The following are the outputs of the real-time captioning taken during the Twelfth Annual Meeting of the Internet Governance Forum (IGF) in Geneva, Switzerland, from 17 to 21 December 2017. Although it is largely accurate, in some cases it may be incomplete or inaccurate due to inaudible passages or transcription errors. It is posted as an aid to understanding the proceedings at the event, but should not be treated as an authoritative record.
***
>> XIIAOFENG TAO: ...science for technology. I share this workshop Internet of things for smart city: Green and sustainability, WS42 for short. You may agree with me that smart city is the future of organization and its foundation is information technology especially in indicating IoT big data cloud computing and so on. This is what IoT is for smart city and how you can improve life quality, foster (?) and the need of sustainability. In this workshop we have divided six speakers to give talks on various topics ranging from smart safety framework, (?), security and policies. I hope you will find them affirmative and enjoyable. And I want ‑‑ let's ‑‑ I want to make brief presentation first.
You can see today's Internet is recognized for three leaders, the first openness, the second one (?) two everyone can access Internet and it's in equal spaces. The second one is collaboration. The collaboration using the Internet has many advantages such as (?), low cost, sometimes almost free and green. Sharing economy that we already have is essentially a collaborative economy where people collaborate. It is expected that as such a person pays collaboration, become collaboration of Internet of things. By adoption of Internet technology. Traditional industry production we are combined to information technology by using, like I said, IoT, big data, AI, Cloud services and so on. Please, next page. Here I give you some examples. Traditional distribution systems, you create a large number of persons to sort and distribute the goods from source ‑‑ from the source to consumers. You can see the left figure, however, today distribution systems start to use IoT technology to facilitate and monitor the movement of goods. You can see the figure. Some IoT‑based distribution systems conserve more than 18,000 items per hour, three times than the person‑‑based system with our around the clockworking style.
You can see at the right figure some Chinese company has also studied trial test for new (?) with trunk driver. This wide trunk can carry five packages at once and 20 kilometers, maybe we are about to see them soon. Last page, please.
Here also, a second example, you can see, consider two solution to meter reading. The left one is the traditional meter reading solution. An employee can read about 18 to 100 gas meters a day in a typical Chinese city. Hundreds of millions of potential houses across the country (?) and burgeoning costs. However, what about electricity meter and water meter? The right one is smart way use IoT technology. IoT enable stable real time data collection from more than one utility meters like gas, electricity, water, and more, we can analyze this meter data to implement ‑‑ implement best management of the city.
The new solution, you can see it's the right side, can be greener by using IoT technology. Here is an example of an IoT‑based gas reader. This is what I use in Shanghai gas group. It is shown that this gas reader equipment is about 100 times energy efficiency and that is GSM‑based reader. The GSM‑based is second‑generation mobile communication systems and this one use IoT technology. This belongs to 4G. Please, next slide.
Here is an example for transportation in a Chinese city. An operator is IoT‑based, camera bay stations, and the government can cooperate to add a new dimension to existing intelligent transportation systems has a great potential to improve the road safety and decrease traffic congestion. Here is the ITS ‑‑ here ‑‑ it is the ITS, the commander center in this Chinese city. For example, traffic ticket and so on are projected on a single big screen at the command center in a real‑time manner. The red hand, finger is a web screen, for safety traffic information like when and where (?) high traffic congestion citizens can look at this Web site anytime?
A. NIR. Next slide, please.
So about three examples together with other IoT applications will lead to a much bigger concept, smart city. Forest research in the year (?) smart city will use intelligent (?) infrastructure components such as transportation, hair care ‑‑ healthcare, education, emergency response, and utilities. IoT and smart safety is expected to hugely improve life quality. Next.
Developing smart cities has so many advantages as imagined in the previous slides. It becomes smarter and smarter. China is a big player in this area. Nearly 500 cities are beginning their smart city plan. Next, please.
So just (?) which can be viewed from the following aspect. Security and privacy, as it's very important. When everything is collected, anyone anyone in the world can collect you how the IoT companies to protect our personal information. This is very important. Second, smart city on the policy, what is the rule of the host maker or government or the country and the consumers. And, for example, (?) also very important. The first one maybe home video where the video will be affected by those new technologies. So this was very important, can we reserve our traditional old customers. Okay, I'll just give a brief, quick presentation. So second I want to invite Ms. Galpaya.
>> HELANI GALPAYA: Yes.
>> XIIAOFENG TAO: Galpaya, Ms. Galpaya. Ms. Galpaya's talk is about mobile network and (?) data. Ms. Galpaya is the CEO of ‑‑ I'm sorry, what is the name?
>> HELANI GALPAYA: LIRNENasia (phonetic).
>> XIIAOFENG TAO: Lirneasia. Please.
>> MS. KAPAYA: I run a think tank called LIRNEasia, we work in Asia. I'm going to take a very different view to this in the sense everyone in the countries we work in, you know, in India, Sri Lanka, IoT's use, everyone is talking about smart cities, yet there's no serious studies we can study. Yes, we can study in Korea, nothing in south Asia in in any meaningful way, we're in the early stages. However, it doesn't mean our cities don't have these problems and we use other types of senses which are not what you fix on buses and not what you fix on weather stations. Next slide, please.
Next. So these are the countries that we're already working. Next. Right. So for us what's interesting is that this huge rate of urbanization in Asia overall, certainly in the poorer countries, and next ‑‑ yes, sorry, previous. And our question is how do we make these cities more liveable. We are behest with traffic jams. We are behest with Dengue and malaria which are spreading seasonable, and are endemic in many of these countries and many other things and how do we use data to answer these questions? Next slide, next. So there's a whole lot of data that you can ‑‑ I'm sorry, yeah. There's a whole lot of data that you can use, and obviously this ‑‑ you know, the administrative data that you can get from government and the commercial transactional data including bank records, all of that, online activity on social media sensors and tracking devices right which is what you were mostly talking about in terms of IoT, the last one. Next slide, please.
