Friday, 20 January 2012

OnLive Cloud Gaming Technology For Hosted Virtual Desktops - An Analysis of the Technology and Business Model That Underpins the Platform

I first started speaking to Paul Weinstein, a Senior VP at OnLive a few months back, we were discussing the cloud gaming technology they had developed that is capable of streaming the most graphics/video/audio rich applications you can think of (high-end video games) over the WAN to users in their homes onto any device using HTML5, but I had little idea that they were going to launch their technology in this way, this soon.

When I first saw a demo of the OnLive technology streaming AutoCAD over the WAN perfectly I was blown away, nobody in our space has ever been able to deliver applications like AutoCAD to hosted virtual desktops users over the WAN like that, let alone latency free versions of the computer games I play on my console. Now OnLive is doing it and they've launched the OnLive Desktop at CES and its AWESOME !

If you go to their website you will find next to no information about this, consider the CES launch their coming out party, OnLive just signaled that they are in the hosted virtual desktop and application space and that they are here to play, but they are not quite ready yet and their platform is after all still a gaming platform. If they were ready, the leading tech journalists would all have a demo desktop to play with and they do not, they are bitching about it.

I have been and always will be a gamer, it's how I switch off after a long days work, I log on and kill foul mouthed children online (or rather be killed by them), I have played computer games for as long as I can remember starting with my Sinclair ZX Spectrum and cassette loaded BASIC games, through to the Atari, then onto the Nintendo NES before moving onto a Sega Genesis. I ignored the first PlayStation and got a Nintendo GameCube instead (fail) and then moved straight onto the first Xbox when it came out, ignored the PlayStation 2 and stuck with Xbox until the Xbox 360 which I still use now.

More recently over the last few years I have also become an avid Steam gamer (TF2 and L4D2 yeah !) which is kind of like a cloud gaming platform in that all of your games are backed up into the cloud, but they still have to be downloaded onto your local machine before you can play them, not a pure cloud play like OnLive.

I am proud to be a good gamer, I can compete and I will own your ass at any game you can name, especially on Star Fox, TF2, MW2 or Mario Kart.

The problem with pure cloud gaming plays as I see it has always been latency, gamers as a group are the most latency intolerant application users you can comprehend, we generally play (at least the more hard core amongst us) first person shooters online with other players and there is absolutely nothing worse on this earth than having an enemy dead in your sights, your finger on the trigger, when things start lagging, it kills game play and this is doubly so for online multiplayer first person shooters.

There is nothing on earth more frustrating than losing a kill because of lag and console games lag sometimes even though you are effectively running a local copy of the game on your console, but playing the game over the network to fight others.

This is the primary reason why gamers have not moved across to pure play cloud gaming platforms en masse, we cannot handle the lag (latency) involved in waiting for that app to push its refreshed display down the pipe over the WAN with any delay whatsoever. But then OnLive came along and changed all that, I tested their gaming platform extensively for free (thanks Paul) by playing Batman: Arkham Asylum until I completed the game and I am pleased to report I suffered little or no lag throughout the entire game, although admittedly I do have a decent internet connection, but still, miraculous.

This has never been possible before, ever, but here OnLive were doing it and the only bad thing about their service is that they did not have enough game titles to keep me interested, that was it, my only negative comment about the service. I remember at the time thinking 'this is friggin awesome, I wish they had more games and I wish my friends were playing on this platform so I would have somebody to play with'.

The AutoCAD demonstration is awesome but I did not get any hands on time, plus I am not familiar with the application really, but 3D rendering of graphics and architectural design worked well for sure.

I did think at the time though that if these guys actually get to the point where they can deliver enterprise applications and desktops in the same way they deliver computer games, nobody would be able to touch them and here they are doing it now at CES.

From the conversations I had with OnLive back then, they were looking to partner with organizations who were operating in the enterprise desktop and application space already, this made sense to me at the time. They had this great technology with better display presentation over the WAN than anyone else and it was only a matter of time before they stopped messing around with AutoCAD and got serious about desktops and enterprise applications, it made sense to me that they would license that technology out to those who were already delivering apps and desktops to enterprise users, through the channel or direct partnerships or however, but I did not expect this.

