Project Blade Runner: Putting it all together in 2019

Project Blade Runner: Putting it all together in 2019

Summary: How far do you think we are from our 2019 vision of the personal computer?


An artist’s rendering of a modular ARM-based Dell BL 2020. (conceptual art courtesy Spidermonkey, Inc.)

By now you're familiar with the Blade Runner Project, a concept developed by Jason Perlow and myself. We tried to envision what the personal computer of 2019 would look like eight to ten years from now, based on existing technology and continuing advances.

We tried to stay as far away from experimental or predicted concept technology. Our forecast has its roots in today's tech. And while our current tech isn't quite capable of what we've outlined, nor as powerful, much of it could actually be built today.

Also Read:

In his article, Jason delved deeply into the technology and systems architecture behind the concept. In my approach, I will be discussing how this new personal computer design would be used by the consumer.

The basic building block of the Blade Runner PC, or BRPC, is the hub. By itself it is a standalone computer, able to handle most of your computing tasks without need for additional hardware. Of course, if you play games, or record and watch a lot of video and music, you might want to expand with an extra storage module and discrete video adapter.

2 terabytes of storage may seem like a lot, but if you're recording video at Blu-Ray quality, that's 50-100GB of data for a 2-hour movie. If you use your computer as a digital video recorder, that storage gets used up pretty fast.

The design is deliberately geared toward ease of use. While there are plenty of people that like to tinker, most consumers do not. They just want their computer to work. To that end, expansion modules simply stack on top of the hub, allowing all stacked devices to operate on the same system bus and use the same power source.

Discrete video adapters may require more power, and thus would have their own power bricks. Since we do not foresee graphics adapters being made smaller over time, we estimated that the adapter would likely require a larger heatsink and fan in the module, possibly even residing next to the main computer stack rather than on top of it--this would definitely be the case if you used multiple GPU units to play intense video games at a high frame rate.

As more devices in our household are replaced with network-capable ones, it becomes easier to control them from a central location--in this case, the BRPC. Your TV, your stereo, household lighting, even your coffee maker, fridge and oven. If the core system isn't enough to handle all of the tasks you throw at it, you could simply add another hub module.

As described in Jason's first article, the entire BRPC performs as an as-needed cluster. Running on minimal current, individual CPU cores and banks of RAM don't actually start drawing power and working until the rest of the system needs those resources.

This is very similar to the way the Toyota Prius hybrid car operates: the vehicle operates on battery power, the engine kicks in only when needed, and then the engine shuts off again when it is no longer necessary.

Ambient noise from the BRPC is also something that will have been reduced to a minimum. With the exception of a discrete video adapter and a DVD/Blu-Ray module, nothing on the unit should make any noise at all. The CPU modules run at low power and generate little heat, which is easily offset by a simple heatsink. The storage devices are SSD--solid state drives--so there are no moving parts there, either.

Connecting external peripherals will be made in one of two ways, either by a next-generation Thunderbolt cable, or by 10 Gigabit Wi-Fi. The Wi-Fi connection, which is at least 20 to 50 times faster than anything wireless that is in use by consumers today, will be fine for any handheld devices, printers, and remotely-controlled appliances around your house.

The Thunderbolt connection would be for high-speed video connections, such as your 3D 4K monitor. You're going to want that kind of bandwidth when you're pushing video that's four times the size of HDTV 1080p.

One of the other aspects of the BPRC is being able to access your data from anywhere. Now, it's obvious that plenty of people have their email in the cloud with Gmail and cloud storage services like Dropbox. But what about private data? These days, it's becoming clear that the cloud isn't really that secure unless you encrypt your data so that even the hosting provider doesn't have access to it.

Enter the personal cloud. You have all of this storage at home, and unless you're familiar with setting up your home network so that you can access your data from anywhere, you don't want to be bothered with the trouble of learning. (If you think I'm wrong, ask your grandparents why their VCR flashes 12:00--or why they even still HAVE a VCR.)

The idea for the personal cloud stems from two scenarios. One, to access your data from a central location no matter where you are. Two, to have a backup of your critical home data in the event of a disaster.

I had originally thought of the concept of using a fireproof home safe to backup all data from your BPRC, but the material required often inhibits wifi signals from penetrating it while maintaining a cool, fireproof environment that protects your data. It might be possible using a physical Thunderbolt connection that would likely melt in the event of a fire.

We decided that the data equivalent of the Japanese capsule hotel was exactly the kind of solution we needed. You could head over to your local FedEx-Kinko's and rent space in their facility.

They would plug in a tiny storage module with a web front-end that you could access from anywhere, and could be configured to sync to your central data at home. Ostensibly you could buy your own storage module and configure it for them to plug in, or you could buy one of theirs.

Next: More secure, less mess »

You would not be tied to a single operating system. Since everything runs under a hypervisor similar to the VMware vSphere Hypervisor, you could run iOS/OSX as your primary OS, and run Windows and Linux applications seamlessly on the same desktop through a separate Windows Virtual Machine launched specifically for those apps.