However, in the absence of sensors and tracking devices humans are the best tracking device, walking around in cities and rural areas. Mobile subscriptions are very high, even if people are not using the Internet, they are using a digital signal in the form of a call detail record which isn't relocation registry and that trace is left in the telecom operator's network so thanks to high mobile penetration we actually have a digital map of something that's going on in these cities. Next.
So we actually use the call detail records in together with some population data, we use historically ‑‑ historical data (?) in order to protect people's privacy and from multiple mobile operators in the countries we work in and I will present an example in Sri Lanka. Next.
We can skip this. Right. Next slide, please.
So how do we actually use it to solve city problems? One of the problems is because cities are so fast changing and sensors are down once every ten years and if you're lucky we get all our data out another three years later, there's no really real‑time estimate of where people live and what population densities are. These are incredibly important to identify where diseases might spread, for example, also density really matters, and on your left‑hand side is the population density based on the government survey, census that was done last in 2012. In the middle is the population density data for the same sort of geographical breakdowns based on our data, which is much more real time, and as you can see, the mapping is very accurate, and on the right‑hand side is even more granular mapping which you can do especially in high density areas where there's more than one cell tower. So you can actually get quite granular data, administrative data, through the sensors ‑‑ census and surveys do not capture to identify where real density is at any given point in time. And this is important for cities. Next slide.
You can also identify where people are at any given time of day, which is kind of like before, the density, but more importantly, where people come from. So this is a zoom‑in on the city of Columbo which is on the southwest course of the commercial capital of Sri Lanka and as you can see, depending on the weekday or weekend and the different time stamps you can see the population hubs. The red areas are where the people are coming into, so you can see there's an influx of people into Columbo and the more blue it is it's where people have come from, right? So you can see at this level of resolution where people have come into the city and then we can zoom down and see more specifically from where they have actually come granular areas. This is important in transport planning, so we use a more granular level of this to work with the Columbo mega palace plan which is one of the biggest projects in Sri Lanka to help plan to put the big multimodal transport hubs where people should park and get on trains, for example, what is the congestion and so on at any given time and therefore more advanced traffic routings can be used. Next slide, please.
So just to give you why this is a good idea, I mean, it's not like you can't get origination destination data for people by doing surveys and we do that, we stop people on the road, the government, in heavy traffic and ask them to do surveys or they come home. You can't be doing these surveys in traffic so then you go to people's house. These are beset with memory problems with average trips taken and so on. So on the left‑hand side is our data showing where people ‑‑ the population, sort of where people live and come from and on the right‑hand side is the transport survey data Travis funded by the Japanese government. You have a pretty good mix. And, in fact, you can identify population or density hubs that you didn't actually get in the administrative data through the actual call detail record base analysis that you see on the left‑hand side. Let's go next. I'll skip this is where people travel and come from in the interest of time. Next, one before, please.
This is to understand the impact of transport. So we built ‑‑ one of our ‑‑ our second highway ever was built Columbo connecting the airport into the city. What happens when you build a new highway, do people travel more, do they travel further, what happens to mobility, are people better off or worse off? These are questions we can answer and incredibly important in deciding whether you should do more infrastructure investment. Next slide.
So, for example, the E3 is the expressway that I talked about and we can see that in the expressway the traffic increased and much more on other routes so you get a sense of how much the expressway was used compared to other alternative routes and then you can do local level surveys to understand what actually that (?). Next slide.
We can also go further. We can understand the level of localized travel, incredibly important in planning local travel which includes not just buses but our tuk‑tuk and auto stations, that take people into the big bus stations and other areas, so this is not just to understand one city to another city, this is at a more local level how do people move around and to identify sort of transport ‑‑ people's transport networks, where do most people go, these are the routes that should be optimized. Next slide. Understanding traffic conditions, this is, of course, new in some countries, very sort of well done in many developing countries. So we can't do this with the mobile network big data that I talked about we use CCTV footage and recognition to look at traffic conditions. Next slide.
In a completely different example we look at diseases. Malaria has been around for a while but that's less of a problem, the new problem, for example, in Sri Lanka is Dengue, again, a mosquito, a vector‑borne disease, where one of the biggest determiners of how it spreads is where people move, particularly infected people move from one place to another, therefore it spreads when a mosquito bites them in that place so people movement is incredibly important. On the right‑hand side is what actually happened for a given period of time, the actual incidence of cases in Sri Lanka for this period of time on the study.
On the left‑hand side is our predictive model using many of the mobile call detail records and some other rainfall data vegetation and a whole set of population data that we did and the model as you can see is actually very accurate about the likelihood of spreading, in fact, our (?) we possibly overpredict in some areas which is darker color than the actual incidence but we think that's a good problem that you have you because at least you've identified where it's likely to spread. At least we're talking thousands of people dying so this actually matters in our country. Next slide.
Establishing data‑sharing agreements is incredibly hard, especially because we are an independent research think tank, establishing multiple operators, getting them to agree to share data is very, very hard because they are worried about commercial interest about where their base stations are would be deep to other operators and so on. These are just examples I showed you was based on data which took us 24 months to negotiate, an incredible transaction ‑‑ at incredible transaction costs. Next slide, please.
However, getting government data is no easier and you think this would be, and governments put huge restrictions on who can access data and what conditions need to be met. For example, just to get granular census data, not to identify any households at a much higher level, you need to write research proposals, they'll give you 20% of the data, then you prove the concept, then you get 80% of the data. However, this kind of big data analysis cannot be done with 20% of the data they're missing the whole point. N equals all is the whole basis of running algorithms to identify patterns, and they don't know how to meaningfully sample and give us even a good sample of data so this is a real problem. Next.
So we really need public‑private data partnerships. If this kind of very basic IoT of humans is work. And certainly the other stuff. Because the government has no capacity, they cannot do some of these things. They need to bring in the experts. And we don't have experts ‑‑ we don't have access to sometimes even the experts, we don't have the capacity either sometimes. We need to reach out to universities. We need to open up the data through APIs so data that doesn't violate people's privacy can be shared. And there is a lot. The debate about data‑sharing is at the moment I think clouded by privacy which is an incredibly important component but there is so much data that can be shared without having to worry about privacy. So we need to separate out these two tracks, have meaningful conversations about what can be shared and how versus what should be extremely highly protected and even maybe kept localized and within the country. Thank you.
