If you measure the power used by a netbook PC when it’s in standby, a frozen unusable state, it uses around 500mW of power. Leaving a netbook on with WiFi connected in an idle state with the screen blanked, maybe with an email program polling occasionally, you’ll see about 10 times the power usage. In PC terms, 5W is impressive but if you go to the smartphone world and take a look at the figures there, there’s a huge huge gap that needs to be tackled. As smartphones become tablets, become smartbooks, there’s a threat that ‘always-on’ becomes ‘must-have’ and that X86-based devices will struggle to compete in casual computing scenarios.
Smartphones are designed from the ground-up around the concept of ‘always-on. From the moment a smartphone is conceived, every element of the design has to be checked for power consumption which is why a smartphone can sit connected to the GSM telephone network drawing power consumption levels lower than 50mw. That’s 1/10th of the power consumption taken by a good netbook when it’s in a frozen state. Impressive.
But what happens when you connect a smartphone to the internet ? You can use cellular data services to achieve a good rate of connectivity by switching on UMTS for example. Switching to WiFi on a typical smartphone brings faster connectivity and, in a lot of cases, lower power that can be done on cellular networks. A smartphone can run a multitasking operating system and remain connected to internet and voice networks in well under 500mw of power, the same as it takes a netbook to sleep. In fact, the best smartphones are running in this configuration for over 24hrs on a 5Wh battery which is an amazing 100mw of power usage. Turn on some background internet activities and it will jump to an average 200mw!
What happens if you take an ARM platform that’s in the same processing power category as a low-end netbook. Put it in a 10 inch screen form factor and do the same test? Actually, it’s the same as a smartphone. The only difference between a high-end smartphone and a smartbook with a 10 inch screenand a huge battery is the screen itself and when that is off, there’s practically no difference at all.
To prove this, I took one of the most powerful mobile computing platforms in a large 10 inch form factor device with 32GB of storage, 1GB of RAM and integrated WiFi. The device has a 23wh battery (about half that of ‘6-cell’ netbooks. The device is the Acer Iconia Tab A500 Android tablet which runs honeycomb. I connected to the Wifi (at 54mbps) and left the device connected with the screen off while it did it’s stuff in the background. Marketplace checks, email checks, Twitter checks and even some GPS usage by Google Maps. A weather service was running, the volume was set to silent and Bluetooth was turned off.
Over a 48 hour period with a few screen-on moments for checking progress (and a 10 minute in-use period as my daughter grabbed it to use a paint program) I measured 46% fall in battery usage of which 5% was due to screen-on time. Take away the screen-on figure and you have 209mw of power usage. The Acer Iconia Tab is nothing more than a smartphone inside!
‘Always-on, Always Connected’ will be a ‘Must-have.’
Always-on tests are interesting because it’s a hands-off test that people think only applies to idle smartphones. In fact, it applies to many computing scenarios. With location, polling, sync, presence, alarms, push updates and of course, cellular voice and messaging becoming the norm in the hand, they will also have appeal on the desk. Not having to wait 5 seconds for a machine to start-up, another 5 seconds for a Wi-Fi connection and another 10 seconds for tweets, emails and other features to catch up is annoying. There’s also a bunch of other screen-off, connected activities that are interesting. Servers for example. By that, I mean computers and gadgets that serve information to the Internet. This doesn’t just cover web servers. Think about internet-connected weather stations and web cameras, in-car data storage and notification systems. Then there are the devices that just don’t need big screens; Connected musical instruments. Digital cameras with 3G. Internet Radios. Low-power internet connectivity is important for these devices.
Related: Social Netbooks and ARM’s Lock-In Netbook Opportunity.
Summary
The point here is not to highlight that ARM is better than X86, it’s to highlight the gap. This gap is currently a huge advantage for ARM-based platform designers.
- The first point is, if manufacturers using X86/PC architecture don’t get products to market with active standby soon, with the help of Intel (the only X86 player trying to tackle this problem) customers will have a chance to experience, and may not turn back from, ARM-based always-on products.
