6 Other types of networks

Prof. Bhushan Trivedi

Introduction

The current network nodes are not confined to computers. Mobile phones are most common types of other members. IOT devices and sensor nodes are other types of members who often join computer networks and exchange information with them. Home networking, or networks which are confined to home and connecting household appliances and devices based at home like TV sets and video recording devices form yet another type of network. Sensors like ones installed in buildings for stress measurements, or cattle fields for monitoring cattle, or jungles for habitat monitoring, etc. are critical for current systems. They are also examples of sensor networks. There is a device, known as RFID or Radio Frequency ID, which can be attached to any real world object to make it addressable network-wide. RFID attached real world objects can be tracked by other types of devices called readers without requiring a line of sight between the object and the reader. Smart cities that are lurking over the horizon, demands a huge amount of data to be produced and processed for the intelligent decision-making process. To generate such data, many new types of networks are needed, for example, a network which is made up of sensors installed in a parking lot, indicating whether the cars are parked properly and generate fine amount otherwise. When such devices are also connected to the Internet, they are called IoT devices. We will brief about those types of networks and what they demand additionally in this module. We will start with sensor nodes.

Sensor nodes

The sensor nodes are a special type of ad hoc networks. The nodes of this network are tiny sensors. The sensors may be mobile or may be stationary. For example, sensors which are designed to monitor environmental or physical parameters like temperature (of the car engine), pressure (in the car tire), speed (of the car) for example, in an automated car. They are also used to monitor the surrounding like bird’s nest in habitat monitoring, camera’s recording in an automated car, video in case of the surveillance camera, etc. They are also used to monitor fields for determining conditions of crops, amount of water or fertilizer needed for a typical case, etc. Sensor networks are also used in battlefields to monitor enemy activities. They are also found their usage in many industrial and consumer applications. For example, smart bulbs and other electrical devices are connected to form a kind of sensor networks which can indicate a bulb or some other device failure automatically to the administrator and also detect an absence of human being and thus report energy waste to improve energy consumption. There are sensor networks implanted in jungles to prevent and control fire, there are sensors which detect pollution at various parts of the city, there are sensors which can detect the possibility of the landslide in hilly areas, measuring the quality of water in urban areas and villages. Industrial segments use sensors to detect metal fatigue and other problems with machine health.

The tiny members of this network are huge in numbers in most cases. The jungle habitat monitoring or cattle field monitoring networks have thousands of such nodes.

Considering the nodes are to be installed and used in large numbers, the sensor node is designed with lesser memory and computing power to keep costs under control.

Let us list the characteristics of sensor networks as compared to ad hoc networks.

1. The ad hoc networks have nodes in the range of 10 to 100 normally. The sensor nodes with thousands of nodes are common.

2. Sensor nodes can be immobile. It is not possible for anybody to remove them and reinstall them. For example, consider the case of sensors installed on a bird’s nest for habitat monitoring, or Implantable medical devices implanted in the human body. When low power and low memory devices are coupled with this requirement, processes which demand less memory and less computing power are essential. For example, the conventionally used routing algorithms prevalent in computer networks are not possible to be used here if they consume more memory or computing power.

3. The sensor nodes may not be individually addressable, and not required to be addressed as well. for example, Jungle fire is detected by the bunch of sensors who reports it, the receiver does not need to know the exact sensor node name or number but only the location. Water quality sensors may indicate that water in some area is not of required quality, that is enough, one does not need to know which node reported so. Let us take another example, if I receive the message from the bunch of sensor nodes that the temperature is 31 degrees, I do not need to know which sensor node sent that message to me. Another point is, when there are thousands of nodes, out of which many have stopped functioning, keeping addresses of each of them and keeping track of who are alive may add substantial overhead which is not needed.

