How Many Bytes To The Gallon? The Data Challenge of Autonomous Vehicles
A few years ago, a blind man rode in a car in Austin, Texas. Normally, such a trip would not make national headlines. However, this blind man was alone in the car. That might sound like the set up for a joke, but it’s actually a pointer toward the future: the car was one of Google’s self-driving cars. After the ride, the gentleman exited the vehicle and entered history.
That’s just one story in a never-ending stream of prognostication, commentary and “gee whiz” quips about the coming era of autonomous vehicles (AV). AVs are burning rubber in media of all kinds (you can find video of the blind man’s spin around the block here). Some of these stories address concerns about AVs and I don’t mean to discount them. Yet, AVs have an air of inevitability about them, don’t they? It’s not a question of “if” it happens, but “when.”
I am just as awed by shiny new objects as anyone. I look forward to spilling my morning coffee on the way to work without danger of rear ending the car in front of me or missing my turn. When will that be, exactly? To ponder that question, we must confront the complexity of getting these cars on the road safely and in volumes that really can change the world.
AVs are not just motorized marvels packed with sensors, radar, microprocessors, and algorithms. They are the most visible parts of a systemic change in the way the automobile industry and the technology industry work together. That change goes well beyond the chips and software in the AV itself. When they happen is dependent on just how fast and extensive that cross-industry partnering happens.
That car will be a mobile node on a network. Think of it as a really big smartphone with wheels. It will produce enormous amounts of data that will travel that network and touch numerous computing resources on its way to the cloud for ultimate processing. Any insights or adjustments to algorithms will travel back down the same pathways to the AV.
Here in Michigan, it might be easy to assume that we will see an explosion of auto manufacturers building their own networks to serve their various AVs on the road. If the industry takes that proprietary build-and-own-everything path to success, we will be waiting a very long time for AVs. To understand the reason, we need to turn to a favorite business school case study: Sony’s Betamax.
Once upon a time (the late 1970s and 80s), Sony’s Betamax felt inevitable too. It was the technically-best videotape playback machine. Sony controlled all of the Betamax pieces. If film makers wanted their content playing on the best machine, they had to pay Sony a licensing fee. If a consumer wanted the best playback machine, there was one source: Sony. Sony owned the entire value chain and designed every piece of the technology.
The forces behind the VHS standard took a more open approach. Everyone could use the standardized approach to build video cassette recorders, encode video for playback and produce tapes. An entire industry ecosystem grew around VHS that could never develop around the Betamax model because Sony owned and protected the entire solution with castle walls and moats. VHS was easier to work with, and because there was competition among VHS playback machines, the price was driven down quickly. That created volume. Eventually, all new content was offered first in VHS because it was by far the larger market. Over time, the Beta version of a popular movie grew later and later until it was dropped all together.
That was 30 years ago, but the lessons are relevant to the future of AVs. By trying to own everything, Sony ended up with nothing. An ecosystem of players from different industries contributing their talents in their own segments built a better end-to-end solution. So it will be with AVs.
The AV itself will be packed with technologies developed by other companies, but that’s just one end of the end-to-end solution. 5G wireless networks might help collect the terabytes of data each car will produce, but those towers will hand that data off to fiber optic networks. It’s easy to say that those networks are a roadway to the cloud, but it gets more complicated. When lots and lots of AVs are navigating in an unpredictable landscape, milliseconds will count. Going all the way to the cloud and back for guidance will take time. Those networks must be married to a computing fabric much closer to the cars themselves where algorithms can be deployed to manage all this autonomous traffic. That fabric then will pass appropriate data on to the cloud for further processing.
The core competencies involved in building the AV, building the 5G infrastructure, building the combined network and computing fabric, and building the cloud data centers are all quite distinct. Building any one of those pieces is a huge capital investment. Consider one company building all of them! Financing that undertaking is beyond the resources of any one company.
To deliver on the promise of the AV era, where a blind person is empowered with personal transportation, will require the best of a wide variety of talents and enterprises. A VHS-style ecosystem will be able to innovate faster than a Sony-style monolith. If we want AVs on our roadways anytime soon, we must rewind the tape to take a lesson from business history.
Recommended for you: Visteon Puts Its Trust In CenturyLink As We Drive Towards An Autonomous Future
This article was previously published on NetworkWorld on September 25, 2019.
This blog is provided for informational purposes only and may require additional research and substantiation by the end user. In addition, the information is provided “as is” without any warranty or condition of any kind, either express or implied. Use of this information is at the end user’s own risk. CenturyLink does not warrant that the information will meet the end user’s requirements or that the implementation or usage of this information will result in the desired outcome of the end user.