>> XIAOFENG TAO: Thank you. Ms. Galpaya is it here, so one question to her.
>> AUDIENCE: Hi. I'm and environmentalist so it's more on the environmental question. I was wondering if the government would be more likely to give you the material, the raw material, if you framed the request as an environmental request, so reducing pollution by knowing the data and sort of the government having a broader transportation strategy, having a more focused strategy because they're meeting the needs of the people.
>> MS. KAPAYA: In this instance, no, in fact, framing it in terms of let's help you figure out the transport problem was much more useful because coinciding with what we're doing is the World Bank came up with this Colombo mega (?) plan, huge grant and therefore, they actually needed to understand where to put the roads and whereby to put infrastructure and nobody knew this and therefore this is one of the reasons they let us in through the door so that policy window might be different in your context, maybe the debate is, okay, listen, this is about the environmental and the government cares. In our case it actually was transport, luckily.
>> AUDIENCE: Just to make sure, the question ‑‑ just short question. You are doing great efforts to digitalize the smart cities everywhere in your region. But how to do it with ensuring privacy and human rights is the issues of citizens who are ‑‑ I know that is, but the system works automatically. Could some cities, for example, be against the data to be collected, the data to be analyzed, for example, the geography opposition or something else? Thank you.
>> Ms. Galpaya: So in this case, for example, we don't, actually get anything from the telecom operators anything about where they work or live or their names or ID cards or anything like that, right? Of course there are ways algorithms can determine where people live with great certainty, at least down to like a cell because these are the base station records that we use. If we try we can. We've decided to get it historical basis so it's not real time. Before they flush it out of their system we get a trillion records and then we go again and get a trillion records. This is not something that actually has been solved because even the data that we have it is possible at some point in the future in combination with other digital traces like credit card information, you know, may be solvable although all of these data traces are nonexistent with 4% penetration, right? So, as I said before, we need to understand which data is sensitive and which is not. However, it's not just about privacy, it is also about inclusion. So if, for example, you look at the Boston city deployed a street bump app which is on the smartphone, the meter identifies the bump on the road a pothole if you're driving around, this is a great app to use you see the pothole you send a crew and fix it this was a great thing. Then the privacy side came over, this is hugely problematic you're identifying where people are and et cetera.
Then third the other side of the lobby came up and said, you know what, rich people who live in back bay have smartphones, poor people who live on the south end don't have smartphones so suddenly you're going to identify potholes in the rich areas and identify those potholes fix those potholes and you're not going to identify the potholes in the poor areas because they don't have smartphones. We need to go to the next step and say yes, I understand there is a representation problem because smartphones are with a particular type of people and they need to know that. Cities need to update resources so I don't see anything wrong with a city saying I will use that with the people that have smartphones but I will use my old real‑life crew which actually looked to find potholes I will deploy them into the poor areas so I know I'm not going to get a signal there because you updated your resources because you only have one engineering van to go all over the city, now you only have to go to a small part of the city. It's a lot more than saying privacy or not, inclusive or not. Okay? I'm sorry, I have to go.
[APPLAUSE]
>> XIAOFENG TAO: So I'd like to invite Mr. Aihua Wang, Mr. Wang is the president of Australia computer society and he's talking about experience of developing smart city in Australia. Mr. Wang, please.
>> AIHUA WANG: Thank you, Mr. Tao, thank you for the Chinese society for inviting me to speak today. My topic is about IoT and smart cities in Australia and particularly I'd like to take a different focus because she's going to talk about, smart security and spectrum of subjects about IoT and smart cities here on today so I'm going to focus more on the challenges particularly who owns the data from IoT. So my background is I'm a lawyer. I specialize in privacy, intelligent, intellectual property law, I'm a technologist, I'm the president of the society I'm on the board of the international federation for information processing, IFIP, that was created by UNESCO years ago and based in Austria. I physically live in Sydney, Australia but we're part of the international society of computer societies around the world. So professor Tao when he mentioned about what IoT is and where it's going on I'm particularly focused on Australia perspective, Australia being a developed country, as you can see from that slide, the penetration of IoT devices and mobility devices are pretty high. One of the statistics from a recent government report indicate that over 90% of Australians will be online by 2017 which is this year, really. And the average Australians own 24 devices already which are connected online. So these are some of the statistics. So we are heavily utilize devices in the Australian landscape. Also the next slide talks about McKenzie report looking at the benefit of IoT in the Australian context and its contribution to the economic well‑being of Australians.
Last year the Australian government announced the Australian smart city initiative to make Australian cities and region areas more liveable and using a more recent technology and smarter technologies in IoT. But as everybody knows, there are no real smart cities in the world. Yes, we have many projects to do about smart cities and IoT but those are coming up but they're separate, discrete projects, there are no really functional smart cities that I can see today but that will be changing very shortly.
So in terms of the Australian landscape on smart cities initiatives you look at data for transport in Sydney and many cities around the Australia so you can get real time information about buses and trains and getting from A to B or even tracking the buses and trains and the time they arrive at your destination. We're also looking at street lighting, meters, trails of autonomous cars, and many, many of those are smart city projects.
Australia in a recent ranking from the company easy park just last month rang Melbourne number 10 and Sydney number 12 in smart city assessment. Rated Melbourne and city very high in terms of 4G penetration, citizen participation in both Melbourne and Sydney's pretty high at 9.82% out of ten, 9.55 out of ten. So in terms of also smart penetration, 9.3 out of ten. So Australia fare pretty well in terms of using mobility devices and IoT but at the present time pretty poor in terms of clean energy and environmental protection.