4. The power consumption is more critical in sensor networks than the ad-hoc networks as in many cases, it is impossible for the designer to recharge those node batteries, for example, sensors implanted in the human body. Another example is sensor node is placed in a bird’s nest to monitor the growth of its offspring. There is no point removing that sensor after two days and recharging and fixing it back in the nest. Sometimes the power consumption issue takes the highest priority and other decisions are seriously influenced by that decision. For example, if the IMD (implantable medical device) is expected to run for 15 years and the battery implanted barely allows that, we forgo security measures (like making sure the command is coming from the designated doctor and nobody else by providing a security service called authentication), or others do not learn about patient’s medical condition (revealing patient’s detail is against the law in many countries) when the IMD is communicating to the doctor’s mobile or laptop (by providing a security service called encryption). Most ad hoc routing algorithms use intermediary nodes to relay traffic, the power consumption issue might demand them not to take the shortest route passing through a node whose battery is dipping. The conventional routing algorithms do not consider power consumption as a factor in deciding the route and thus other routing algorithms may be preferred.

5.Unlike ad hoc networks, the sensor nodes are placed in hostile terrains like jungles, battlefields and subject to damage, capture and harsh atmospheric conditions. The routing and other algorithms to manage communication must consider this fact.

Smart Home and Smart City Networks

Having a network at home is quite an attractive idea and many models are appearing on the horizon which caters to the needs of home residents. They are still in infancy, but quite interesting to warrant discussion. There is no serious market product yet. Home networking is not confined to devices like laptops, printers and other home devices but one which includes household appliances and other devices like TVs and Microwave Ovens, ACs, refrigerators, washing machines, bulbs and heaters.

Smart homes are able to manage the homes in a clever manner, to control TVs, security systems, and appliances from even a remote location using your smartphone or a laptop.

What could be the advantages of having a home network apart from the ability to control remotely over the Internet? Here are a few examples for you to learn.

1.Your refrigerator reminds you that you are overeating, or you are eating food not good for you.

2.Your front door security camera informs you that some stranger is trying to break in, or the bedroom camera informs you that the babysitter is taking a nap instead of doing her job

3. When you are looking for your spectacle, or car keys or something similar, the device can make noise for you to notice

4. A phone receiver realizes that you are upstairs and transfer the call upstairs.

5. When there is nobody in the room, for a while, the lights and the fan are switched off.

6. The room temperature is adjusted automatically with requirements.

7. The house is on fire, the fire alarm rings the fire department, provide location and take necessary steps, like spraying anti-fire gas.

These scenarios appear like a fairy tale or straight from a Rajnikanth movie, but they are actually possible with current technology. Unless they become cost effective and viable, it is hard for these solutions to become commonplace. Interestingly most such devices need an Internet connection to communicate with others. When such devices are also equipped with an ability to connect to the Internet and communicate with it, it is known as IoT or the Internet of things. When such device helps an entire city getting enough data and served based on that data, they are smart city IoT devices.

Let us take a few examples to understand smart city IoT devices. Sensors in water pipelines measure and bill citizens for the amount of water they consume, additionally they also instruct the government officers to manage water levels accordingly at the water sources. They can provide some estimation of the amount of water needed in a specific case and can be used in a predictive analysis. The power consumption of a typical home can be monitored on the fly and the residents receive the data anytime they want. The congestion information is made available based on road sensors and travelers are informed well in advance to avoid congested roads (we do not need traffic updates that we listen to popular radio channels then, you can provide your own path and you may get a precise information about traffic on your path and not a generic traffic update which is useless for you most of the time.)

One may ask a question, when will such solutions be viable and commonplace. There are already some products around and smart city movements taken up by governments but it is still hard to assess the time by which these solutions become commonplace and available to an average citizen. Building a home network, in a true sense, is more challenging than building an office network. Building a smart city network is even harder. Unless such networks fulfill a common user’s demand, they will not in a mood to accept such networks. Here are some challenges listed.