They were after all a gaming company, what do they know about delivering hosted virtual enterprise applications and desktops I thought ? In retrospect this was more than a little shortsighted of me, from a technology perspective at least, if they had high end computer games over the WAN licked, then they sure as hell can do any enterprise app or desktop we in the space can throw at them and they are.

After their announcement, the penny finally dropped. They do not need a partner, they have the entire platform to deliver Windows 7 desktops and applications live over the WAN perfectly already built, it looks like a gaming platform but with a little imagination you can see it delivering apps and desktops just as easily. Sure it still looks like a platform a gamer would use, but once their UX people spend a little time developing it into a more 'corporate' looking platform they are in business.

Autodesk, the company behind AutoCAD were one of the original investors into OnLive, it makes sense that they had exactly this in mind the whole time. They have cloud-ready DC's capable of storing, synchronizing and rendering high end graphics applications ready for delivery over the internet, they have a presentation virtualization technology capable of delivering their applications to PC's, Mac's or any iOS/Android device, they have a wireless joypad technology which would be really easy to convert into a wireless keyboard technology for use with tablets.

Later investors include British Telecom, Belgacom and AT&T, all major international telcos and I can tell you that these guys are not interested in computer games, even if they are willing to distribute them in their respective regions bundled with their broadband packages, nope these guys have something else on their mind.

I am the first to call it and I may be wrong, but I think OnLive is a hosted virtual desktop and hosted virtual application company in stealth mode, masquerading as a cloud gaming company, I am watching Steve Perlman and Paul Weinstein now and shaking my head in admiration, because NOBODY saw this coming and I read enough blogs and tech press covering our space to know that nobody saw this move from OnLive coming, I mean they are a gaming company right ? I was talking to them about this months ago and even I did not see it coming at the time.

So let's assume they are a hosted virtual desktop and hosted virtual application provider in stealth mode, let's treat them as such and take a look at them from that angle starting with their technology before moving onto the moves I would make if I were them.

Take a look at the video below of OnLive Founder and CEO Steve Perlman giving a demo, he KNOWS their tech is better than anything Citrix or VMware have got for throwing apps and desktops over the WAN, he makes some digs at them as the presentation progresses, so let's start right there and take a look at the individual pieces they have put together to complete the puzzle.

Display Protocol

OnLive have their own proprietary video compression hardware that works in collaboration with standard GPU and CPU chips, you need to have one of their proprietary chips on the endpoint device to really get a great user experience, it's why they send you a dongle to plug into your TV, that chip needs to be present and this should ring some big bells with the more tech savvy amongst you.

It reminds me of PCoIP and no I do not mean the imitation software version of the same name that VMware licensed and have spent so much money developing, I am talking about the real Teradici PCoIP which is STILL the best way in the real world that I know of to deliver a 'true PC' experience over the WAN under low bandwidth, high-latency conditions. Same drill, Teradici PCoIP needs a proprietary chip on the endpoint device (they now embed them into sweet HD Samsung monitors) in order for the display presentation to work properly, hardware driven accelerated video compression is the name of the game here and software cannot reproduce this hardware based display presentation no matter how many millions of dollars are pumped into it, I think VMware are waking up to this if they have not already. You gotta have decent hardware driven video rendering on the client to do it properly over the WAN.

By the looks of things, OnLive have their own version of the Teradici PCoIP tech that performs as good as if not better. Now I know a lot of you will not have seen the REAL PCoIP out in the wild much and no its not a VMware tech, they just licensed it because they needed something to compete against Citrix HDX with. My own company tuCloud have been a longtime Teradici partner since their very early days and I am telling you now that if you want a real high definition user experience, forget any software based technology, you absolutely need the hardware acceleration on both sides, the server and the endpoint, nothing else comes close.

Teradici can still deploy better looking, feeling, performing desktops with high end applications (yes AutoCAD) better than XenDesktop or View could ever hope to, but of course hardware is expensive unless its produced in volume and software is (comparatively) cheap so they had to try right, hence the VMware/Teradici partnership, the thinking being let's take the best display tech out there and reproduce it with software.