So last year the Australian government has launched an initiative to fund IoT and smart city projects around Australia by investing in 50 million of that 28 million so we've been allocated to a number of cities and projects around Australia including in the city of Darwin, 5 million to make this Darwin city smart with smart lighting, parking, wireless and CCTV cameras. In addition to that, 5 million's been allocated to a city north of Sydney, Newcastle in making the city smart and making it like a living digital map to experiment on the smart city concept in the many areas that we've already canvassed so I don't want to go through that detail but you can see from the slides those normal projects of IoT smart cities, those are now starting to roll out in cities like Newcastle in Australia. Of particular interest to this topic because we're talking about green and sustainability living, recently just a couple weeks ago nearly a million dollars was granted to a regional park in Perth in terms of sustainability and green ecotourism so utilizing IoT sensors, satellites, and drums to better manage the park diversity, wetlands, in real time so they can be used for the management and the running of the park which is very close to the city of Perth so this is just another example of what's happening to the smart projects in Australia.
Asia‑Pacific from a recent ranking from Vodafone, barometer indicated the region is one of highest adopters of IoT. As you look at the slide 36% of Asia‑Pacific companies have adopted IoT compared to 27% in America and 26% in Europe. In terms of companies also committed to collaboration in IoT Asia‑Pacific at 92% compared to 72 and 77 in America and Europe.
So what are the challenges of IoT and smart devices, as you know, there are many. One of the participants asked about privacy today. Unfortunately I'm not talking about privacy. I have a whole presentation for an hour just on privacy alone. So the six minutes that I have would not allow me to talk much about privacy but to mention that obviously that's an area of interest to a lot of people.
So what are some of the city challenges? I think one of the primary ones about the debate of interest between the public and the private interests, if you go in to build a new smart city or adapt an old city to make it smart, who's going to decide on what, when, and why some of these projects get to go about? Obviously cities have evolved in some instances over centuries so there are laws and governance that have evolved in a piecemeal fashion over decades and hundreds of years. So we have to look at that and how do we change those structures and regulatory system to enable smart cities which require very different thinking. And the other issues also as some of the speakers mentioned today, some of the social economy issues and who are going to finance some of these projects in the smart cities. From a strength perspective the Australian government has allocated 50 million to help cities around Australia start on those projects.
And obviously I think one of the speakers is going to talk about critical infrastructure so when you have IoT devices connecting to our critical structure of energy, water, electricity, transport, even hospitals, you can manage the reach of cybersecurity hacking, reliability of broadband or Internet to have them all connected because if something should happen to your telecom infrastructure how would those devices still function and how would the city function? So particularly in this presentation I'll talk more about the ownership of data. As you know, I was in a couple sessions yesterday where they talked about data, who owns data being some of the important issues coming up because IoT is all that data. We're collecting lots and lots of data. But who actually owned the data collected by those IoT devices? We don't really have a framework at the moment to deal with those questions, but those questions are coming up, and the debate is already here. Yesterday I was on a talk on the World Trade Organization about localization of data and regional (?) of datas across national boundaries. Ultimately the question is who gets to control them, who gets to use them, and who actually gets to own them.
So data, from my perspective, is like you imagine in the days of the Spice War when the Europeans came to the East to look for spices: Peppers, cloves, nutmeg and how high they were, in terms of spices they were like the currency of the past, now data is the currency of the future. So how do we ‑‑ the burning question is how do we balance the needs of the public the private and the private citizens in the quest for the economy and the use of the data.
So the interests are extreme in some cases, it's been debated widely in this forum and (?) and other forms around the world.
So currently most of the laws around the different regions of this planet there are generally a reluctance to accept the ownership of data or even can see an interest in data. So obviously with this debate that's going on we need to create new mechanism to facilitate the management and exchange of data among a complex web of stakeholders. As IoT's my devices I'm going to create more and more of those. So we have seen that with platform users like Uber, Facebook, Airbnb, and Amazon they all revolve around data at one point in time. One of the issues with data is like droplets of water, how do you control droplets of water? Once it's out there it merges into seas and water, how do you get it back even if you wanted to, how do you manage the flow of data like water in the river system and into the sea. So this is some of the big question and the big challenges that we've got to work on towards the future.
When people talks about owning data they're talking about the traditional concept of the right to possess, but as data is, they say can be used many times over is not ‑‑ it can be utilized over and over again, do you actually need to possess data in the concept because you can actually make multiple copies. Who gets the right to control the use of the data? Who gets to use it? And who remains to control it for the future?
So these are some of the big questions that we're all struggling with, even now with the IoT and data emerging so there are competing interests in those perception, a value of data, ultimately the legal framework may have to consider attributing some value or rights of control to data because those things are currently high on the agenda.
In the Australian context, from the IoT and devices, security devices concept about security, I know that (?) is going to speak on that, I have one slide on that concept. The Australian government is looking at giving IoT devices rating cyber kangaroo star rating similar on the energy rating on appliances so when people buy those devices they know what the ratings are for those particular IoT devices. So our advisory committee will be presenting their recommendation on this very concept towards the end of this year. So this is one of the developments that we currently have so on that note I don't want to consume all the speakers' slots so thank you, I'd be happy to answer any questions during the panel.
[APPLAUSE]
>> XIAOFENG TAO: (?) as you said, and I also think smart city be a major driver of the economic benefit of a lot of things. Of course there's some challenges for example, of privacy (?) balance between performance and reliability and a balance between private, that sector and the public sector. And next I want to invite Mr. Nu, Mr. Nu from academy of communication technology. Talks about the development of smart safety.
>> LINCOLN: Good morning, my name is Lincoln and I'm assistant researcher from the CCIT and I'm very honored to be here.
>> UNIDENTIFIED SPEAKER: (?)
>> LINCOLN: I'm going to introduce the development of China smart city. In professor Tao's PowerPoint you had Chinese smart city. I hope I can show you something different. Next slide, please.
First of all, I'll give you an example of China's smart city I'll show you general practices and finally I'll talk about the future. So next slide.