1.Home and city networks need to be more robust than normal networks. A normal user cannot debug a malfunctioning refrigerator which is not cooling and requires the reboot. A smart city network might not be treated with care by citizens and require a much robust design and installation than the conventional network. A home user is less likely to accept messages like “this program has performed an illegal operation and likely to close” or a metro railway commuter “Ticket printing device malfunctioned, the program needs to close and will take 5 minutes to start all over again and print your ticket”

2. The biggest challenge, in both cases, is cost. It is hard to convince an Indian customer to buy an IoT-enabled ₹ 30,000 refrigerator compared to ₹ 20,000 normal refrigerator. When there is not a big market, the cost of devices is unlikely to come down which introduces a vicious cycle. Unless there is a widespread use, the cost won’t come down and unless cost comes down, there isn’t likely a big market.Implementing a city network will also have similar problems; for example, passing their (additional) budget in their respective councils. It requires substantial justification in terms of future benefits to sell this idea.

3. Security issues are on the rise. When this issue is confined to computer networks, and somebody is stealing victim’s files, it is one case, when the issue is spread to an IoT device and somebody is playing with the victim’s glucose level, it is an entirely different case. Somebody reading my mail is definitely a concern for me, but somebody who can break into my house by exploiting some vulnerability in the house network is altogether a problem of a different level. Somebody taking an entire city on ransom (like shown in one Die Hard series of a movie) by breaking into a city network is a problem of even a higher level.

4. Those who are managing conventional networks deal with backward compatibility issues, upgrading software, databases and apps, and installing and uninstalling them regularly. Expecting a city servant who is managing pump services for water, to upgrade and reinstall OS for the manager software is not likely to work and the political parties insisting on so are not going to win elections. Similarly, when a new shelf is to be added in a refrigerator, and the manufacturer demanding an upgrade of the refrigerator OS (and charging for the same) is not going to have better customer relations.

5.People expect home devices to continue working for years without real maintenance. An OS getting obsolete in six months or two years and the vendor stopping support for that OS after a few years is the idea that is not welcome.

6. The home network is preferred to be wireless as there are many issues with wires already (power wires, telephone wires, desktop wires etc.). However, the wireless solutions are less robust and less secure, so one must strike balance.

7. Making the devices concerned to talk to each other demands additional protocols. For home networks, a protocol known as G.hn from HomeGrid Forum and ITU-T is likely to be primarily used at physical and data link layers but there is no clear cut winner (there are many competing protocols, ZigBee, 802,1h, MQTT-SN, SMCP, XMPP-IoT etc. On the other hand, at a higher level, many IoT devices use conventional internet protocols and avoid using any other protocol. However, it is not easy to use conventional TCP/IP for these devices. Lack of standard protocols inhibits mass production, introduces interoperability issues (home router of company X does not work with TV of company Y), and prices cannot be reduced. This further inhibits the demand and use of such devices. There are many attempts to tailor existing protocols for low-power wireless communication by the very large number of IoT devices for city networks and beyond. One such protocol is LoRaWAN (Low Power Wide Area Network) which is designed for IoT devices spread across a large geographical area to communicate wireless using less power. One more is 6LoWPAN which we address in module 32 and 33.

8. One more issue is about customer’s mindset. A normal IT user easily accepts software flaws and hardware breakdowns. For example, if software malfunctions, an IT user reinstall it or try solving it some other way without considering the vendor responsible. Even when a vendor provides a link, the user knows how to download and install a software update in most cases. A normal user holds vendor responsible if the supplied product is substandard or not working as per expectation. Meeting customer requirement in this segment is much harder. For example, if the customer has not patched his home system and burglar broken into the house using that unpatched system, a customer will always hold the home networking company responsible for it, exactly opposite to the IT industry where the customer is held responsible.

9. Problems related to the power outage can significantly hinder the spread of such systems, especially in villages where this is a common issue. No one would like to stay outside the house in hot afternoon when there is no electricity and thus the door opening system is not operational.

