In our podcast today, we're talking about the IoT trends that will require colocation data centers and cloud services to play a major role. The automobile industry will be our focus and here to share her thoughts is the founder of Silicon Valley-based TechLAB Innovation Center, Sara Rauchwerger. Subscribe to Sara's next podcast in eHealth, Energy, Gaming, Retail, Smart Manufacturing, Smart City or Smart Buildings.
Ben: Welcome to the vXchnge podcast series. I'm your host, Benjamin Hunting, and today we're talking about the Internet of Things and trends relating to it that will increasingly call for colocation data centers and cloud services to play a major role. With us is Sara Rauchwerger, Founder and Managing Director of TechLAB Innovation Center. Subscribe to our blog to hear next month's industry discussion with Sara on eHealth. Thank you for being here today, Sara.
Sara: Thank you very much for inviting me.
Ben: Now, connected cars are something that's very much in the news on an almost constant basis these days, and I think what sometimes gets lost in the discussion of autonomous vehicles is the sheer amount of data that's going to need to be transmitted in order for the connected landscape to function. And not just from vehicles but also traffic lights, signals, and sensors on almost every mile of roadway. How is it going to be possible to manage all of this data?
Sara: That's a very good question. Well, amazingly living in Silicon Valley has changed so much. I had a podcast with you guys last year, and what we see at TechLAB Innovation Center which is a tech startup center, are pieces of the future. Our topic today is autonomous vehicle, which has taken center stage, even though that vision is more than 60 years old, it has been in the works for more than 20 years recently, really thanks to DARPA (defense advanced research projects agency).
But if you go back 60 years, we didn't really have the technology. Even 20 years ago we really didn't have that technology. It's become much more sophisticated, so now we can actually start to leverage on it. Over the past year, just at TechLAB, there's a company called Teradeep which is an early stage startup. They've been developing what's called a deep learning software and hardware solution, which is basically the principle of the whole autonomous vehicle.
So I've been watching how a machine learns to recognize things as silly as a cat, a dog versus a human all in real time, to then, funny enough, a year later had Detroit knock on our doors to want to buy the solution from them. You're probably starting to understand the correlation. What's happening in the industry is that more and more applications are being built for real time feed of information, collecting enormous amount of data.
To manage all this data, we've been building open platforms that help facilitate easier adoption. But the key component to make all of this work is really analytics. It's helping build what I consider the ecosystem or the peripheral product and services that will support this new industry. Take autonomous vehicles for example, where you need to have a detailed mapping which shows your vehicle roads. The vehicle has to make the right decision, traffic signal communication, sensor technology.
Much of the sensor technology is part of the IoT as you mentioned Ben. Amazingly enough, in Silicon Valley, IoT was the hot space last year. Now we're in a completely different space where we're looking at other technology and how to interconnect everything. I'll have to admit the technology is nowhere near perfect yet. The companies like Google, Tesla, and others who are following suit are taking this challenge.
So to answer your question regarding the managing of data, some of that data will be processed locally on their interconnected devices, while the larger data such as the analytics, will place new loads requirement on data center.
Ben: It's interesting that you mentioned machine learning and specifically deep learning, because on top of storing and managing the data flow, you still have to analyze all the information that's coming in. As you mentioned, that has to be done in real time. Do you see machine learning and deep learning being the most likely candidate to rise to the challenge, or do you think we're on the cusp of something else where deep learning, machine learning are the entry points into a new level of technology that will handle this information?
Sara: We're actually jumping a little too far ahead, and I will tell you why. When you hear my answer you're going to be quite surprised. The real challenge for autonomous vehicle and infrastructure that we're talking about is really urban city services and solutions that will facilitate the autonomous vehicle viability in the first place. Cities and counties are now forced to look into what we call the smart cities. We do have one city in Silicon Valley that has taken that challenge and through partnership with the technology sector, the city of Palo Alto. The city is in the heart of Silicon Valley right next to Stanford and has committed to taking this initiative.
And the first thing they've done is just as a test, they converted all of their analog traffic signals to Internet protocol sensors, which provide more intelligence to help control the network. They then strategically, for example, place sensors to collect data on the road. And what they're doing is they're trying to solve real time city problems, which helps drivers, and in the future the autonomous vehicle find, for example, parking spaces. What is happening is that the sensors are sending data to a cloud resource to inform the car of the nearest parking space.
Electric vehicle ownership in Silicon Valley is growing very, very fast, but the problem is charging stations. They're becoming an important element of the infrastructure. How do you deal with that? There's a force pushed by cities, counties, and the state to mandate policies for charging stations. One company that's well-known is PlugShare. It's one of the dominant players. They have an app, and it allows the drivers to find the nearest charging station. And from my own personal experience, because I have an electric car, they're always full now.
So the end result of this challenge, it forces the city to evaluate and plan future urban development, accounting for the technology change. So that's what we're dealing with. The data center in the infrastructure, but it's the rest that needs to follow suit.
Ben: And it's interesting too, the set of the ecosystem, to me that talks to the data transmission infrastructure, because it's good to know the cloud can handle this type of analysis and storage of the information that's coming in. There's only so much bandwidth available for transmitting information, and we now have hundreds, if not thousands, if not millions of small sensors and devices that are using the same data transmission infrastructure to feed the cloud for all these autonomous vehicles. And that would seem to dictate that question of scaling is really going to become very important.
So do you think that we're on a timeline where we're going to have to dramatically expand bandwidth availability before we can fully transition to an autonomous vehicle infrastructure?
Sara: To answer your question, the current data transmission infrastructure is actually capable of scaling by leaps and bounds. Our existing infrastructure is not 100% yet impacted. However, as more and more complex software is entering the market, solutions will have to be developed. Today we're actually using mobile phones to drive data center evolution.
But with everything that's evolving with autonomous vehicle, it's a huge ecosystem and all the key players are playing in it to build that infrastructure, to be able to make this work. So using mobile phone, sure, it doesn't impact the system, but if we're going to develop that whole ecosystem, we're going to need to have something out there that's a little bit more powerful, and that's where the data center evolution will have to change as well.
So as our market is driven by higher demand of these products and services because they're already penetrating the market, what will happen is the performance will be reduced. So we need to increase the performance. At the moment, technology is not an obstacle at all. If you can dream of a process, you can implement, just like the success and failure of startups. This is what I see in our center. The core of the autonomous vehicle is actually human behavior. It's how we use the data and analytics to manage and change the human behavior.
If you look at global cloud computing marketplace, it is growing so quickly. We're estimating, if you look at market opportunities, not we, but if you look at what's being estimated right now, $80 billion by the end of just this year, that's a huge market, which tells us that computer intelligence is getting better and better every day. There are more products and services that are getting pushed very, very quickly to allow us to expand into many industries including medicine and science, every home of human endeavor. Basically we're revolutionizing the human machine interaction. So it's not only the autonomous vehicle and its peripheral ecosystem, but it's leveraging on the infrastructure.
Based on what I see and hear, there will be a gradual growth in the sector over the next few years. And if you listen to what Google is telling us, because they're pushing very hard. If you listen to Tesla, although we did have one accident, but we're going to have many more because we're learning and we're going to have to learn until we get it perfect. They want to bring this out to the market in the next two or three years. You can just imagine what's going happen on the data center side.
Ben: Sara, this has been a fascinating conversation. Thank you very much for taking the time to speak with us today.
Sara: Thank you as well.
Subscribe to our blog to hear next month's industry discussion with Sara on eHealth.