Before updating the most recent statistics I would like to talk about two things, the first one, next slide please is the broadband. I believe that broadband is a key foundation of city development and smart city. So in 2013 the state of China elevated broadband to national strategy, named broadband China and it aims to promote broadband penetration all over China and now it's around 70% and also have a goal by the end of 2020, we expect that 20 megabytes per second and 4mpps for Europe and this goal is almost achieved next slide, please. There I'd like to talk about the policy. With colleague turbo, that means positive policies do generate great power for smart city development and China's government has been playing an important role in the smart city construction. So to promote the smart city program the central governments formed across ministry working group which cochaired by the national development and the reform commission and the cyberspace administration of China and you can see the participation from other 24 ministries. And in April 2016 the working group clarified its missions, core values and operating principles during the first meeting and here is the policies like the guidelines for Internet (?), guidance for big data, planning for promoting e‑governance. Next slide, please.
Finally here are the statistics. The total number of China's smart city has reached 386 now, including all provincial and subprovincial cities, 70% for prefectural cities and 32% of county‑level cities. So next slide. Because of time limits I'll just pick several exemplary practice. Next slide, please.
The first one is Internet plus government service. As we know that China has a lot of population and people have to queue up for the government service, and as you ‑‑ usually one issue needs approvals from different agencies and it's really time‑consuming, so, for example, in Guandong province by the end of 2015, 50 Chinese agencies have been involved in Guandong e‑governance system and more than 1,600 online services are available covering nearly all of the government services. Next slide, please.
So the other one is for the medical service, we take the (?) city as an example, as a city of Guangxi province. So we need the municipal government to be government‑led and market‑oriented operation mode and established telemedicine services and a collaboration platform with enterprises which court an inboard Cloud hospital and these are working with (.) And next slide, please.
We also viewed some platform for migration management, these are real time. We can see the real time population flow and so where are they going, where are they from and what do they do and what should we do to improve service for the people.
Next slide, please. Smart plus it's cool, bus arriving time, totally under control. We also upgrade these platforms through the bus companies, they could know the bus location, condition, driving time and flow of passengers. We can use these data for some analysis like how to optimize the bus routes, how to provide better service and how to increase profit and even how to improve city planning in a bus station construction. Thanks. So we talk a lot about a lot of achievements by the government but we cannot ignore the power of the Internet companies and actually in China is ‑‑ the big Internet companies do contribute a lot to smart city development. So, for example, (?) provides an Internet related to various government cities and Ali Baba provides we call it three entrances plus one platform (?) city service platform. And for the transport service, we have the DD platform, DD is the Chinese Uber and DD provides service in 400 large and medium‑sized cities and 1 million taxi drivers and 110 million passengers have registered at DD platform and 3 million of them are active per day. So the average amount of order per days 5 million and one's picked at 50 million. Next slide, please.
So now we move to the future. So for the future we have three important tasks. The first one we take (?) as the example as we know the GDP ranks low amount of provinces of China. But fortunately more than ten cities in (?) provide ‑‑ in (?) province have been involved in (?) program which grants (?) a leading role among Chinese and these data driven smart cities are supported by Cloud service we call it seven‑plus end Cloud service include government surveys industry, tourism and so on. We also have the M application platforms. That means they put one intelligent applications are illegal and or platform the (?) Cloud system for data application. So this. So we will do the same for other underdeveloped regions just like we have done for (?). So next slide, please. So the second is to further diversity to ‑‑ diversity of development path. It seems like to educate different people according to their natural ability. So at present we have three development type. The first one is the comprehensive development motto, the like Beijing and Shanghai, these cities are large‑scale and with a strong economy and with high city developments index. And for these cities smart city is a prime development strategy and implementation is easy to achieve because of good foundation and condition. And the second model we call it the industry's pooling motto like the second tier cities and some coastal cities like (?) which have great potential to achieve comprehensive development and it's important for this city to develop registries with local advantages and to prepare the development of key areas and bring more favorable resources from outside. So the third one is the ‑‑ sorry. The third one is the fullout model. It's like Google and (?) there's more and medium‑sized, it's important for them to improve infrastructure construction and application development and enhance the foundation for future. Next slide, please.
So the third is to ‑‑ is to promote cooperation of government agencies and the private sectors and during the 13‑five year that means the 2016 to 2020 the estimated investment scale for smart city construction exceeds a this will thresholds is around 1.64 to 6‑point ‑‑ 6.70, 5 trillion, and we can see the huge gap between the investment input and the expected goal so some business models are largely depending on government investment and management and we have launched and query and based on the results 32% cities suggest the fund‑raising from the market and 40% of provinces suggested to promote the public‑private partnership mode. And the 39 cities have put together the programs. And we also get clear support from the Chinese government. The Ministry of finance, they issued an announcement related to the promotion of government and the (?) corporation and (?) issued guidance on promoting the government and the (?) model. That's all. Six minutes.
>> XIAOFENG TAO: Thank you. Experiments in China and improve the integrity, policy impact plus and some future program. Okay, I'd like to invite the next speaker. Our next speak is professor Mikhail Komarov, the higher score of economics Russia, consumer, management with the use of the Internet. Please.
>> MIKHAIL KOMAROV: Thank you very much, professor Tao, thank you very much for the invitation to this wonderful session, this wonderful workshop focused on smart cities. But I also would like to remind that CT belongs ‑‑ city belongs to citizens so it's quite interesting to see and to understand change is happening in terms of behavior of cities who are actually customers of different services, right, provided by the city government, provided by different businesses or different companies and so on so ‑‑ next slide, please.
>> Next slide.
>> MICHAEL GINGULD: So as it was already mentioned today customer centricity and IT development actually brought us the special sensing solutions and applications which are widely spread all around the world in different cities and outside and, actually, also connected to our smart devices, you know, called smartphones, right, but I would emphasize we should be focused on mobility, mobility of customers, mobility as a resource so we have that resource. It's not just about allocation but it's about context while we are asking or requesting service. It's about context actually how we understand the current (?) if we're okay with the rain outside or not, which way, actually, we will go toward our local grocery store and so on. Mobility as a resource defined and those described as a more social approach because we trust to sell one's recommendation, right, when we are choosing one or another service. So we should explore those characteristics of mobility in a more smart way and actually, of course, we have a tendency of exchanging humans, right, because we have mobile predictive services we have mobile consultancy, we have mobile management, logistics, it's interesting so I'll tell you more about it but I'd also like to remind you that today there was I would say big step towards revelation of let's say Uber, you know, business models because you probably heard that the high court said that Uber is not Internet company, it is a taxi company which means that they're going to pay taxes. I mean incidence of use services, incidence of humans using new platforms. There should be some consequences in terms of taxes, of course, right? We will see how it will go, right? But anyway, trend is here. You know, we are trying to exchange humans, we are trying to have peer‑to‑peer connection and so on. Next slide, please.