Despite all these issues for mass adoption, and experts’ opinions are divided on it, there is no doubt that the home networks will soon be commonplace. Whichever company succeeds in coming over these obstacles and comes with a product that fulfills user’s expectation will earn a huge customer base and thus, money. For city networks, it is much more than a bigger scale home network. A home network, being owned by a single party, there is no likelihood of mismatch between two device speeds and capabilities. Unlike that, a post department computer may be very old and the lines are of low bandwidth while some other department might have much better set of computers and lines. A user accessing them can feel that difference quite frustrating, especially when his own device is high-end and thus his expectations are such. Another point is, it is not likely that a networked home appliance is used without care and broken every now and then. A city network element, for example, a water meter, might just be broken by a careless employee and do not even report that. A city network of devices must be more fault tolerant and robust compared to a home network. Another difference is of scale. The home network may not generate data which can swamp the network but the city network can, based on the amount of devices it is getting data from and the amount of information input for processing and generating reports.

Networks based on Radio Frequency Identification Tag (RFID)

Other networks that we have discussed so far might take a few more years to become commonplace but not one based on RFID. RFID is being used by many companies today successfully for many of their operations. In this short segment, we will try to throw some light on what RFID is and how it is networked to solve problems.

RFID tags are small, postage stamp size devices which are capable of being identified by any computing device, including our laptops and smartphones by just connecting the computer to a device known as a reader. The best part of RFID is, one can attach it to any object of the real world (garments, sheep, books, passports, birds and spectacles are a few examples of objects already tested), and that object is possible to be tracked automatically like a mobile node in a network. In a way, every item can be part of a network if a RFID tag is attached to it. Once attached, that object can exchange information with rest of the network through that tag. The RFID is a form of a sensor node. It contains a small microchip and antenna so can act as the transmitter as well as a receiver. Usually, it transmits a few things, including its own id, upon query. The RFID tag usually also contain information about its location and few other things needed for their operation (for example price tag of the item RFID is tagged to), it might respond with those values as well.

Another device that is used with RFID is called a reader1, which is installed where the tracking is done (that can also be connected to a conventional desktop or laptop). It can find tags which are in its range and query them for many things including the id of the tag. It reads the response and then transmits its observation to a computer system connected to it and usually stationed nearby. A computer system is equipped with a RFID software or RFID middleware can take an action accordingly. For example, if one buys a shirt from a store with RFID tag attached to it, the computer at the billing desk read the observation from a reader installed nearby and prepares a bill automatically filling information about the make, size, and price of the shirt. The accountant only needs to add a name of the customer if required. Once the computer prepares the bill, it can use it for further processing like make necessary modifications in the inventory database. Databases like Oracle2 has that middleware which one can use to program for such processing.

Interestingly, most RFID tags neither have a plug nor a battery. They are charged by radio waves supplied by readers. This is a very clever move considering our need for a power efficient solution. Such RFID tags are popularly known as passive RFID tags. Other types of RFID tags, on the contrary, have a power source (and thus plagued by all problems of sensor nodes we discussed earlier) and are known as active RFID. Thus RFID has two varieties, self-powered (active) and powered by readers (passive), one can use the variety which suits their application. One more type of RFID tag is called battery assisted passive. It has a battery on board which gets activated by readers, this is best of both worlds kind of solution used by a few vendors, not in vogue yet, though. A passive tag is cheaper and also last longer and thus are preferred over other two options. Tags may be read only, which contain factory assigned a serial number that might be used as a key field for database search. In this case, there is no circuitry to read anything on the tag, they only transmit.