You need a chip on the endpoint that talks to a chip on the server, Teradici have one in the REAL PCoIP and so do OnLive with their proprietary video compression chip. They can display apps and desktops better than VMware or Citrix, full-stop. OnLive are doinitright.

Datacenter & Network

This is one of my favorite subjects and I wrote about this a year ago, except I did not mention any specific technology providers in my article, although if you are tech-savvy enough you would have guessed I was talking about Juniper cloud fabric. See the thing about all this next generation cloud DC and networking tech is that only one company has actually gotten this right and that's Juniper. I was pretty amazed that Cisco, a networking company, could screw up the idea of what a cloud DC and network does, but am not going to go into it here, if you are interested do read about it here.

To summarize, the 10G cloud switch fabric revolution has arrived and it's not being delivered by Cisco, this is Juniper tech and only Juniper are doing this properly. I had this explained to me by a personal mentor of mine and the man I consider to be the father of modern day cloud, Mark Medovich, the sole patent holder on VMwares vCloud technology who is now working at Juniper Networks as their Chief Architect. Mark thinks that for a networking company, Cisco got the idea of cloud completely wrong and he went to work at Juniper doinitright.

Why am I talking about Juniper? Because OnLive partnered up with Juniper back in June 2011 to upgrade their entire DC's, all five of them scattered across the US. Think of your cloud DC wet dream and this is what OnLive have built, they have the new Juniper cloud fabric, the switches, the gateways, the routers and all underpinned by Junos operating system for managing the whole thing on a single platform. The OnLive DC setup is my cloud computing wet dream, partly why my own company tuCloud is partnered with Juniper right now instead of Cisco, they do cloud DC and network better and the father of modern day cloud is helping them do it. Juniper recognise that with the 10G switch movement, the network becomes a viable bus in a DC, the network and the DC are one and the same, THIS is what cloud is about, scaling out as well as scaling up.

They have five tier-3 DC's scattered around the US, crammed full of the most advanced technology out there, 10gig switches, optical networking, you name it they got it. Simply put, very few cloud providers on earth have a DC/network setup as advanced as OnLive. When it comes to cloud DC and network OnLive is doinitright.


I can only talk about their platform from a gaming perspective here, I have only tested it by selecting and playing games on it so far but I can tell you (as you can see in the video) that desktops and applications can be just as easily delivered using this platform, instead of selecting a game to play you would simply select your desktop or just a standalone app.

The whole platform is HTML5 based, yes that HTML5, the killer of Adobe Flash and beloved of mobile gamers and fancy web designers everywhere. We are starting to see this technology in the hosted virtual desktop and hosted virtual application space from Ericom and more recently from DinamiQs, although VMware have been having a go at HTML5 with their AppBlast and Octopus projects since early 2011, its HTML5 that powers OnLive's application delivery platform and let's them put their hosted virtual applications and hosted virtual desktops onto any device, anytime, anywhere.

They could not have chosen a more forward thinking tech to base their platform on, HTML5 is clearly the way forward, I was lucky enough to be treated to a presentation of DinamiQs HTML5 based cloud platform in December 2011 by their founder and CTO Erik Westhovens and I am telling you it's bad ass.

If you had to choose a next generation technology to base a cloud platform on and deliver hosted virtual desktops and hosted virtual applications off, it would be HTML5. Once again OnLive are doinitright.

So Much Doinitright But What Are They Doing Wrong?

Now we have talked about how OnLive are doinitright, have ticked all the right tech boxes and built themselves the dream cloud hosted virtual desktop and application platform, datacenter and network infrastructure with their own proprietary presentation display hardware tech, let's take a look at their business model and see how it fares when it smashes up against the cold harsh reality of the enterprise desktop and application space.

First of all this HAS to be an enterprise play, if they are talking about Windows 7 desktops there exists no Microsoft licensing mechanism currently in place for OnLive to sell their service to a retail market, not unless they want to do terminal services like all the other (already extinct) wannabe hosted virtual desktop providers out there. The only way this works with Microsoft Windows 7 desktops is if they build PRIVATE (non multi-tenant) desktop cloud infrastructures and their customers have Microsoft volume licensing or software assurance in place.

