or perhaps the incentive will be cheaper costs to purchase games through this service than in a store.
Rights management is becoming such a big issue with "non-tangible" goods where you aren't buying a physical product, but rather the rights to use something like code. It's still a developing issue that is attempting to find the right place and type of use that both benefits the user, and makes money for the person selling it. Give it time, and I think things will work out pretty well.
Once the beta starts, and we hear more impressions on how it all works from people like us, I think we can then decide if this is going to work or not. Until then, I'm going to be optimistic because I have NWN2 here that I've never been able to play because I'm too cheap to replace my broken video card since I should probably upgrade my whole computer anyway. If I can play the latest and greatest on a $400 machine, I would love it. I want to play these games, but I can't afford a new computer and the games I want to play, but this service may be able to do that.
Yes, it's not for everyone, but think about how much the PC side of the games industry has declined in the past few years as games required more and better hardware that most people weren't buying? How many games would people have bought if they could be run on any computer?
As far as crashes, etc. They aren't taking a retail game and installing it on these servers. It sounds like the companies are writing specialized installs for them that would take advantage of the setup they have. Bugs would be much easier to iron out as they can write for a specific piece of hardware (similar to consoles) rather than any PC out there with all sorts of different drivers and configurations. It could really allow for a lot less patching for PC games, and invisible patching for us as we don't have to wait for the 2GB patch to download during the time we wanted to play. They patch the server, and anyone connected to it benefits from the patch immediately.
Anyway, that's a huge wall of text, but I'm pretty excited about the prospect, and I am going to look forward to hearing more as this gets closer to release.
My guess is like with he's created something that's ready to sell to another company, but not ready for market. There are just too many factors that will make this not work, and all I've seen from them is slick marketing talk.
It would also mean that the service would probably have to wait, possibly months or more for companies to basically port their games over to the service. Either that or the rest of us who don't have the service could end up having to wait months for it to make its way over to us.
I do like the idea of structuring it as a rental system. So you pay like $10 for the game and you get to play it as much as you want for a week, no other service or subscription charges. That could be worth it for some people. Not for me, personally, simply because I like to own among other reasons.
Plus, with the companies that are already on board, it seems that much more tangible.
I can't imagine that porting to this service is too much different from just doing a regular PC game, so I don't think it would really delay anything by months. But I know nothing of game development so who knows..
Anyway, right now, it's a cool idea, it's got some hurdles that they think they've managed to, well, hurdle, and until the Beta players start talking and then eventually it's released, we won't know for sure if it'll work or not.
Right now PC gaming is for PC savvy people and that's about it. Sure, I'm smart enough to work my way around issues that arise, but you can make a lot more money selling your games to the much larger audience of people who aren't that technical about it.
Look at Nintendo - with the DS and especially the Wii and the simple yet entertaining games available, they've made $$$ by appealing to the casual market first. My wife and I have friends picking up the Wii for stuff like Wii Fit, and these are people who otherwise had zero interest in games of any kind outside of maybe solitaire or Pogo.com.
If you can get people to play PC games without having to purchase a crazy expensive machine, and without having to spend an hour or two browsing random internet forums to find an obscure method of making your game work, then why the hell not try it?
Sure, right now we should all be in the "believe it when I see it" stage to some degree, but damn it sure would be nice to see this get off the ground and work the way they say it does.
Other commercially utilized remote clients such as Citrix, X11, RDP, or VNC don't perform nearly this well. If anything, Videoconferencing companies should be showering them with money to license their technology.
From a gaming side, what would be more amazing is when companies don't have to develop games that skew towards the lowest common denominator, making games for the highest end hardware without compromise. I can forsee games that model each and every bit of environmental change (ala Molyneux's original promise with Fable) and MMOs that are more action oriented without concern for ping times between client and server.
