In 2018, nearly everything is in place for a descent into a Ready Player One-style dystopia. The way things are going, we should be destitute and beaten down by climate change long before the 2040s. All we’re missing, for now, is the technology. People are already more than happy to spend huge percentages of their lives hooked up to alternate realities (social media, MMO games, binged television) but these are all rudimentary compared to Ready Player One’s immersive, hyper-lifelike OASIS VR universe. Those looking to escape the grim, cash-squeezed drudgery of day-to-day life through fully-immersive VR will have to wait, for now, until someone invents a way to get 75% of the country’s population on the same platform.
For this weeks Giz Asks, we reached out to a number of VR experts about the plausibility of an OASIS-like platform coming onto the market, and how much computing power would be required to sustain it. Pretty much all agreed that, tech-wise, we’re not even close—but none disputed the idea that it could exist, at some point down the line, and many were sure it would, especially now that certain tech moguls have taken up the cause. By then—if our natural disasters and market crashes continue on schedule—the technology should be ready for a marginally less awful life in virtual reality.
VR Laboratory Manager, National Institute for Aviation Research, Wichita State University
If you want a black and white answer: no. But I would say we are about 80% there. Any kind of VR universe would still employ the common principles of any online game, where the bulk of the content is downloaded onto the local machine, and the only data being sent back and forth is player position and interactions with the world. Now as more people get on, the more data there would be—but a gigabit internet connect would be able to handle that. The main limitation is downloading the content. You don’t really want to actually download it, since updating something like on every person’s rig, would be a nightmare, and the world could be thousands of terabytes of data. And those downloads would easily kill a home broadband connection, mainly since the ISP, which oversells their capacity, wouldn’t be able to support such downloads by that many people.
The technology to fix this is starting to be developed now, with network companies like Cisco developing what they call “The Fog.” Similar to traditional cloud solutions, but localised across the network. So your neighbourhood might have a fog device with the VR world setup on it, and all it is doing, is streaming to your local device. The actual use case they are going for is VR sporting events, which would require a live stream of a VR feed (most likely a 360 video stream to start with) to thousands of devices at the same time. Traditional hosting services would be hard pressed to keep up with that demand. Instead with a fog setup, the game would be loaded/streamed to the fog device from a central server, and then that would distributed down to the user. It’s basically the Walmart distribution model, but for gaming.
Director of Research and Development Integration, Institute for Creative Technologies, USC and project leader of the Integrated Virtual Humans group
There are a couple of challenges there. First of all, the computing power to realise such a detailed world—the solutions we have right now for both VR and AR really are in the beginning stages. You need a lot of local computing power to create something that’s compelling, that people can see, that have a wide field of view, that can track what the real user is doing and compare that to what the character is doing in VR.
What becomes interesting is indeed what kind of server infrastructure you’d need, and there are a lot of current games that provide an example. Think an MMO like World of Warcraft, for instance—there’s so many millions of users on it, and what they do to keep track of all those users is divide them up in different realms, so per server you only have a few thousand people playing. The moment too many people come online, you can actually notice that it lags.
There are only a few games that allows all the players to be in the same gameworld; I think the best known one is EVE Online. It’s a game set in the future, and you are basically the captain of your own spaceship. They have a lot of interesting methodologies to allow all those people to share that same virtual world at the same time without putting them on diff servers.
One of the things they do is, if too many people come in the same location, they slow down time, so the computers have time to catch up and calculate everything that’s going on. I think in order to really realise that, you need to be smart about what you compute with. To get everybody in the same virtual reality world, the connection needs to be enormous. There are lots and lots of challenges trying to do it right now [with just a few people], let alone with billions of people.
The Ready Player One dystopia also shows an enormous gap between the rich and the poor, resulting in a very unequal balance on how resources are shared. It seems likely that even if energy is very scarce, a VR infrastructure will be developed for the wealthy regardless. Once the infrastructure exists, companies will try to find ways to target customers who have less disposable income. To what extent that would be feasible depends on many factors. Technology can be a great equaliser, but there certainly are situations in which expensive tech divides the rich and poor even more.
Director, Medical Virtual Reality, Institute for Creative Technologies and Research Professor, Dept. of Psychiatry and School of Gerontology, University of Southern California
The dystopian future portrayed in Ready Player One may be due to a dramatic reduction in available traditional fossil fuels, leading to the collapse of the use of internal combustion engines as a practical method for transportation, with concomitant poor planning on society’s part to create a feasible alternative energy source. Thus, those circumstances would make it nearly impossible or extremely costly for people to travel and navigate the real world for work and play.
That situation may actually have given rise to the need for in-home experiences to keep the population engaged and quelled, since serious individual travel as we know it would be off the list of options for everyday folks.
Thus, the OASIS may have emerged from a need for people to have experiences without leaving their living rooms. The energy consumption to run the OASIS may be significantly less than having billions of people driving to work, to the movies, to parks, to visit friends, to go to bars, etc. So the OASIS may have risen as a solution to address the energy crisis, and as a form of social control to keep people from thinking too much about what they were missing in this world of limited real-world options.
