Perhaps you have been following the huge legal battle between Aereo's streaming TV service and several major over-the-air (OTA) broadcasters, but Aereo's other problem has received little attention. This nonlegal problem is what plagues so many projects: power consumption.
If you are not familiar with the Aereo approach, it grabs OTA TV signals, converts them to streaming Internet format, and sends them wirelessly to your PC or other device via the web, antenna, and DVR (shown below). I was under the misconception that Aereo offered a PC-like dongle or adapter at the user end that received and tuned OTA signals, converted them to a digital stream, and output this via a USB interface, but that's not how it works.
I don't see the virtue of the Aereo approach -- it seems like a complicated solution to a nonproblem, and it consumes significant wireless data capacity -- but I'll let users and the market decide.
The Aereo system captures OTA broadcast signals and converts them to streaming video for web users with any device.
The Wall Street Journal reports (subscription required) that each Aereo subscriber requires a transmitter and a small antenna setup at Aereo's area-server facility (shown below). This a lot of upfront hardware and fixed costs, and each user setup consumes five to six watts. Do the math, and it's pretty clear that, if this company succeeds in getting tens of thousands of subscribers, that's some serious power it needs to provide and fund.
A single Aereo streaming antenna (top) and a farm of these antennas (bottom) give some indication of the hardware complexity
of this streaming TV service.
The article points out that cable set-top boxes also consume power -- typically 15-25 watts per box. The difference is that the cable subscriber, rather than Aereo, pays for that power, and the inevitable dissipation is localized, rather than centralized (even if the power generation comes from elsewhere, via a wall plug). As Internet providers know, when you start concentrating megawatts of power use and dissipation in a central facility, you accrue big advantages but bigger headaches.
It's one thing for engineers to have a power budget (which all projects have), where you are restricted to using a certain number of watts (or mW) to maximize battery life and/or minimize thermal dissipation. But in those cases, you have a fixed maximum. It's a very different power challenge when the amount of power you have to supply and dissipate goes up linearly with the number of customers you are trying to serve.
It's easy to think of the Internet as a convenient wired or wireless port to which you simply connect, but there's a huge amount of physical infrastructure behind that spigot: server farms, routers and switches, and intra- and inter-facility cabling (some copper, some wireless, and now mostly fiber). All this hardware consumes power that needs to be supplied and dissipated.
For an excellent, eye-opening perspective on Internet physical facilities and their requirements, check out Andrew Blum's recent book Tubes: A Journey to the Center of the Internet. It's a well-written, informative, insightful, respectful, and even admiring reveal of what engineers and others have accomplished to make the Internet ubiquitous and just there for so many of us.
The Aereo approach -- like any system architecture in which hardware and power costs grow linearly with the number of users -- makes me admire the virtues of the simpler broadcast model, despite its limitations. For basic broadcast, a single transmitter can serve a limitless number of users within a given range, and there's no need for additional hardware and power at the source. Perhaps power limits, rather than legal strictures, will be the real constraint on services such as Aereo.
What do you think of Aereo's approach from a technical, nonlegal perspective? Have you ever had to deal with system-level power issues that scaled up and got worse as you succeeded and got more users or customers?