1  The reader is also called an interrogator.

2   Oracle calls it OSBS or Oracle Sensor-Based Services

The RFID tags are low non-volatile memory devices. They include a small transmitter and a receiver. A reader, which is much more powerful in terms of memory and processing power, acts as a master during the communication while tags act like slaves. This is an example where all devices are not considered equal in a given network. The reader is considered boss and all devices connected to it obeys its commands. The reader often has multiple antennas (tags have one antenna), to talk to multiple tags at the same point in time. Reader sometimes broadcasts an encoded radio signal to interrogate tags surrounding it. The tag, upon receipt of the signal, responds back with its identification information. for example, an EPC (the electronic product code) is one such standard used for tagging inventory. This information may be a unique serial number (for example, identify a student by an enrollment number so RFID contains that unique number on student’s ID card), or some product related information (for example, batch number in case of medicine or accession number in case of book in a library, a production or procurement date, and so on.)

RFID comes in three different varieties from the viewpoint of the frequency. A UHF frequency tag is 900 MHz range (we will look at the entire electromagnetic spectrum in module 11 where we will see what this range mean) in the US and around 400 MHz range in region-2 (which includes India). The advantage of this range is that the signal travels several meters so can be picked from a reasonable distance. Tracking driving licenses and permanent pass-holders at toll stations are excellent examples of using such RFID tags. Tracking livestock and wildlife also is done by using RFIDs using frequency in this range now.

The other variety is called high-frequency tag. They are 14Mhz range and can work for a very short range. This is ideal for identification cards like driving licenses, college and corporate I-cards, passports, credit cards, and so on.

An older technology called low-frequency RFID uses lower frequencies which can travel long distance but does not carry much information. This was useful for tracking Livestock as one needs signals to carry over a large distance but with not much information.

RFID is used for two purposes primarily, identification and tracking. Some example is in order. An assembly line product, for example, an automobile, can be tagged. Once it is tagged, the progress of that item can be tracked for the progress. Clothes can be tracked, inventories of them can be taken without physically counting them (Pantaloons in India has a similar system in place), Livestock and pets can also be tracked using a similar technique. RFID tags can even be implanted within the human body to help automatic authentication and tracking. Though such proposals lead to privacy related concerns but are within the reach for those who wants to implement this system, technically. Similar concerns are also raised for other RFID-based solutions. For example, if the attendance system takes into account the identity cards of students to fill attendance sheet automatically, a student must carry one in the class. Unfortunately, some other device which snoops the movement of the student can get vital personal information about him/her. It is hard for RFID tags to be equipped with latest cryptographic algorithms with limitations of memory and processing power. The solutions like passwords and weaker solutions like pin numbers are usually deployed which are easy to break. However, these are common issues and researchers are working hard to find solutions to all these interesting problems.

One interesting application of RFID is to replace barcodes. For example, books in the library now contain RFID tags so the librarian can take their stock without manually bringing the book in front of the barcode scanner. In supply chain management, the items containing RFID tags can be counted, invoiced and packed without any manual intervention. Readers connect to RFID-attached devices and get information from them. In the case of passive tags, readers also charge them using radio waves. The RFID does not require to have physical contact or even line of sight requirement (as in the case of optical readers to read barcodes for similar services). RFID uses electromagnetic fields to transfer data from the tag to the reader. However, the barcode has one typical advantage over RFID tag, the cost. So replacing barcode is not yet common, but using them at other places like driving licenses and passports etc. are quite attractive applications.

Internet of Things (IoT)

When RFID helped one-way simple communication, some solutions demands two-way communication with many additional features. For example, when two vehicles come nearby, sensors embedded in those vehicles should help avoid the collision of those two vehicles. Sensor nodes, which had the capability of monitoring and relaying monitored information over to a computer attached also requires to be extended to communicate over the Internet. IoT is an answer to both puzzles. It is a sensor like a device with the capability to send its information, but not to a nearby reader but over the Internet, to any corner of the world. Obviously, an IoT device is a superset of both these things and likely to replace both RFID and conventional sensors in future.

