The Smart City would have to plan to harness the power of Internet of Things (IoT) by laying a strong but flexible foundation, capable of growing with the global emerging standards in IoT techno space, writes Nawin Sona Natesan, Managing Director, Maharashtra State Cotton Growers Marketing Federation Ltd, a special article for Elets News Network (ENN).
IoT : Heard of but not seen
The Internet of Things is in a way a misnomer. It is not, as popularly believed, just connecting more things to the Internet. It is more of “Internet by Things” and can be an “Internet for Things” too. In IoT, the Internet is not based on the TCP/IP protocol stack, designed by and for humans or human intervened communications. The IoT universe would be a largely M2M universe, where Machine to Machine communication would be the norm rather than the exception. Imagine failed street lights “ordering” their own replacement, clogged drains calling the cleaning machine, where vehicle breakdown automatically run internal/connected diagnostics and call a tow truck. IoT would indeed build a Smarter City. But we need to appreciate the difference between IoT and the traditional Internet, which I choose to call, “Internet of Humans”, or “ IoH” for lack of a better word. With the advent of IPv6, 340 trillion trillion trillion (not a typo) devices can get their own IP addresses, unlike IPv4.So, there is a lot of space. But for what?
IoT : A Re-introduction
What do we mean by the IoT? Very simply put , it would be “A Network of Networks, of very large number of continuously or intermittently, low powered , connectable devices, sensors, transducers and actuators, connecting over a large variety of network technologies, delivering raw unprocessed data to be processed and analysed, at edge or near edge or remotely, through cloud based algorithms to produce meaningful information on the state or change of state of the physical or cyber physical system being monitored and or acted upon, and may be used as part of a cyber-physical system, to act in a decentralised manner.” This is my own definition and I have tried to capture the wide gamut of the elements involved.
First of all IoT is not just a network, a WAN or LAN, it a network of networks. For example, a person would himself or herself become a small network, a PAN ( Personal Area Network) - the mobile, a laptop, a pad, smart watch, fitness reader, new age wearables or the clothes themselves together. And this network in turn would connect over the Internet to off-body devices, servers and cloud based applications to derive meaning such as health status of the user, movement, geo location etc. But when this PAN interacts with other networks in the Networked City, it would intermeshing with other applications and contribute to gather meaningful information to add value to quality of life, quality of work and leisure, as well as health safety, personal safety, transportation efficiency among other things. One can see the power such a system could have in emergencies. A medical emergency on the personal network, say a heart attack, could auto trigger an ambulance call, to the nearest smart hospital, transmitting the geo location, the ID and possibly help him be tracked. Auto emergency message may also go to family and friends, and the ER could also be alerted. Lives may be saved.
All this is practically and theoretically possible. But it is not easy. Here we switch gears into the pragmatics and look at the issues involved.
A2F Framework: I have listed the very basic components, for mnemonic convenience, what I call, the A2F Framework.
A - Analytics- All device data need to be analysed in real time, that is a “Velocity” problem in Big Data terms, i.e. Streaming Analytics. The challenges are many, including data loss, irregular data, intermittent transmission. Data Modelling becomes the key challenge. Various network data models have to interact and be Machine Readable. The Analytics should emit meaningful approximations, predictive and prescriptive information to the user, without which a city level network of networks would be meaningless. So time, context, reliability, interconnectivity are sine qua non.
B- Bandwidth- all this real time analytics requires large bandwidth across different networks, it is a “volume” problem in Big Data terms. The backbone must be robust and expandable.
C- Connectivity- The network may be Wi-Fi, wireless, 3G or wired internet etc. 6LowPAN, 801.15.4 (Low Rate Wireless Networks), 801.15.6 ( Wireless Body Area Network), ZigBee and other protocols are available. Wireless mesh networks, bluetooth, radio frequency also give a wide variety of connection options depending on power, transmission distance, data cycle and data representation requirements.
D-The Devices- First of all, we should be clear that the IoT infrastructure is not a mobile-based or a tablet-based one. The following are the main features of IoT devices:- servers and cloud based applications to derive meaning such as health status of the user, movement, geo location etc. But when this PAN interacts with other networks in the Networked City, it would intermeshing with other applications and contribute to gather meaningful information to add value to quality of life, quality of work and leisure, as well as health safety, personal safety, transportation efficiency among other things. One can see the power such a system could have in emergencies. A medical emergency on the personal network, say a heart attack, could auto trigger an ambulance call, to the nearest smart hospital, transmitting the geo location, the ID and possibly help him be tracked. Auto emergency message may also go to family and friends, and the ER could also be alerted. Lives may be saved.
Low power - mostly would operate on sleep mode, or very low Energy, for reading, transduction, and transmission.
Very low computing ability- are not designed for in-situ computation, i.e. these are non-von Neumann devices, without the RAM and the chipset to run any complicated algorithms. They would run on CoAP, Constrained Application Protocol (instead of http).
Size- they are largely (no pun intended) small devices.