I don't care how good their platform is, they bang up against the same Microsoft licensing wall as the rest of us do in the real world, you cannot rent a proper desktop OS in virtualized form and deliver those over the WAN to people who want to rent ten of them by the month, Microsoft do not let you and I wrote extensively about this in March 2011 in my analysis of Microsoft's licensing strategy, which is still batshit crazy if you ask me, but it's their market dominance to lose.

My own company tuCloud provides thousands hosted virtual desktops to the Federal government but we cannot rent five to a small business that could really benefit from them, Microsoft do not want us to and unless you are a server desktop provider deploying slices of server OS and pretending they are desktops to your users, you never will be able to.

Microsoft do not seem to be budging on this one at all in any way and if Citrix are not allowed to do this, neither are OnLive.

So we are assuming they are going after the business desktop and application market here and the conversation continues until they hit the glass wall erected by a dogmatic enterprise space where NO disruptive techs are ever used unless they have VMware or Citrix stamped on them.

The truth of the virtual desktop and virtual application space is that there are LOTS of disruptive tech companies out there vying to get their techs into the enterprise desktop space, but bang up against the glass wall that is corporate America. I am generalizing here, but unless a technology comes from a tier-1 provider, it does not get any traction in the corporate desktop space, ask Wanova, moka5, Teradici, Virtual Computer or any other number of smaller disruptive players with really smart tech that would burn Citrix and VMware if they were given a chance.

The general consensus is that Citrix and VMware buy the best of the bunch, so it's best to wait until they do and then use them, this is an attitude I do not see changing anytime soon in the enterprise space.

There are so many barriers to entry for a new tech into that space because so many enterprise level organizations have a predetermined list of criteria that a vendor must meet in order to ever bring their technology into production. Things like size and stability of the vendor, annual sales although my peers and I have never understood how that translates into solid infrastructure design or architecture.

OnLive are going to have to fight the same hard fight that everyone else does when trying to enter the corporate desktop space and even worse they are a gaming company, can you imagine the fun Citrix and VMware sales reps are going to have poking holes in a computer game company when it's mentioned by senior corporate execs?

Then you must consider if they are actually geared up at all in any way to properly support desktops, applications and users, I find this highly unlikely and reading through OnLive gamer forums talking about this, they had to delay the release of a bunch of new games in order to dedicate engineering resources fully to the desktop launch.

Hosted virtual desktops used by business users need support, 24/7 telephone support from a person who knows why Office crashes all the time, worse upgrade to a new version, takes like 2 service packs before it's bug free. Word will crash half way through a mail merge, Excel dies when print preview, PowerPoint crashes when you insert huge vector images accidentally into a presentation, these things all need support and lets not even go there with Outlook, viruses and everything else users delightfully conjure up.

Any sane business would need much more support than OnLive's existing gamer customer base needs who mainly get it through a support forum and rarely live, I think OnLive will work through partners to get their technology into the enterprise and smooth the path to traction, they just are not geared up for anything else.

Can Everyone Stop Being So Impressed With Windows 7 and Microsoft Office on iPad Please?

Seriously this is nothing new, the ENTIRE hosted virtual desktop space has been offering this functionality for at least 2 years, Citrix Receiver and VMware iPad View client can do this beautifully and so can a whole bunch of other techs, it's called consumerization of IT and we can even deliver desktops to Androids believe it or not. Get over it!

I noticed in all the day old articles covering the OnLive Desktop that they made a big show of their tech working on tablets and I just yawned, they are like the rest of the desktop space was two years ago, super impressed at their new toys and eager to demonstrate them working on tablets.

The rest of us have long realized that this is nothing but a gimmick to show prospective customers who are all wrapped up in their shiny fondleslabs, at least those who have not savvied up in the last two years as most technical buyers have. Streaming Microsoft desktop apps to an iPad does not a revolution make, but a little bit of tinkering and the technology could be repurposed magnificently.