With some current and future mobile phones able to process 720p video, a mobile client would effectively circumvent the limitations (other than bandwidth caps and whatnot) of the platform. While twitch games may not be suitable (lack of buttons, spotty connection, etc), imagine playing something like Pixeljunk Monsters or Dawn of War II on your iPhone/WinMo/Android phone, with multiplayer options.
Edit: But I don't want to put the cart before the horse. Let's all hope the beta works as advertised.
A few points, first, you are right, this kind of capability would have a lot of uses. Which brings up a good question: why would you start with such a hard case, i.e. action oriented video games?
Second, lag will be worse as far as ping times go, but MMOs could be helped by removing client lag in crowded areas.
Third, console games, especially HD ones are closer to being budget limited than they are to being limited by the hardware. If this OnLive led to more revenue for developers that would do more to make room for more in depth games than the hardware improvements would.
(Please do not gift. My game bank is already full.)
I don't know about you, but I'm making split-second decision all the time in Dawn of War 2, being a fraction of a second too slow leads to losing entire squads, or not getting to victory points fast enough.
It would also make the game hideously irritating since it has to pause until all players are synched up again, extra lag would NOT be a nice thing.
I would assume that their choice of market is because: 1. Enterprise software vs Commercial software development is tough and yields little reward since the audience is small 2. It's just more fun.
What I meant in terms of client/server connections in MMOs is not the connection between the user and the Onlive servers, but rather, the client and server inside the NOC itself. What if an MMO developer were to put their servers next to Onlive's servers, allowing the client computers direct access not over the Internet, but within the LAN?
I agree that the cost of producing a high definition game is a limiting factor, that's why I didn't mention increasing graphical improvements in future games. Making games more in-depth (say, in terms of advanced AI interaction or variables in choices) as you and I agree will require more that just GPUs, but it does require more in processing power and in disk space, something that a netbook can't provide locally.
Agreed, that was a bad example. I was just thinking of a RTS and since that's what I was playing recently, it immediately came to mind. A better example would be a Strategy game, such as Disguea or a turn based RPG. Even something like Valkyria Chronicles would be nice to play on a phone, provided something could be done about readability of text and whatnot.
The one thing I haven't even seen mentioned yet is that there is at least 80 ms of latency just to start off. Then you have the round trip latency from your onlive box to the server located somewhere in the US. For me on dsl to ping google is about 70 ms, so lets estimate 100ms. Now you have 180 ms of latency in any game. That's 11 full frames in a game running at 60fps. You also have the regular control latency on games which I think averages around 2 frames these days, which normally doesn't matter, but in this case it's added to a higher number. Now we've got around 13 frames of latency from when we see the result of our action on screen.
With that kind of delay this system really isn't going to work well for anything that requires reactions even close to real time. The best case scenario for this company is probably that Time Warner will move from being an investor and just out-right acquire them. You will then be able to play checkers, chess, and settlers of catan on your cable box.
Even complicated turn based games are out given the 1:1 scaling ratio required on the backend because of the technical requirements of the games, compression, and streaming.
"Read twice, post once. It's almost like 'measure twice, cut once' only with reading." - MetaverseNomad
If anything, if this service does take off and lots of people are playing it, this might make the ISP's more pissed off because now alot more bandwidth is being used, and you know how lazy these shits are in terms of upgrading their infrastructure, and instead will enforce harder caps.
Xbox Live: Kunohara
I mentioned above that I think this service will only see profitability if bought by Time Warner, or another cable company. In that case it doesn't matter as it will be done over TV cable.
However if they somehow do come to market and people actually use it to buy/rent modern games from it, then yes the caps will be a problem. Comcast, Qwest, and Charter all have caps. I'm sure that many other ISP's do as well. Playing these games over them will really speed up people hitting their bandwith limits.
Also something I didn't even take into account in my previous post is traffic shaping. If done to the packets for this service, it will add even more latency to the game. In the US this isn't too much of a problem yet, however Canadian and British gamers probably would find the service unusable if it found the ISP's ire.