Director, The Game Research and Immersive Design Lab, Ohio University
I think the problems with server/internet connections could be significant, but not insurmountable. To get a sense of the current obstacles, all you have to do is look at bandwidth requirements for online games and realise you will have to scale that up for VR. VR environments require higher fidelity graphics and network connectivity than games, streaming video, and most other forms of digital media. A laggy experience would be a deathblow to any VR experience.
I think the challenges with creating a Ready Player One “Oasis” scenario are quality and reliability of network connectivity. I am e-mailing you from the Ohio University main campus in Athens, Ohio. The actual campus has one of the fastest and most reliable network infrastructures in the state. If you go two miles outside of town and beyond, you will find large quantities of households that are lucky to get a dial-up Internet service. Those households won’t be able to participate in the Oasis. And of course, without a healthy, protected energy grid, the fastest internet speeds are always going to be somewhat unpredictable.
I don’t think currently the biggest concern about someone creating the Oasis is network or energy reliability, but something more fundamental. VR-ready computers aren’t cheap. The headsets required are expensive and uncomfortable and may need to be replaced often. You also need to allocate space in your house or workplace to comfortably use VR without fear of tripping over something or running into walls. We have a long way to go in making VR, and something like the Oasis, sustainable. Despite all of this sounding pessimistic, I am actually quite optimistic about the future of VR. Prices will come down. Reliability will improve. If someone creates the Oasis, it could revolutionise the way we play, learn, and work.
Professor, Computer Science, University of Nottingham
A concrete example [of what we’re discussing here] would be an online music festival in which, say, an audience of a hundred thousand avatars joins together at a virtual stage to watch virtual performers and sing and dance along together as part of a ‘mass live’ experience.
This is a tricky proposition, and perhaps the biggest technical bottleneck is the internet. Can we squeeze enough data down its pipes to make this festival scenario possible?
Each participant at our festival delivers a stream of movement and audio data—let’s assume around one megabit per second, just for the sake of argument. If every virtual festival-goer sees every other one live then each of 100,000 people receives 99,999 streams from the others. That’s 99,999,000,000 MB per second flowing around the network and making its way along your local internet connection to your headset. Hmmm, that doesn’t seem so possible.
Of course, just sending every avatar every other avatar’s data stream may be a pretty dumb way to approach the challenge, but the point is that the amount of available data we might potentially need to ship around the network more or less grows with the square of the number of participants involved.
A better solution might be to send each participant only the data that they really need to be able to enjoy a compelling live experience. What might this be? You’ll presumably want a high fidelity view of any performers. Then you’ll want some live interactive feeds from a relatively few people ‘nearby’ in the crowd —a mix of friends, family and a few interesting strangers perhaps —so you can sing, dance and talk with them. And finally, you may require a general impression of the behaviour of the wider crowd—a sense of their movement, noise and mood. Of course, the system will need to be very smart about personalising your feed according to where you are standing in the virtual crowd.
We call this approach ‘Level of Detail’ in VR. You don’t need to see and hear everyone in the crowd at the same (full) level of detail all of the time. After all, that’s how it works in the real world. I can see that there is a person on the other side of the stadium and perhaps even see that they are dancing, but I can’t read their face or make out the individual words they are saying.
Lecturer, Games and Visual Effects, Staffordshire University
A low resolution 360-degree experience, as is available for most current headsets, requires at least 25 mbps. However, for resolutions comparable to HD TV, this jumps to 80-100 mbps. Further, a 4k experience needs as much as 600 mbps. We also need to bear in mind that the current global average connection speed is 7.2 mbps, so, we have a way to go in terms of infrastructure supporting the kind of VR seen in Ready Player One.
Having said this, Facebook is making good headway into solving these issues. They are doing this by using clever algorithms and techniques such as Field of View streaming, which reduces the amount of bandwidth required by around 25%. This technique concentrates on decoding only spaces within the field of view of the user. In addition to this, they are utilising a 360 degree video encoding technique called pyramid geometry, which reduces file size by 80%. Also, by using cube maps, Facebook has also reduced bitrate and storage without affecting video quality.
There is no doubt in my mind that the associated problems with delivering these experiences will be solved very soon. Human beings are using social media platforms at an increasing rate. The trouble with current mobile devices is that they mainly require text based communication. If the problem of freely available VR experience was to be solved, then people will eagerly engage in this more fluid, interactive experience. Clearly, at least in the short term, the problems of VR deployment around the global internet will be solved by just such clever algorithms as have been mentioned. Concurrently, Infrastructure will continue its evolution, and follow up behind these otherwise novel solutions.
Professor, Computing and Intelligent Systems, Ulster University
It is not a simple matter of extrapolating bandwidth from today’s demand into the future. For instance, just over 10 years ago, there were only around 2 billion connected devices. However, by 2020, we expect there to be over 50 billion. Yes, the Internet of Things is a real thing. Of course, not all devices are equal. Video is the bandwidth killer. Video by 2020 will consume more than 80 per cent of all internet traffic, with every second nearly 1 million minutes of video content crossing the internet. That is competing with any World Wide VR platform.