But first I would like to shortly tell you that one of our research projects which we conducted with my colleagues at the school of business informatics was focused on agent‑based modeling of social level things, what is social level things, so network of shared services by people and devices via social network. So actually both of those agents, humans or devices, they exchange information, they exchange data, right? So main focus was on ‑‑ main focus was on behavior, actually, of humans, while exchanging the data information with devices. And the idea was to make, you know, modeling of, you know, agent‑based modeling where humans, they are presented as agents, right, as well as smart devices. Actually, they also can establish network connections with each other and with devices itself and so we tried to research, tried to answer to, you know, the question will that connection shape humans' behavior or not.
A theoretical study shows that 90% actually, as you can see here, average resource request from the humans towards services, towards, you know, smart devices, will be reduced by the influence of those smart devices. Because they all connected, you know, with each other through the network and, actually, they might influence on all be here they can shape it and it's quite interesting because you can think about some smart TV solutions you can think about some default settings, how often do you change them, how often you actually I don't know change or not download some ‑‑ or not use some preinstalled software or some apps or something like that. So that's shows our behavior changing under such influence. Next slide, please.
Another interesting case was conducted, you know, research was conducted recently with my former master students, big data systems program, was about researching and actually trying ‑‑ managing, managing customer flows, but not in store, out of store. So managing, you know, customer behavior, managing customer flows with the use of mobile technologies, of course. And the idea was actually ‑‑ next slide, please ‑‑ idea was to utilize, you know, location, right, of potential customers or actually existing customers. That's why I won't be focused on privacy. Next slide, please.
So we tried to analyze possible use cases how we can actually change someone's behavior working one best back home or using one route back home to the office how to change the behavior how do you ask customer to use another. And, of course, you understand, you know, there should be value for the customer. So value for the customer, what? So providing additional discounts for some products that customer would like to buy in one particular shop. Probably on the way home back, right? Next slide, please.
And so we, you know, we try to analyze also how ‑‑ some external factors, my influence on the customer behavior like, you know, whether some other things additional services and so on. Next slide, please.
So there is some potential use cases which we try to explore, so when we know that those customers, they are members of special loyalty programs in retail, right? And so actually we know their rules because they have installed preinstalled application. And so they, you know, they agree to share their allocation with us ‑‑ their location with us and we know actually when those customers, you know, go to the shop, to the retail store, and, actually, what do they buy there, right? So we know at least the products there so how we can change their behavior. Of course, you know, first we should analyze our retail stores. Please next slide.
And if you know that one store, you know, has too many products of one particular, I don't know, too many products in one store and less products in another store so we can manage, you know, logistics by using our customers. So why customer would go to another store in order to get additional discount. Then we would use, you know, delivery company delivering one product from one store to another, right? So that's where we have value for the companies and value for the customer getting discount. But it is quite interesting ‑‑ next slide please ‑‑ you know, there are different use cases, or three use cases, you know, we also can predict that some ‑‑ you know, while a potential customer or, you know, a customer would go to another store for instance, so we would probably recommend him to go to our store, you know, within the retail chain in order to buy, you know, products the customer prefers with some additional discounts. Next slide, please.
It's quite interesting because customers, they like being identified in the shops. And, you know, they would ask and, actually, there was research about it, right? Next slide, please.
I won't go in details ‑‑ next slide, please.
Models there. Next slide, please.
Trying to identify mobility patterns of our customers or existing customers. Next slide, please. Next slide, please. Develop special software. Next slide, please.
Go through some test bets, right, identifying routes and identifying, you know, placing some shops and now we ‑‑ next slide, please. And now we're in the process of discussing actually implementing this idea of managing customers, managing customer flows with the use of, you know, those customers, right, reducing expenditures and logistics, reducing, you know, or increasing number of sales for the local stores, right? But there are, of course, different limitations. Actually, there is no ‑‑ you know, there is no limitation mentioned in this slide because it depends also on the country, on the different mind sets, because some people, they probably wouldn't go to another store far away, you know, from their traditional route even for the discount, right? But we know there are so many people who would do that so that's why we can explore it and that's how we can change, you know, our potential customers' behavior, right? Thank you very much.
>> XIAOFENG TAO: Thank you.
[APPLAUSE]
>> XIAOFENG TAO: A very interesting topic. Effect on consumer. Okay, I want to invite next speaker, the speaker is Mr. (?) of (?) technology, His talk is about IoT. Please.
>> UNIDENTIFIED SPEAKER: Thank you very much. My topic is about the vulnerability and hacking IoT. That security matters about and thank you very much for being at this session. And I would like to thank also ‑‑ thanks also to RIPE NCC who has invited me to this Internet Governance Forum.
So my topic is about how the IoT security governance can be managed in general. This is ‑‑ why this IoT is a big issue, because this is not traditional, there are a lot of nontraditional approaches that we are introducing in IoT systems. This is ‑‑ we are breaking the security rules which are traditional are introduced and this is a different way and different challenges that we need to overcome, and these are both security challenge, technology challenge and also the ‑‑ in general, the security new way of handling it in untrusted environment.
Why IoT is different? The first thing is the cost. We are thinking that we are deploying the millions of devices and per device, the cost to bring down, the cost is very critical. While bringing down the cost, we would like to have the highest smartness in these devices, so from one side it's a kind of trade‑off when we are talking about this. How much hardware, how much intelligence to put into these devices and from the other side to keep down the cost. The cost is very critical because the scalability of these systems are very high.