Let me be clear, NOBODY does any serious work of a Windows 7 desktop that is being displayed on a tablet, no matter how good your pinch and zoom is or how good your wireless keyboard is, it's a gimmick to show in demonstrations but practically useless and I do not want to deep dive into this argument right now, but those of you who actually have hosted virtual desktops they can access on an fondleslab and tried to do any real Microsoft app work will get me.

I think this is a reflection of the innocence OnLive have upon entering OUR space and crossing over from the gaming world into the enterprise desktop world, it's like they are babies with a shiny new toy, but still not quite sure of its positioning or place in the world yet, although they do know they have something big, really big that could potentially disrupt the rest of the space around it.

We shall see and boy would I love to be Steve Perlman right now, the gamers call him The Emperor and with a name like that how can you not like the man? The gamer community respects and admires the man even if they are feeling a little ignored at the moment.

The Emperor Napoleon springs to mind and his rule that beer was for the boys, whisky for the men and brandy only for champions.

I am betting that Steve likes a good cognac.

But the gamer in me screams WTF are OnLive doing? Why o why are they getting into the enterprise desktop space when they can OWN the online cloud gaming space? It galls me to see their wonderful shiny platform with absolutely no decent games (bar Batman: Arkham Asylum) or gamer community around their platform to speak off. Sure they have a hundred titles, but nobody wants to play them and your little sister already completed them.

Why are you delaying the release of new games and pissing off your entire customer base who are beginning to feel like guinea pigs, this could have been done a lot more smoothly, days after their CES launch analysts and journos still complained that they cannot get a hands on.

WHY would you want to sell boring (in comparison) desktops and boring applications to office workers when they can be selling games to the gaming world? In my mind they cannot do both things well at the same time, they must choose between either dominating the cloud gaming space or going after the enterprise desktop and application space and you can argue that there is more money in desktops over the long term, but have you seen how much money is in gaming?

The worldwide gaming industry is the fastest growing component of the international media sector, worth almost $50 billion in 2011 and predicted to grow to over $70 billion in 2012. That's a billion dollar market that's there NOW, not some early adopter market full of vicious competition and potentially game changing technology, they have a much bigger market in gaming to go after NOW and much less competition really.

Forget the other pure play cloud gaming companies, they are pikers, the real competition is good guy Gabe and his Valve/Steam monster and don't think for one second that the guy does not realize that the internet is way smarter than any of us. He is sitting there right now, billions in the bank looking at OnLive, their tech and thinking they are not really serious about the cloud gaming space, how can they be when they are about to begin trying to sell hosted virtual desktops and hosted virtual applications to office workers in a market that is still nowhere near mainstream enough to make serious money from yet.

As a gamer, I feel completely let down by OnLive, they have so much potential and so much to win in the gaming space, wtf are they doing in our space messing around with hosted virtual desktops and hosted virtual applications even if they do have some pretty slick technology?

As a startup (they are a startup after all) you cannot do two hugely ambitious things at the same time and go after two huge markets at the same time no matter how much financing you have, a startup company needs to focus on its core market and core business which begs the question, what is OnLive's business?

Are they a technology provider like Citrix or VMware now? Are they a hosted virtual desktop provider like tuCloud or Desktone now? Will they get bought, oooh the drama!

I wonder if they have worked it out yet, and if they have, if they are sure of themselves. One thing is for sure, they have in their hands a big fat shiny diamond and whatever they decide to do with it will determine their company's (and investors') fortunes for the long term.

I hope to god they choose gaming, I REALLY do not want to have to download games or buy expensive consoles anymore and I REALLY do not want to have to compete in the hosted virtual desktop space against that kind of tech, at least until I get my own HTML5 rig up and running in the next few months and other services suck.

Whatever they choose, I wish OnLive the best of luck and I am really looking forward to seeing what their next move will be as well of course to test their desktops like every other tech commentator in the space.

Originally written by Guise Bule, tuCloud Chief Executive Officer and Fist of the North Star at the Desktop Superhero Alliance.

SOURCE: tuCloud Blog.


  1. This comment has been removed by the author.

  2. Hosted virtual Desktops can be used for using applications, files and remote working from a variety of internet enabled devices.

    Hosted Virtual Desktops

  3. Bluehost is ultimately one of the best hosting company for any hosting services you need.