Embedded or near embedded systems- may be part of a water supply network or say pollution or traffic telemetry network and may be built into other more complex devices, such as cars, metro rails, or into infrastructure like buildings, roads, bridges, street lights etc.
Sensors, Transducers or Actuators- could be either sensors, temperature, humidity, emission, vibration, etc, transducers, or actuators. They may be automatic or remotely controlled based on simple algorithms. For e.g. signaling devices, say when garbage threshold is crossed, a signal may be sent to the garbage truck, or spill detection in dry environments.
Life cycle- they may not be permanent devices. The life cycle of the device may be short or long depending on the design. For long life cycle devices, associated issues include retiring of devices and device identities, security, environmental sustainability, recyclability.
Networked- the devices are essentially networked. By networked, we have to understand that this does not mean the Internet.
E- Energy: How to power the devices is one of the key challenges in IoT.
The power based classification would be:
- Connected power
2. Battery Powered.
Ambient Energy- based energisation of devices
2) Transducer based power - For e.g. passive Wi-Fi based energy harvesting is a breakthrough technology wherein the device will be powered by other energy. Vibration of a bridge may generate enough mechanical energy which can be converted into the small energy requirement for a embedded sensor. One day maybe traffic noise may be converted into energy!
F- Fog Computing: After all these years of being in the Cloud Computing, the next stage in computing is already forming for IoT. And it is the Fog. Fog Computing layer, is the closest, decentralised layer from the device, where computing will take place, Graph Theory and Network Theory terminology, Edge or Near Edge Computing, etc. The IoT devices being low computing ones, would just transmit and transfer the data to the nearest data center- which is called the FOG. This small data centre could be for a building, connected to larger cloud for further collation or analytics.
IoT and Smart Cities: The flurry of activities gearing towards Indian Smart Cities, the need is to understand doing a Smart City project is not simple. Nor can it be a piecemeal, organic, fractured, completely privately funded collation of activities. In other words, startups are important, but they are at the end of a large value chain. And the State role in IoT is huge; not just supporting a cottage industry of IoT systems, but to establish the foundation.
Role of State: The State has, in my opinion the following major roles:
Establishing the standards for devices and networks. IETF and BIS need to talk. Master Blueprint for the IoT networks- infrastructure, land allocation, connectivity corridors.
Security Standards- with large public data and service delivery depending on IoT communications and large outsourced devices, each device and data transfer and access protocols must align to State and industry decided security standards.
Interoperability Norms- I am not saying the state should support one single standard, because it is a war out there, and the IoT space is a primordial soup of alliances between different types of M2M communications, security standards, and the State has to be careful in choosing one technology versus the other, lest it gets locked into a system which becomes incompatible in the long run.
Disaster Recovery and BCP- which I think cannot be over emphasised. As a public service, the DR and BCP incentives in policy should be very strong and supported in tax incentives and such.
Security and Privacy- the state should be guarantor of privacy. And security standards must be enforced to present large scale attack on public services.
What is the road map for Smart Cities?
I see Smart Cities in two forms
1) De Novo Smart Cities (DNSC), say like DMICDC or Amaravati or New Raipur, etc
2) Retro-Fitted Smart Shahars (RFSS)- Ajmer, Nagpur etc.
The challenges are very different. The road map would be too.
A: De Novo Smart Cities (DNSC) IT Masterplan should be made.
Major System wise masterplans must be made - Transport, Communication, Smart Grid, Water Supply and other Utilities. I strongly feel IoT or data infrastructure is at once a technical policy issue, a town planning issue in terms of land allocation, a infrastructure issue , in terms of building code, road specifications.
B: Retro Fitted Smart Shahars (RFSS) The Smart Shahar, is a pretty organic thing. It has grown by push and pull of economy, politics, people, and technology adoption by citizens is often demand driven, or supply driven (market). My strong opinion is that instead of costume cladding the city with smartness , by way of giving citizen level services based on IoT, which is tempting to do, the city government should work on the back end services with IoT as a support system.
As a first stage, utility services can be made more automated, self diagnosing, with faster fault detection, faster fault correction, with better preventive maintenance, with better cost and energy efficiency, and the citizen benefit would be an outcome instead of a goal of IoT. IoT is a power tool, a whole arsenal, really, to tackle inefficiencies of old systems, and would augment human effort, support decision making, and also give large level of meta-level system state understandings like traffic, water, waste, energy for example. The A2F Framework would be helpful in charting out the roadmap.
Conclusion: The Smart City, or the Smart Shahar would have to plan to harness the power of IoT by laying a strong but flexible foundation, capable of growing with the global emerging standards in IoT techno space. It should foster open, competitive growth, with interoperability, security and most of all environmental sustainability, and be careful about what kind of standards and technologies, and devices would be encouraged in the city.
(Nawin Sona Natesan, is an IAS officer of Maharashtra Cadre, in the rank of Secretary to Government. He has done professional courses on IoT and Big Data, from Computer Science and AI Lab, MIT, USA. He has contributed to a book on Big Data, published in USA. He also codes in R, and formally learnt Machine Learning.)