- The second point is that this is a screen-off gap. Current screen technology is killing ARMs advantage in the ‘in-use’ scenario where screens are larger than 7 inch. It reduces the ARM advantage from 20x in idle to about 4x with a 10 inch screen being backlit. When the devices CPU is being actively used, the advantage drops even lower to around 2x. [Acer Iconia Tab 4W. Samsung NC210 8W)
- Finally, the Acer Iconia Tab A500 is a good example of low-power internet connectivity. It’s likely that other devices in this ARM tablet segment hit the same figures.
Keep an eye on high-end ARM-based platforms over the next year or two. Honeycomb and iOS are leading the way into the professional space with their software and application ecoystems and you might find that this always-on advantage starts to lock people in soon.
The problem is that “always on†and “always connected†will always be less efficient then the alternative. Of course, I want to be able to get in and out of my house, but I don’t want a big hole in the middle of my walls. Thus, I have a door. It allows me to have a hole in my wall whenever I need it which is big enough to pass through, and it magically reseals. In fact, my door comes with advanced security/privacy features like a lock which makes make it possible for me to chose who else can create the hole and when can they can do it.
You’re kind of cheering for the Internet to be like television WAS. If your TV wasn’t always on then you were missing out on the constant stream of content being pushed to you. However, TV became better than that when we added computers to them which provided PVR functionality. The PVR became a more humane interface to the TV content. Why would I want my mobile device to be stupider than that? I agree with you. I want a mobile device that CAN be always on as efficiently as possible, except it’s not as efficient if it MUST be on all of the time. ARM hardware is great. I’ve been a developer for a long time and occasionally consult on some really neat embedded projects. The software support for the embedded side of ARM is great, but the software on the consumer handheld side is such a pathetic joke that I don’t even bother subsidizing it with my cash. ARM hardware is not incompatible with Personal Computers, but it seems like the software should always be. If basic open source drivers were available for popular ARM hardware, then something like Debian with xmonad and uzbl would humiliate today’s “tabletâ€, just ask the people in the Efika MX community.
The simple fact is that I’m going to power my device off all of the time. If I am in a meeting then my devices are absolutely off. If I don’t feel like being around radiation then my devices are absolutely off. If I’m in a studio with sensitive equipment then my devices are absolutely off. If I just feel like it then my devices are absolutely off. Clues for the sensible approach for computing can be seen in the way that RSS feeds and software updates happen. When you’re ready to have more content pushed to you, then go check for updates at the location where they’re published. I really think that ARM hardware could have a great future (although, like others, I think that the diversity of ARM licensees could be ARM’s downfall), but the “always on” possibilities of ARM MUST be complemented with “on demand” realities of the hardware. There’s also a social component that’s just not worth getting into very deeply with all of this “always connected†nonsense, but I have found that people who use their mobile phones in an “always connected†rather than “on demand†fashion tend to be nearly incapable of functioning in simple social and conversational settings. In recent years, I’ve fired a number of clients who just couldn’t get a handle on their own maturity with respect to these issues. For most people, the positive opportunity that “always on†represents is used to merely subsidize staying incompetent rather than moving towards a place of greater competence.
that’s you, From your post your neck deep in uber-geekness and on the cutting edge of whatever it is you do.
Now take the average person who doesn’t care about 95% of your post. They want there content there , people are by nature lazy if it can be done for them vs they have to go do it themselves. There going to have it done for them. guess what they don’t want to wait either. That’s the most efficient for them.
Simply put people will take the path of least Resistance almost every time. Always on push content is path of least Resistance.
If no “always on”, then your phone won’t ring when someone calls you, or SMS you. If that’s what you want, then you don’t need a cellphone. And I do have friends who don’t use cellphones, and some are very technical otherwise.
But the commenter seems to push his own ideal into others. Incompetence has nothing to do with a need for “always on”. This is a temperament difference in being an extrovert or an introvert. You need to understand that people are not the same. Even though you are so smart in technology but you don’t seem to have a grasp of human psychology.