Smart solutions like smart thermostat (which maintains temperature, in house or car etc.)3, or smart home (which adjusts temperature, lighting, music, TV channels, etc. according to dweller’s preferences), or smart city (which caters to varying requirements of the citizens like smart transport, smart congestion avoidance, smart electricity management etc.) are three examples of smart solutions of increasing complexity. All of them need solutions containing some hardware and software which typically provides a service. We call such a solution an IoT device. IoT devices like an air purifier, smart lighting, ovens, washing

3 According to one report, an average Germen drinks about 150 liters of coffee every year. One interesting application was developed at Fraunhofer Institute of Physical Measurement in Germany. The institute innovated a coffee mug with an IoT device, which is part of the mug, which is powered by the hot coffee itself and sense the hotness of the coffee to inform the kitchen if the coffee has gone cold. machines and refrigerators can communicate their state to the user over the internet to enable him to make decisions.

An IoT device, like a passive RFID device, can be autarky, i.e. generate power from the atmosphere. The footnote 3 is one such example. Another example is wristbands which monitor body temperature, blood pressure etc. which generate power from the difference of temperature between human skin and the surrounding atmosphere.

An IoT device is like a miniature computer, having almost all capabilities of a computer including running a program4. They can act as agents for others. For example, an intelligent shopping IoT device can learn about a user’s buying behavior and suggest offers which users might be interested in, thus acting on behalf of a sales assistant. Another example, surveillance cameras which can learn on their own, and judge an abnormal situation on their own to report owner or police over the Internet, acting on behalf of a security guard are already developed.

There are a few challenges with IoT deployment. Let us discuss some which are considered critical at this point in time.

One critical issue is of privacy of information. An IoT-enabled glucometer and insulin pump can help the doctor understand the condition of a diabetic patient, or a pacemaker reading enables a doctor learning about the condition of patient’s heart, is just fine. However, if a third party learns about patient’s condition in a similar manner, is not desirable. One needs to have deployment platforms ready with the guarantee of observing privacy policy completely.

There are deployment and economic issues plaguing IoT system growth. The hardware solutions are hardly economical and deployment of hardware devices in any location demands expertise which is rare.

The IoT devices that we have currently are not context-aware. Taking intelligent decisions like where and when the accident happened and whom to inform depends on a lot on the intelligence and context awareness of the IoT devices attached to vehicles, for example.

Security is anyway important but for IoT, it makes more sense as it touches a physical as well as a cyber realm. Somebody who can hack into victim’s bank data and steal some amount is one thing and somebody instruct the insulin pump to inject more insulin than needed to inflict hyperglycemia, the low blood sugar level, which can faint the patient or

4 Typical high-end sensors used to run programs embedded within, the rest of the devices which we discussed had no program execution capabilities. even kill him, is altogether different. People have demonstrated how a device can be used to fail a car’s brake based on typical signaling process, is another example.

Smart things like smart TVs, refrigerators etc. are being used by hackers for launching various attacks as these devices have the much limited form of security. Smart TVs from Samsung and LG have found various adware and spyware installed on them, for example.

IoT, unfortunately, overlaps many predefined and so far considered isolated, areas. For example, it tightly integrates hardware and software and also many times have a design that fuses multiple network layers together5. This situation not only confuses the designers, also have many protocols from many vendors competing for solving the IoT puzzle. The sanity in conventional networking solutions, X is working on Ethernet cards and Y is working on IP layer solution, can work on their own, is just missing when a layer is found to work as both IP as well as Ethernet (i.e. it looks at the routing table, and without constructing an IP packet, it constructs the frame, for example).

Researchers are working to find solutions to these problems and how the IoT revolution marches forward largely depends on the answers to above questions.

However, IoT is applied to solve many problems of the real world. So far, they are tried in Environment Modelling, healthcare, home automation, manufacturing, academics, infrastructure management, homeland security and many more areas. The spread of IoT is so significant that we will be discussing them in more detail in three modules, 31,32 and 33.

