It looks like Google is now planning on adding a fully-fledged satellite network to its Project Loon balloons and Titan Aerospace solar-powered UAVs. According to reports, Google is on track to invest $1 billion in 180 low earth orbit (LEO) satellites and infrastructure.
The project is apparently being run by Greg Wyler, former founder of O3b Networks (The Other 3 Billion) in which Google is a key investor. O3b has run tests with 1,500lb satellites but inside sources say that Google will be using smaller satellites around 250lb in weight. O3b currently has 4 satellites in orbit, with another 4 due in July and a further 4 later in the year.
The number of satellites may later double to 360, as Google is also a key investor in Jersey-based business L5, which trades as WorldVu Satellites. It has secured spectrum rights to build a Ku-band (12-18GHz) constellation of 360 satellites, operating between 800 and 950km from the Earth, beginning deployment in 2019. The spectrum was bought from the now defunct SkyBridge Wireless, which had won the spectrum in the late 1990s but had failed to launch its project due to the high initial capital requirement.
A LEO satellite differs from a geostationary satellite as it does not stay in a fixed location above the earth. They orbit at an altitude of between 160km (99 miles) and 2,000 km (1,200 miles), encompassing an orbit of between 88 and 127 minutes respectively. A LEO constellation requires dozens or hundreds of units to provide global coverage. Higher orbits require fewer satellites but the tradeoff comes with bandwidth and latency compromises.
The LEO satellites will provide better bandwidth and latency than the geostationary satellites. The LEO deployment is very similar to the system used by Iridium Satellite, which provides satellite voice and data services worldwide. The Iridium project was initially pushed by Motorola, but the project spiraled out of financial stability and declared bankruptcy in 1999 after just nine months of service ‘ where private investor backed Iridium Satellite bought the assets (which cost upwards of $6 billion) for $25 million.
Google might spend up to $3 billion if the reports are correct. It can’t expect a return on its investment any time soon, but that is not Google’s aim. If we take its ‘Don’t be Evil’ motto at face value, this is an altruistic endeavor by a multi-billion dollar company with cash to spend. But in reality, this is a valuable enterprise to secure new and future customers of Google’s web services in the near future.
Unlike similar LEO satellite projects, Google is not looking to profit directly from the satellite connectivity, at least not if it has any sense. Rather, it is undoubtedly looking at the 3 billion unconnected people around the planet and seeing them as potential internet users with an enormous number of potential page impressions and ad revenue that they could generate. After all, with a nice bit of marketing on the ground, Google can ensure that it is the first website that the new users will visit ‘ what better introduction to the Internet than the Google homepage and a search for local news and weather reports?
So ideally, the Google model would use the Titan drones for their intended disaster-relief connectivity, the Loon units to provide a drifting but ubiquitous global coverage, all with a redundancy package of LEO satellites for users on the ground to fall back on. While this is wonderfully optimistic blue sky thinking, we feel it is proper to be somewhat more pessimistic than other outlets. For us, the economic arguments for this don’t make sense as a comprehensive global deployment.
The argument for connecting the next 3 billion internet users makes perfect sense in the short term. With little effort on their part and the support of local governments, it’s true that Google could roll out its Loon connectivity to all of Africa in a matter of months. However, the rosy view of this feat of connectivity always fails to consider the long term usage of these newly connected people, who very soon will want to start watching video on their newly connected devices (themselves an issue, as users need compatible ground receivers and computers/mobile devices). Satellite connectivity can’t handle this, no matter how much money is thrown at it. There’s no way to backhaul an entire continent’s video usage over satellite. This is where wireline and broadcasting infrastructure comes in ‘ but only after Google proves to the wireline investors that there is money in the new markets.
Google would then only have a few years in which it will be the only ISP in these new markets before the cable and telecom engineers move in. This is certainly long enough to win good-faith and influence consumer behavior in ways that can’t be bought, but Google can only connect these mass markets in the short term via satellite and balloon before the cable players move in. Google itself could lay the required physical infrastructure, but if that is the long term plan then is it really necessary to go the satellite route first?
The satellite connectivity would of course be massively beneficial in the interim period while the wired backhaul network is constructed, which will require lengthy projects. But once the wires arrive, the balloons and satellites will pale in comparison to the latency and bandwidth of traditional wired network architecture.
Google’s connectivity projects make the most sense in areas where it is nowhere near cost effective to install wired networks. Islands, rainforests, mountainous regions and populations spread out across vast expanses of land, these will all benefit from the wealth of information that a basic internet connection can provide them with, as data cables have not been built to bring these remote users online.