And because of the cost, there are many other side effects that are affecting to this cost issue. This is a low level of computational resources that we are putting into these devices because of this keeping down the cost. From the other side we are having the low physical security because these devices are unattended, they are distributed among a very wide special distribution, and they are remote, they are physically accessible by ‑‑ could be even public. And these devices should be secured in this environment. When we are saying on unattended devices when we are saying accessible devices we mean that anyone can tap into this device, anyone can tap not only the device but also the environment and the communication means, whether it is wire communication or wireless communication and to secure these devices in this unsecure environment. Another thing is the power supply. These devices should be self‑sufficient and these devices should be ‑‑ miniaturization is a high component of this because we would like to have tiny components that will be very chip which we can anywhere in a large number and deploy them and connect them to the local proximity gateways and hubs and through these gateways to come out to the Cloud and then collect all this information, huge information, because we would like to have a lot of knowledge about different things to do a better analysis and better information visualization or processing or intelligence.
And another thing is the mobility that is different because IoT will not work in our general mobile environment. It will not work in 3G or 4G. We need 5G because we need millions of connections per tower and simultaneous connections should be very ‑‑ on high speed, 1 milliseconds of delay latency will be necessary, and it will be necessary to have a very high amount of physical simultaneous connections, and this is what 5G supported.
And then administration shift. This is a different approach. Administration shift means that we are ‑‑ would like to put some intelligence in the devices, some plug and play features in these devices, in such a way that these devices will be deployed not by techies, not by technical peoples, but by anyone who can buy these devices and deploy, and they can configure on their own, they can put different types of sensors as per their needs, as per their system application requirements, and this will work on its own so that we would like to delegate a certain type of administrative function ‑‑ functions that regular administrators have been doing. So we would like to have some kind of features to put into the system so that people can select different types of sensors, they can do their own conflagrations without being techies and this is administrative functions that can do a loose configuration and there should be enforce certain rules for this to be done properly.
So IoT security challenge is the first thing, as we said, it's violate a traditional single point of entry concept which means that when we are talking about the traditional networks, this ‑‑ the local network is having a single point of entry whether it is one single device or maybe redundant devices for which you are entering there is such a limited number of entry points to your network, whether it is a local area network or it is a secure wide area network you have a limited number of entries for which you are going into the network and then you are protecting your network through these interfaces.
When you are putting your IoT sensors, whether it is on a municipality level or it's ‑‑ could be your land network level, you ‑‑ LAN network level, you are opening holes, you are having a different approach you are opening holes, security holes, in your network which means that it was totally unacceptable in a traditional approach to do this. So you are putting some devices, whether this is an actuator or this could be a sensor, these are the source of information or a controllability point that you can control or manage or this source of information can cause harm or can cause vulnerability from that point, and this can be in your network. So the traditional approach of securing everything as a castle and having a limited number of entries to protect them you are widely spread, you are just open there was clear definition of trust environment and untrusted environment and there is a DMZ in between and you are opening the whole network and untrustful network and you are working in the new network where everything is untrusted.
To achieve the low cost of the system you need to handle the high complexity of full security because the design, scalability and usability point of view, you need to add additional functionality, having these limited resources so that these additional functionality can create a plug and play systems between different components and different layers of IoT. When we are saying different layers we mean sensor, physical layer, physical layer or actuator layer and gateway and hub layer and Cloud, central server layer, and also, above that, also analytics and artificial intelligence software, different layers.
So in each layer because this is a tightly integrated system, from sensor to artificial intelligence all the chain consists of different layers and all these layers should have a kind of additional functionality that you need to have and this is a design challenge of the system.
Another thing is that you need to rely on limited competitional power and low memory which means the functionality of the traditional computers, the generic purpose computers you are distributing among your system. So you cannot use the generic purpose computers, generic purpose embedded controllers, if you don't want to have to plug into your embedded controller let's say HDMI display adapt or display to see something, you don't need to have this hardware, right? So you have a limited capacity to build in to bring down the price and this is a challenge to distribute your functionality among the system, among the distributed system.
Then connected greet and time enclosure. Sometimes there could be a critical application and real time working so that regular embedded space unit connectors may not work, you need to collect the data in real time and this is another challenge that the speed function is critical and then you need to have this kind of connectivity in each and every segment and any kind of privacy breach ‑‑ and the growing mobility. So these are the cyberattacks well‑known in different paradigms, confidentiality, data service, data ability and integrity, in this aspect think bots are added, things are component things can be source of attack. This is hacking different layers. We are well‑known the communication type of attack, this is where you are compromising the data to the packets or the and UTM devices usually are handling this but there are also software attack that also you can ‑‑ it's a threat of how much effort usually the hacker is putting in to hacking some system and what is the benefit they can get out of it.
So communication attack is the cheapest one that anyone can do even ‑‑ there are a lot of scripts and bots San software that you can ‑‑ the software attack requires more effort and this is the more power is necessary to do this. And the recent hardware attack which is the most costly one and hardware attack when we are saying hardware attack we are talking about when you are cracking the system through the debugging devices or maybe you are putting extra hardware into your device to reach your accessing your ‑‑ having a back door to access to the system.
So what we are governing in IoT, we are governing all this inter‑‑ all these features that are closing the damage. The traditional approach, we're damaging data only. We are talking about the damages that can cause physical harm because of the security and vulnerability issues. And this could be the firm way, hardware, or embedded scripts at different levels and there is also the functional safety challenges that are coming with the self‑driving cars and there are a lot of efforts in this area when IEEE developing the 26262, the FUSA phonetic), functional safety standard. And also, of course, there are a lot of things in the wireless area.
So the governance, our strong belief is that there are a lot of things in technical area and we are thinking that the governance ‑‑ the key component of governance is the standardization. Standardization that are worked by the technical people. So technical people should resolve this issue first. This is the key component. And the second thing is to force the standardization to the vendors who are producing these components. And in a summary, I would like to say that the security measures where we are having about the standardization issues, this is ‑‑ there are a lot of efforts are going on now by IEEE, by ITU, by ISO, a number of standards are developed. But this will not work if these standardizations will not be forced to the vendors. Thank you very much.
