Idea No. 1: Semi-Dumb Device + Split OS Design
Imagine a semi-dumb device: it has touch screen, microphone and speakers (plus, perhaps a couple of buttons for power and volume). By itself it doesn't do much. However, connect it to the cloud, and it's a fully powered smart device. Now imagine it's optimized for networking (like a TV) that its battery can last much, much longer -- days, weeks or even months. Sounds good? Read on...
The semi-dumb device provides an instant-on experience. You can use biometrics to wake it up or to unlock it. Once your identity is verified, the device then connects you to the cloud where the "real" OS is. All exchanges with the remote OS are secured. In a way, it's like remoting to the cloud-hosted OS, but the session's experience can feel more local to the device.
For example, you can rotate your device and the remote OS renders a re-oriented screen for you. You launch GPS and the remote OS knows where you are. You attach a peripheral, the remote OS knows. You snap a photo, and it's stored in your cloud-hosted drive.
The truth is, the semi-dumb device has a mini-OS of its own. Half part of it is designed to boot the device, and to enable operations that are local to the device like hardware level security, I/O, networking, peripheral/state detection, etc. However, the other half of what it does is optimized for communicating with your remote OS session. Thus, the semi-dumb device can behave most of the time like a TV that is tuned in to a single channel: your device's remote session with the cloud-hosted OS.
The semi-dumb device exchanges states with the remote OS. The remote OS calculates through the device's states in order to run programs, launch applications and/or render the appropriate UI/UX. All, if not most, computing power will be on the remote OS. The semi-dumb device is there to mostly show the results and to exchange more I/O.
This split OS design can let the device use standardized or proprietary data exchange and communication protocols with the remote OS session. The protocol can define how initial states and deltas can be exchanged efficiently, reliably and securely.
This split OS design can re-invent the camera app, which can be 50% local and 50% remote, with the more CPU/RAM intensive processing delegated at the remote OS's side. The split OS design can also re-invent the GPS app, which can be 50% local and 50% remote, with the more CPU/RAM intensive processing delegated at the remote OS's side. These are just examples, and I'm sure more apps can be re-invented to apply the split design.
Most important here really is how "lightweight" semi-dumb devices can be. They will not need bigger or better CPU, RAM, storage and battery. Therefore, they can be made smaller and cheaper, making it possible for semi-dumb devices to be friendlier to the environment. They can also be designed to be share-able such that, using biometrics, whoever's handling the device can access their own unique remote session from the cloud, for example.
Dumb terminal? Client/server? Zero/thin client? Cloud-hosted OS? The semi-dumb device is potentially a revolution in computing. It seems quite smart to start R&D for a semi-dumb device and to back it with a split OS solution. It is possible. Someone just needs to prototype it and to start selling it.
That is, if being online all the time is reliable, stable, and fast... really fast.
The semi-dumb device provides an instant-on experience. You can use biometrics to wake it up or to unlock it. Once your identity is verified, the device then connects you to the cloud where the "real" OS is. All exchanges with the remote OS are secured. In a way, it's like remoting to the cloud-hosted OS, but the session's experience can feel more local to the device.
For example, you can rotate your device and the remote OS renders a re-oriented screen for you. You launch GPS and the remote OS knows where you are. You attach a peripheral, the remote OS knows. You snap a photo, and it's stored in your cloud-hosted drive.
The truth is, the semi-dumb device has a mini-OS of its own. Half part of it is designed to boot the device, and to enable operations that are local to the device like hardware level security, I/O, networking, peripheral/state detection, etc. However, the other half of what it does is optimized for communicating with your remote OS session. Thus, the semi-dumb device can behave most of the time like a TV that is tuned in to a single channel: your device's remote session with the cloud-hosted OS.
The semi-dumb device exchanges states with the remote OS. The remote OS calculates through the device's states in order to run programs, launch applications and/or render the appropriate UI/UX. All, if not most, computing power will be on the remote OS. The semi-dumb device is there to mostly show the results and to exchange more I/O.
This split OS design can let the device use standardized or proprietary data exchange and communication protocols with the remote OS session. The protocol can define how initial states and deltas can be exchanged efficiently, reliably and securely.
This split OS design can re-invent the camera app, which can be 50% local and 50% remote, with the more CPU/RAM intensive processing delegated at the remote OS's side. The split OS design can also re-invent the GPS app, which can be 50% local and 50% remote, with the more CPU/RAM intensive processing delegated at the remote OS's side. These are just examples, and I'm sure more apps can be re-invented to apply the split design.
Most important here really is how "lightweight" semi-dumb devices can be. They will not need bigger or better CPU, RAM, storage and battery. Therefore, they can be made smaller and cheaper, making it possible for semi-dumb devices to be friendlier to the environment. They can also be designed to be share-able such that, using biometrics, whoever's handling the device can access their own unique remote session from the cloud, for example.
Dumb terminal? Client/server? Zero/thin client? Cloud-hosted OS? The semi-dumb device is potentially a revolution in computing. It seems quite smart to start R&D for a semi-dumb device and to back it with a split OS solution. It is possible. Someone just needs to prototype it and to start selling it.
That is, if being online all the time is reliable, stable, and fast... really fast.
Comments
Post a Comment