"Read twice, post once. It's almost like 'measure twice, cut once' only with reading." - MetaverseNomad
If I had to stare at a screenful of blurry text all day, ow my head
Can trade TF2 items or whatever else you're interested in. PM me.
There is no way to remove the ping time between your computer and theirs. The controls will ALWAYS lag, and it will often be a noticeable amount. This isn't like networked gaming where lag is taken in to account by the server and all this prediction stuff comes in to play, this is introducing a timeout between manipulating your interface and what you see. Can you imagine playing a twitch shooter where every movement you make is a tenth of a second behind what's on the screen? You're dead before you even see what killed you.
I'm pretty sure this is going to end up just like the Phantom...impossible to pull off in the real world.
EDIT: ALSO...
Cloud computing works on the idea that not everyone uses peak resources, so you shift the load around. Games always use the full resources of the system. You can't have two people using the same computer to play crysis, nor can you spread it out over time, as gaming happens at very specific times during the day. It will be very hard for them to economically have an affordable subscription that is actually usable and won't find you waiting behind a huge queue when you want to play.
3DS Friend Code: 2707-1614-5576
PAX Prime 2014 Buttoneering!
This is good. A more in-depth and descriptive version of the scaling problems I mentioned.
This is much different than the Phantom though. The Phantom was essentially a computer that ran nothing but Steam that you hook up to your TV. It's not a terrible idea. It was a little ahead of it's time, but the main problem was that the company was run by very corrupt people who just wanted to steal VC money.
"Read twice, post once. It's almost like 'measure twice, cut once' only with reading." - MetaverseNomad
I'm not saying it resolves the issue, but it does seem to answer that one point in some small way.
I think we will have to wait and see before we know exactly how hard it's going to be to control
Of course, caution and skepticism is fine, even warranted in this situation, but you have to admit, the prospect of this technology is exciting, and if it works, it'll have effects beyond the game industry
Also it seems like a lot of people didn't catch that presentation they did. It should be up on youtube or whatever by now, so everyone should go watch it, it answers a few questions at the very least
PSN ID : DetectiveOlivaw | TWITTER | STEAM ID | NEVER FORGET
There is no such magical server hardware that allows one to have the processing powers of multiple gaming rigs for less than the price of multiple gaming rigs.
Especially considering that games are designed for consumer hardware, not server hardware.
EDIT:
"Speculation on how many milliseconds of lag" is grounded in reality. There will be a minimum network latency no matter what. This is just how the internet works. Data must travel between many points before it arrives at its destination, and each point takes time to process it. You can remove all the latency on the server side, and client side, but the network latency can only be removed by placing the servers closer to you. Which is fine if you live near a peering point like Atlanta, Dallas, or LA. It will SUCK if you live in Minnesota and want to play games.
3DS Friend Code: 2707-1614-5576
PAX Prime 2014 Buttoneering!
That's what they were saying though, if I understood correctly
Using these weird custom built servers using powerful GPUs and shit, one server can run multiple games at once, since running a game doesn't require all the processing power of a single computer or GPU or whatever
Look, I don't know science, but I got the distinct feeling that they have some idea of what they're doing
But again, none of us knows what the fuck
edit: yes I know jon but some people are saying 22ms and some people are saying 180ms
I don't know but there seems to be a bit of a discrepancy there
PSN ID : DetectiveOlivaw | TWITTER | STEAM ID | NEVER FORGET
Here is Tycho's writeup:
Personally this whole thing seems comparable to being told that someone has used phychophysics to unlock cold fusion and is starting up a company to make lawn mowers that don't ever need to be refuled, and would you be interested in investing in said company.
(Please do not gift. My game bank is already full.)
Reading and listening to their stuff, it seems that games that use fewer resources, they'll run multiple instances of the game using VMs and such, but they're still running regular hardware. For "big" games, they'll have to using a one-to-one server thing. I still think they're going to run into bottlenecks in the timesharing based on the narrow timing of people playing games in the afternoons.