Many users are afraid of viruses, trojans, and being taken advantage of by malware on the internet because they don't know how to watch for it or are unwilling or unable to learn how.

The security method for the BPRC system is simple: All operating systems are static. They cannot be written to. When an application is run, it uses a copy of the OS to execute. The operating systems do not communicate with each other directly. And when the application closes, so does the native OS it used.

One added benefit of a system like this is that it forces application developers to write self-contained applications that have everything they need within their own software directory, instead of spraying files all over the hard drive like a broken firehose.

These days, more and more people are using portable devices as their primary computing devices. I myself use an Android XOOM tablet for most of my work and play, resorting to the laptop when I find my tablet browser isn't compatible with a specific website, or when I want to use specific applications that simply aren't available on tablets.

The BRPC lends itself to working with portable devices on a more integrated scale than ever before. Over the home Wi-Fi network, your device automatically syncs with your central system, updating data, documents, files, pictures, videos, music, and anything else you've configured to maintain.

Put your tablet or smartphone down on the desk, and your induction charging place-mat recharges the battery. This place-mat also provides power to your wireless keyboard and mouse.

I know that previously some folks took issue with the concept image of using a laser keyboard projected onto the desktop. Keep in mind, this was artistic license taken by our talented conceptual designer. Most people, myself included, really prefer the tactile feedback of a physical keyboard. A solid desktop doesn't lightly push back against your fingers the way the keys of a physical keyboard do.

It's important to note that while futurists like to make a big deal about voice interaction and hand motions to control computers, the keyboard has existed for over 200 years since the invention of the typewriter. Voice dictation and voice control of computers has really only been viable within the last 20 years. It certainly hasn't wiped out keyboard usage, and I doubt that it will for many decades to come.

I know that some folks, like me, also have a problem with cables. A computer, monitor, printer, router and cable modem already create a tangle of cables. The design of the BPRC is meant to shed most of those. One power cable for the hub and modules stacked on top of it. One power cable for the monitor. One thunderbolt cable from the hub to the monitor. One power cable to provide power for the induction desktop.

The modem and router will likely be one unit with its own power brick. No more network cables, USB cables, charging cables for devices. Or for cats to chew on.

When we first started discussing this project, it was important to strike a balance between what existed, what was possible, and what would be coming down the road. I tended to be more conservative than Jason, automatically discarding anything having to do with stuff like quantum computing.

Everything we put into the system design exists today in a more primitive state. Sure, there might be a capability for terabit internet speeds, but there is no way that the telcos and the cable providers are going to pay billions to replace their entire infrastructure when they can continue making money hand over fist on their existing networks.

It's easy to speculate about what the future holds for personal computing technology. We think we're actually very close. What we put together here is meant to tie it all together so that it integrates and interoperates the way we want it to, instead of the way it doesn't quite do it the way we want it to right now.

We had a lot of fun putting this together. I hope you had as much fun following us along for the ride.

Topics: Operating Systems, CXO, Hardware, Mobility, Networking, Storage, Telcos

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • I see it a little different. I think that package-on-package technology

    will get so cheap that all displays and devices will have their own computer with various types of wireless networking out of the box. It would be nice to have a standard for the connectors so you could easily upgrade the computing module.
    • RE: Project Blade Runner: Putting it all together in 2019

      @DonnieBoy Essentially, this is what was designed, we started with a powerful multicore SoC with integrated RAM. If you added a storage module, you'd essentially have the "Hub" or the base architecture. Everything else we proposed simply builds upon that.
      • Ok, but, I think they will also include storage (flash) in the package-on-

        package. It might eventually come in the size of an SD card. It would be nice to have a standard like there is for SD cards.
      • RE: Project Blade Runner: Putting it all together in 2019

        @donnieboy In the original article, I talked about how completely self-contained devices as you describe would likely be the most popular version of how the [b]systems architecture[/b] would be implemented. With this architecture, we are not just accounting for regular end-users, but we are also accounting for creative content professionals, programmers, engineers, etc that have higher requirements. The architecture is also extensible to the datacenter.
  • Also, multiple operating systems for consumers are not necessary. They need

    a web browser and a standard run-time for native applications. Maybe Portable Native Client (PNaCl) will emerge as a standard run-time for these kinds of devices. With Oracle in control of Java, it looks like Dalvik will be out.
    • RE: Project Blade Runner: Putting it all together in 2019

      @DonnieBoy We cannot predict the outcome of that litigation. But the judge has already asked that Oracle's claims be limited to 3 patents, not 120.
    • RE: Project Blade Runner: Putting it all together in 2019

      @DonnieBoy Remember that for the most part, as I explained in the first article, we're not really talking about multiple operating systems in the same sense as we run things like VirtualBox or VMWare Workstation today. We're talking about a "Shell" OS that manages the applications and the user interacts with (which could be some form of Windows or Linux or even "Mac"), and the applications THEMSELVES are self-contained in VMs that run on their own JeOS with their own libraries, so they could be written in anything. The JeOS could be Microsoft's Midori, it could be iOS, it could be a stripped down Linux kernel, etc.
      • How about if the VM includes the ability to convert intermediate code to

        local machine code? So, the VM virtualizes the interface to hardware, and, you could have an OS compiled to intermediate code (like for PNaCl), the OS could be generic, run anywhere. Of course it would have to adapt to different screen sizes and input methods.