We also have a long way to go from where we are currently according to Akamai’s 2017 State of the internet report which states that the global average connection speed is 7.2 Mbps. It is actually a 15% increase compared with one year prior. At a country level, South Korea has the highest average connection speed at 28.6 Mbps while Singapore has the highest average peak connection speed at 184.5 Mbps.
If we ignore AR experiences which require slightly lesser bandwidth and attempt to calculate the bandwidth requirements for proper immersive VR on a network, we need to take into account the screen resolutions of the headsets. The leading VR headsets at present typically are OLED with resolutions of 2160 x 1200. We can ignore for now that 360 degree 8K, 90+ FPS HDR stereoscopic displays are just around the corner. Unfortunately, each VR stream must also be duplicated twice to stream individually to both eyes.
To experience a low resolution of 360 degrees VR, we would require at least 25Mbit/s for streaming. For the ‘retinal’ 360 degrees experience, VR would require around 500-600Mbit/s. Most network engineers use the rule of thumb of Virtual Reality requiring 5 times more bandwidth than HDTV if you wish to create an immersive user experience. We also have to factor in the latency aspect. Latency is the delay from input into a system to desired outcome and basically affects how responsive our network experience is. Lower latency is better and ideally we want to have motion to photon (MTP) latency of less than 15ms.
Of course, in Ready Player One there is a fossil fuel shortage, so some may doubt that we could run a global scale VR network. At the moment, a typical VR ready PC uses about 220 kWh. If you entered the VR world for about two hours a day, then this would cost about $300 per year. That is reasonable. We also know that the electrical efficiency of computers doubles roughly every year and a half, so as computing power has increased over time, the amount of energy needed per computation has gone down. It turns out that improvements in energy efficiency are driven by the same techniques engineers use to make microprocessors more powerful, including cramming more transistors onto a chip. This is likely to continue, so we can expect power requirements to drop—therefore, a global VR network is still feasible even with an energy crisis.
In some ways the numbers are moot. What we find is that the internet is organic. It has repeatedly found a solution to problems of capacity and scale. We now have network backbone giants like Cloudflare who have 15 Tbps Capacity and 151 global Data Centres and route much of the traffic online. A content delivery network provider like this with regional nodes combined with futuristic technologies such as Edge Computing, 5G and P2P 3.0 could very easily meet the capacity challenges of a world-wide VR paradise. Yes, the reality of the situation is virtual.
Professor, Communications, Stanford, founding director of the Virtual Human Interaction Lab and author of ‘Experience on Demand’
In my new book, I argue that the roadblock to great networked VR such as “The OASIS” is interactional synchrony. In the 1970s, psychologist Adam Kendon, working at the Bronx State Hospital, uncovered a principle that continues to guide much of the research and the applications involving nonverbal behavior today. What Kendon discovered was that body movements—subtle shifts in posture, eye-gaze and gestures—not only dance to the rhythms of speech, but they also perform in reaction to the movements of others. In fact, these movements, both subtle and obvious, are correlated between people at a highly intricate level.
Ever since Facebook’s purchase of Oculus in 2014, creating hardware and software that can reliably achieve a high level of “social presence” for all users of VR has been the technology’s holy grail, and many dozens of companies, small and large, are currently attempting to solve aspects of this fiendishly difficult task. And it remains one of the most difficult problems in VR—how can you put two or more users in a virtual space and allow them to interact in a human way, with each other and with their virtual environment? How do you capture and convey the subtleties of human social interaction, in the movements of the face, in body language, in eye gaze? Once again, the challenges posed by VR remind us of the richness and complexity of our human experience, because to understand how to make our avatars feel real, we have to know what we humans are doing—consciously and unconsciously—that makes our daily encounters in real life feel real.
If virtual travel and telepresence is going to replace physical travel, it will have to devise a system that allows virtual people to be as nonverbally rich as their hosts. “The OASIS” will need to replicate in virtual bodies the complex choreography of body language, eye movement, facial expressions, hand gestures, and physical touch that occur—often unconsciously—in real life social interactions. In my opinion, the roadblock of tracking and rendering gestures is higher than any other problem related to computing power.
Professor, Electrical Engineering & Computer Science, York University
Imagine building some sort of interactive VR space similar to the OASIS in Ready Player One. Having a single space in which everyone on the planet can interact with everyone on the planet (or 75% of everyone on the planet interact with 75% of everyone on the planet) is not practical with today’s technologies. The Earth’s population runs at about 7.3 billion, so 75% of that is around 5.5 billion and having that many agents interacting in a common simulated place is just beyond our ability to do the networking. But observe that for most of the OASIS as depicted in the book and the film this is not necessary. Consider how massively multiplayer online games like Destiny work. Players do not interact with all players all the time. Rather they interact with a small group of players and this ‘micro world’ can be simulated separately from other micro worlds and the effects of each micro world integrated into a common database when required. This is what makes games like Destiny practical. In Ready Player One, the OASIS is split up into 27 sectors, each sector into zones, and so on. So until everyone wants to show up one one corner of one planet to take down IOI we can break up the simulation so that we don’t have to stretch networking and centralised simulation resources that much.