For lag: As an example, my home internet happens to have the same "home" as my TF2 server. I get an average of 10ms latency for bare pings between here and there. It's less than 15 miles away, and only touches four routers. It just gets worse from there.
3DS Friend Code: 2707-1614-5576
PAX Prime 2014 Buttoneering!
The company with the most lawnmowers is the company that wins.
http://blogs.msdn.com/shawnhar/archive/2007/12/14/network-latency.aspx
Obviously, prediction algorithms and shit effect that stuff, and the concept is that they will be close enough to you to handle stuff. From here at work in Reston, VA, pinging the Speedtest.net server in DC, about 15 miles away over a Verizon Business T1 averages about 250ms roundtrip, but there is a lot of traffic going in and out of here, so I guess I won't be gaming from work.
Using VNC to remote into my home system, I am getting an average of 105ms roundtrip to the same server, which is probably about five miles closer there, and this is the result of the speedtest.net test on their page:
So it's a pretty fast connection.
None of this speaks to the type of latency we'd see with this service, but it does go to show that even just jumping across the river to throw out a ping can take a tenth of a second.
TIME TRAVEL IS NOW!
WHERE WE'RE GOING, WE DON'T NEED ROADS!
(1000 miles)*(1609.3 meters/mile)/(3e10 m/s) gives 5.37ms. And that's just for one-way, with no switches, on a perfect slightly cloudy day.
P.S., I hope I don't offend anybody with my lack of significant figures.
1ms is just the time to encode the video and push it down some fat pipes.
The 1ms claim is the time it takes to encode the stream (a pretty significant claim in itself), not the latency between the client and server.
Edit: Since my post is redundant *Shakes fist at agoaj*, I'll address something else:
As jon stated (and myself in a earlier post), even if they managed to get racks of servers with multiple PCs on a board, you can't timeshare all of them. High end games were built to take up as much resources in a single computer as possible for maximum performance. Thus you run into a scenario of several thousand people attempting to run Crysis and having to scale your network center appropriately to meet demand. You can cut costs to heat and power by deploying more efficient design (aforementioned PC on a board + physical case layout), but it's still a 1:1 relationship for remote computers and each client has a pretty high X associated cost that may be hard to profit from. I don't doubt they can do it (nothing appears shady within their ranks), but the business model, along with the ongoing latency issue does seem suspect.
Cloud computing means the information is stored on their servers, you don't have anything. It means major privacy issues. Everything that you do in a game, and also in your private life would be at someone else's reach, potentially. This poses a huge risk I'm not willing to take.
Also with the fact you DEPEND on them for even getting anything or playing any game at all, whenever you're bored, it means that they can charge you far higher than if you had bought a high end PC. This is like X-Box Live but worse, pretending to save money but really you're paying $50 a year for the past x years, it adds up. This would be far worse, you'd burn through the same amount as a high end PC. And don't think that games wouldn't be tiered. You want to play a hot game with great graphics on release? That's extra than playing 1 year old games on medium graphics. Of course they'll do it.
This also means that small companies won't be able to get their name out there, a lot of word of mouth is generated by sharing CDs with friends, playing a free demo etc Already you might have heard of the PS3 really crippling small publishers by charging for all bandwidth, a small publisher simply cannot get its demo out on the PS3.
You have to realize that this cloud computing thing means you give all power to someone else. This from a consumer point of view is not very good at all.
What is really the ONLY benefit? Really, the only one : No more need to upgrade computers. Do people really understand how the pricing will rip us off? Do people really care about games on demand? You can already have that with services like Gametap anyway.
Think this cloud computing, cloud gaming thing through, personally I will never ever buy into such a scheme, ever.
As for upgrades, am I mad that I can't play Empire : Total War? Sure. But I'm still loving it playing old games again, or maybe Drakensang and it's lower requirements
So,
* what do the switches and ISP's add?
* assuming optical backbones, what is the cost of fiber<->copper conversions? (edit: gotta be negligible, right?)
* how much does the typical household router add?