In May 2011, for ZDNet's 20th anniversary, Scott Raymond and I wrote a series of articles that attempted to predict what the personal computer of the year 2019 might look like. We called it Project Blade Runner in homage of Ridley Scott's classic 1982 science fiction thriller, which in itself is considered a major film for inspiring futurists.
In the almost two years since the publication of those articles, much has changed in the personal computer and mobile technology industry. When I re-read my original piece, it's clear I did not take into account a number of important trends in semiconductor technology, as well as the speed of cloud services adoption.
Let's go over what I feel is a more accurate assessment of what the future of personal computing is likely to look like 10 years from now.
In the original piece, I talked about the concept of a "Hub", or a modular platform for device design, essentially a Lego brick systems architecture.
While I still think this is technically feasible, I believe it is unlikely that we will end up using powerful, expandable systems with large amounts of localized processing and storage in 10 years. Many if not all of the things we think about as "localized" computing resources will be instead distributed into the cloud.
Instead, I see the mobile device as the "Hub" and the center of the end user's computing experience. I also think the issue of expansion and equipment preservation will be completely anachronistic in 10 years' time.
People will own smartphones that have powerful systems on a chip (SoCs), which will be completely disposable.
Additionally, the very idea of system upgrades in 2023 will be completely foreign, because all computing devices will be completely sealed and unrepairable.
I do think the computing experience will be modular, but not in the way Scott and I originally thought. Instead of "Lego bricks", I see smartphones, tablet screens, and keyboard housings sold as a matched component set, much in the same way as stereo and home entertainment systems used to be sold.
I believe it is unlikely that we will end up using powerful, expandable systems with large amounts of localized processing and storage in 10 years. Much if not all of the things we think about as "localized" computing resources will be instead distributed into the cloud.
Earlier this week, I spoke a bit about Canonical's Mark Shuttleworth and his vision of having a complete system of integrated smartphone, attachable tablet touchscreen, and keyboard.
What will make this possible is a concept I am referring to as "Unified Computing", whereby the smartphone, tablet, and desktop operating system becomes the same platform and software development target.
I see a matched set of smartphone "brain", 10-inch to 12-inch tablet touchscreen and keyboard/docking station costing around $500, with an expected system lifetime of two to five years.
Many of these systems will be sold by wireless carriers and will be subsidized with data contracts, so there will be a high incentive to replace them as frequently as possible.
Unified computing will also enable device manufacturers to consolidate their product lineup.
With the exception of Apple, which will undoubtedly continue to make form factor changes at a highly accelerated rate, I see companies like Samsung using technologies like mobile device virtualization to release matched component sets that will run a customer's choice of mobile operating system, whether it be Android, Windows, BlackBerry OS, Ubuntu, or any other system that may come along.
With the exception of the software that runs on the smartphone "brain" component, Samsung, HTC, LG, and any number of companies can refresh a single lineup of matched component sets once a year, and not require 10 different models, all customized for a specific OS platform as they are today.
With this strategy, they can offer all of the platforms, with the same exact matched components.
Need a Samsung Android phablet? Buy the basic 5-inch smartphone "Galaxy Brain" and order it from your carrier with Android pre-loaded. You've decided that you need Windows on it instead? Click on an icon, and the firmware for Windows will be downloaded to it.
Need a 13-inch touchscreen display and keyboard so you can do productivity tasks? Buy the matching display unit and keyboard. Need a cheap 27-inch 4K or 8K monitor to do creative content work? Just buy one, and use the high-speed remote display wireless link capabilities built in to your device. It will just work.
You need wearable computing capabilities? Buy a wireless eyewear or wristband peripheral, which by 2023 will be vendor neutral.
So what happens to all of your personal data and storage when you migrate between matched component systems? Nothing. It lives in the cloud.
Regardless of what device you have, it's always there, and cloud federation services will allow you to mix and match as well as move the data back and forth between cloud storage and service vendors as you need it.
In the original "Blade Runner" article, I talked at length about applications and storage. Much has changed in the way I think about these two concepts since I wrote that piece.
Originally, I thought legacy applications, ie windowed desktop apps designed for the x86 architecture, would be run localized using a specialized processor. While legacy applications will still exist and still be very important in 2023, they will no longer reside on the local machine, which will be ARM based and only run full-screen mobile applications installed from an App Store.
Instead, they will run in the cloud, using desktop as a service (DaaS).
There are a number of reasons why this is likely to happen. First, many ISVs are going to move to a subscription-based model, and will encourage their customers via strong pricing incentives to move their application data to remote versions of these apps running on their desktop clouds, which will run on re-branded hosting provider and public cloud infrastructure by some of the biggest names in the business.
Private cloud versions of these legacy desktop apps will also be an option for those enterprises and institutions that warrant it, and require the strictest form of isolation, away from highly populated multi-tenancy public clouds and with much higher SLAs attached to them.
Once boxed, one-time purchase software goes the way of the dodo — and it will, despite the kicking and screaming that is happening today — these legacy Windows applications will all run remotely but display locally, using technologies such as Microsoft's RDS/DirectAccess/App-V, Citrix's XenApp, and VMware's Horizon.
And, of course, along with the remote apps and the desktops, the vendors providing these services also include integrated data backup, threat and anti-malware monitoring and elimination, routine maintenance and application upgrades, and also quick provisioning and re-provisioning, if required.
Besides Microsoft, Citrix, and VMware, it's also possible that we will see other desktop session protocols emerge in more common usage, such as Red Hat's SPICE, and it would not surprise me at all to see Apple develop its own version of remote desktop in iCloud for use with ARM-based "Macs".
Running remote applications will provide the same level of desktop fidelity and capability with discrete GPUs that exist today, even with the most demanding workloads for content creation. Powerful servers running in the datacenter using virtualized GPU clusters will render complex 2D and 3D workloads remotely and pass the information over the internet to the local ARM device.
None of this is science fiction. There are ISVs that are already doing this, and the technology to accomplish what I have described has been proven and has reached a fundamental level of maturity that will see rapid adoption of subscriber DaaS applications in the next 36 months.
While there may still be a need for certain niche apps to run on localized x86 systems, the classic high-end PC workstation and Macintosh will almost completely disappear.
Will the PC of 2023 be your smartphone, connected peripherals, and your cloud? Talk back and let me know.