>> XIAOFENG TAO: Thank you.
[APPLAUSE]
>> XIAOFENG TAO: IoT security issues in detail. So now it is time for Q&A. Your questions, your comments. Please.
>> UNIDENTIFIED SPEAKER: My name's (?) MAG member from China. The future of IoT and this (?) is interesting but after listening to the panelists we are also aware there are lots of challenges the bandwidth, hardware, software, hacking by enemy states or terrorists, you know, the host city, the key infrastructure, the key platforms that canning attacked by terrorists, the consequences can be catastrophic. And that is one thing. And another thing is that to ‑‑ to set up such kind of system, we have a problem of data sharing. Some data is generated by the public system like the traffic control, vulnerable surveillance owned by the governments. But we also need data from those commercial platforms like DD, taxi, et cetera. I'm wondering how the governments, the city governments or national governments share those data, do we need to put legislation in place or the government just struck a deal with each company separately? How do China and other countries are doing this in this regard?
>> XIAOFENG TAO: There are two questions. You can give a brief answer. Second question I want to (?)
>> UNIDENTIFIED SPEAKER: You know, I think that to ‑‑ when we are talking about how government should handle these issues, I think that there should be a kind of approach where the government should take the road of confluency issues that should be forced by the vendors because there are a number of standards and if someone is compromising, even from hardware or from software, they should be punished, this is my belief, because government cannot go into the technical details into the system how they work. But the standards are forcing this security issues from tech ‑‑ technical perspective.
>> XIAOFENG TAO: Thank you. Please, Mr. (saying name).
>> UNIDENTIFIED SPEAKER: The Australian government has adopted open data policy and sharing, so the premise is most public government data should be available unless there are some other concerns like privacy which limits their availability. But in terms of private and public enterprise‑sharing, I don't, at the moment, see much of that happening, but it would need to happen. But the issue is some private company's thing that data they collect are they proprietary information? Some companies actually trade with those information, they actually use that as a way of, if I mention data, they actually sell or license the data for a fee, so there are implications which is kind of complex but that would have to be managed as those issues are being dealt with so it's a matter of negotiation with those private companies for access to data.
One of the things I could say is that if the government is allowing licensing of operators to come to the country to operator with their data that they'd like to get access to maybe it's part of that negotiation to ensure that it's embedded in the contracts, in the agreement to start with so that the data could be used or at least determine as to who owns the data and how that could be shared and controlled.
So these are some of the mechanism framework that you can start thinking of. But currently I think it is really doing with private enterprise and it's a matter of negotiation to start with, thank you, zoo Zhao, thank you. Are there some questions from people?
>> AUDIENCE: Yes. Actually, we do have a very active participant online and quite a few questions. We got seven questions now. And I tried to (?) on this question. I found there are some questions, it has been on‑site about security of (?) our speaker already answered actually, it's a question (?) from Pakistan, have a question from (?) about the measure of security and NOTTY. So I think we need, we have to have time to talk about other questions. It has been divided into two categories. Some of the questions comes on the environmental, the participant that until now which mechanism or process can be used in smart city to face environment, pollution, and the (?) and what's the value for this challenge, for the future structure of smart city and shall I finish?
Other question focused on the people. They want to know the effect of IoT or the people, like replacement of human resource and also the people living last developed area or country how they can adapt to the change brought by the IoT device.
>> XIAOFENG TAO: Okay, okay, thank you. The first question is to you. Can you answer the question on the point of pollution?
>> Yes. I don't have any material about pollution.
>> XIAOFENG TAO: Okay I think the first question about pollution, actually the countries are (?) for example (?) census work, maybe a possible solution for pollution. The second one, second one is for the effect, effect, people.
>> Especially the last.
>> XIAOFENG TAO: The same story happen several time in the history. I think ways we can get solutions or get opportunity to rephrase this kind of people's opportunity. Yes. So please. One question because we're almost out of time. One question.
>> It's me, okay? Thank you very much for your presentation. I have a question ‑‑ sorry my name is (saying name) I'm coming from France, and I teach supply share and purchasing so I'm very interested in the limitation of the industry and the behavior.
The key question is there is a big difference between datas and behavior. Behavior related to your predictive compartment and if we go further, it ‑‑ it is related to your soul, your personality. So we don't deal with data, we deal with the nexus of the humanity so who is able to control this, and the second question: Where is the right for anonymity? Because we are a civilization of movement. And our data, we can follow our data through the ‑‑ through our Internet action on the web and also through the Internet of things. So thank you very much.
>> XIAOFENG TAO: Thank you, please.
>> So, yes, quick answer.
>> MIKHAIL KOMAROV: I'd like to remind you of development of trains, jets, I'm not talking about cars, right, and just imagine how many people first are not using trains, jets, you know, cars, right? You all remember that very nice picture, you know, I think New York, right, where we had horses and carriages there. And after several years, you know, after that we had cars almost everywhere, right? And so the same happening here, you know. As humans we are not used to, you know, to change our behavior and adopt some, you know, new things unless we understand the value. And so I think the main idea is trying to ‑‑ to explain the value, you know, of IoT, right? Of course we should also be careful about privacy and personal data protection. That's why we should ask people that usually would like to participate in a project or not, right? Thank you.
>> XIAOFENG TAO: This lady first, please.
>> AUDIENCE: Thank you for the session. My name is Natalia Footage (phonetic), I'm a researcher based in Washington, D.C. My question relates to 5G and to what extent these technologies will truly be based on 5G only or you also have technologies based on a licensed spectrum, you know, connecting all this.
>> Yeah, 4G, 5G, 5G3 scenarios, one scenario is IoT, massive connection, that's for 5G for MP IoT, 4G, 5G names massive connection for 5G. Thank you. Okay. I'm very sorry. I'm out. Thanks again for the speakers for their excellent presentation. I would also like to thank everyone who joined us this afternoon at this workshop. I believe that now we have an understanding of the community development worldwide from Asia to Europe to Oceania where we can contribute to the development of IoT for smart cities. Thank you, thank you very much.
[APPLAUSE]