        Applications DO need fool-proof sandboxing, maybe running in a jeOS would be the best way to do that.
  • Um, no thank you

    First of all, while this might fly for consumers, business users don't want a computer on their desk that looks like it landed at Roswell. This is only an outward design and can be changed, but your artist's rendering will not sell in the enterprise.

    Secondly, we have seen attempts at modularity in the past, and none of them really took off. With components getting more powerful and cheaper, I would expect a more appliance approach in the future. You buy a cheap little box that connects to your peripherals, or, as Donnie suggests, you buy a display that has the basics built in. Why add the cost of a modular system, which may end up costing as much as the computer itself?

    I also believe that there will still be more traditional computers, with all of their power and complexity, for those of us who will still need or want them. Sure, there will be many who only want a browser and will live in the cloud, but there will be other applications that still run locally. While the market will change over the next 10 years, there will still be more traditional systems and applications running along side the new.
  • RE: Project Blade Runner: Putting it all together in 2019

    Man, I want this concept rig soooo bad right now. I am simultaneously excited for something like this to come out and depressed that it isn't here already. Why did you have to write an article on such an awesome concept! DAMN YOUUUUUU! :)
  • Message has been deleted.

  • Blu-Ray

    I doubt Blu-Ray will be around in 2019...
    • Blu-Ray in 2019

      @rynning Perhaps, perhaps not. DVD technology is over 10 years old and is still here. There's definitely going to be legacy stuff hanging around. And not everyone will have bandwidth capable of downloading Blu-Ray quality video over the internet for a long time to come.
      Scott Raymond
      • RE: Project Blade Runner: Putting it all together in 2019

        @Scott Raymond
        However unlike DVD, Blu-Ray's replacement was here before Blu-Ray even came out. The quality of current downloadable movie content is is also very high. Seriously, download a high def. movie from AppleTV. It is excellent. Videophiles may argue that it isn't as perfect as Bluray, but the difference is nearly imperceptible to most other people.
      • Re: Blu-Ray in 2019

        @Tigertank The issue isn't about the quality of downloadable content, it's about bandwidth. Many people around the world do not have bandwidth capable of downloading high quality video much less stream it. The US government is busy defining broadband as 4mbit+ download speed, which just isn't good enough.

        The telcos and cable companies are fighting to keep from upgrading their networks to be capable of handling the increased load; instead they are choosing to implement data caps instead. At 250GB a month, a handful of high quality movies will use up your entire allowed capacity.

        Until the bandwidth providers are forced to upgrade their networks to match demand, physical media will still be necessary and widely available.
        Scott Raymond
  • Second Building Block

    Your Bladerunner conceptualisation and the release of the Chromebook suggest to me a second building block for 2019: I'm going to call it Solaris for reasons I hope ... you have already started to guess.<br><br>The difficulties in deciding between a Chromebook, iPad, netbook and portable are caused principally by the bundling of screen and processing capability: so continuing your modular design our first decision is to split the two ... resulting in a processing wafer and a screen wafer ... which do not produce a working Solaris until plugged together.<br><br>The size and connectors on the processing wafer are industry standards. There are various monitor wafers which all swallow up the processing wafer on connection to produce a tablet-like device. Now one can choose a monitor size to suit, as we can for desktops currently. We can also choose whether to have a browser based experience (Chromium, Micro$oftium or go on $afari etc.), a cloudy network terminal experience (RDP, CITRIX, etc.) ... or should we plug in the wafer which is our mobile phone: an autonomous, offline computing experience.<br><br>As with your Bladerunner design there are three chances of this happening: fat chance, slim chance and no chance. However Google, not being a hardware company, may be that global corporate with the only motivation to make such modular designs work (that is outflanking all their competitors like Apple, Microsoft, IBM ... simultaneously).<br><br>You didn't get the choice of Solaris? Classic science fiction film plus network terminals designed by SUN. Keep up with the programme please ;-)
  • More computers designed by ID Software

    Seriously it looks like the shipping container in Quake. Computing will not change much in concept but it will in speed and electrical consumption. Look for disk-less hard drives and removable flash to replace all types of ROM. Computers will look like they do today only smaller. Not like a Rubbermaid inspired container.
    • RE: Project Blade Runner: Putting it all together in 2019

      Rubbermaid is right. I was thinking the same thing.
  • RE: Project Blade Runner: Putting it all together in 2019

    If i understand correctly the concept is dependent on the fact that the off chip communication can be power efficient, fast and at the same time "plug and play" like.

    I can also imagine the complexity of the software which can support such scalability in hardware.