Smart Cities

Any city contains many networks of varied types. Water lines, power lines, communication lines, transport structure network and so on. A smart city is not a free Wi-Fi city but a city with all its data and ability to crunch that data for decision making. It has a strong network for information flow from small IoT devices and conventional sensors, citizen’s home computers, and their smartphones, to centralized servers and decision-making points. For example, a city bus driver (a decision-making point) can learn about many computers collected at a specific point (based on data from those computers’ smartphones sending data to the centralized server which diverts the information to the bus driver’s device) so he can move the bus there to pick them up. A water management system learns about a leak and stops distributing water on that specific line. Another example, a smart electricity system learns about customer needs, thus predict and align electricity grid power accordingly. Sometimes this is known as IoE (Internet of Energy) and sometimes smart grid technology. It is about connecting grids to the Internet, dispatching energy as needed in

5  Remember our discussion that conventional protocol stack design is not applied in sensor nodes due to lack of memory and processing power issues. various segments dynamically, link various energy generators (most of them are renewable), all being managed over the Internet. The basic idea is to optimize generation, monitoring, and distribution of power flow in an optimal and dynamic way, based on changing needs of the customer. The IoE allows customers to contribute by using renewable energy sources themselves.

Smart City networks need to provide many services normal networks are not designed to provide. Let us take a few examples. When a surveillance camera picks up thieves breaking into a house, they should call the police who can use map service to find out the exact location of the place of burglary and how to reach there in the fastest way. The call, however, should reach to a patrolling van which is geographically nearby, should have GPS enabled and thus the system can find their location. If there are multiple vans, the system should inform the nearest. Let us take another example of pollution level control. The system should be able to detect pollution levels, congestion levels, and direct citizen in proper directions to avoid congested or polluted areas. Home energy management systems will have the ability to inform customers about the power they have consumed so far and pattern of power consumption and additionally provide suggestions to reduce it. As per one estimate, such smart metering system can save up to 30 to 40% of the electricity bill for an average customer. Citywide academic hubs, video lecturing, and counseling, smart classroom devices can improve learning of citizens and provide them more options they can choose from. Some studies which are based on experimentation on “Study from Home”, requires students to go to the institute only twice or thrice. Not only these solutions save student’s traveling time, they also help reducing city’s congestion levels. Similarly, remote health monitoring of the citizens is enabled by IoT or non-IoT IMDs (Implantable Medical Devices). Government hospitals can keep track of discharged patients, newborn babies, and their mothers remotely.

Providing all these services is not a joke. The amount of data being transferred is huge and demand broadband connectivity. While many IoT devices deployed across the city, local processing abilities and convergence of information from other networks like power and gas and water etc. is a big challenge. The humongous data itself is not a single problem, storing that data in a fashion that the decisions can be made in real time is equally critical. Over and above, one needs a sensible networking infrastructure in place, which can store, process and forward relevant information to the point where it can be utilized by the citizen. For example, a detailed analysis of city buses and how crowded they are, the system can help citizen to choose a typical bus station and pick up a typical bus to avoid the crowd and save time. All information about all buses travelling right now is to be sorted, once a citizen decides a destination from a given location, the system should sort out buses that are of interest, finds out how crowded they are and which is the nearest bus station for that citizen and provide that information on citizen’s mobile in a fraction of a second.

Another need for a smart city is ubiquitous computing. Whatever services offered by smart city system, a citizen should be able to avail from any part of the city. The information might require few different types of networks to communicate. For example, whenever there is an accident on the road, the city system should be able to inform the police (under whose jurisdiction the accident has occurred), ambulance (which is nearby and free) as well as the hospital (which is nearby and also capable to handle the patient) if anybody is injured, relatives of the patient are to be automatically informed, if he is unconscious and not in a position to do that himself, specific doctor in case of a specific injury and so on6. This requires seamless communication between many networks.

One of the biggest challenges of deploying smart city solution is a huge amount of data being generated from thousands of sensors deployed, in CCTV cameras, in vehicles, in hospitals, various places in public service centers etc. Not only the data is huge, is continuously being produced and of varieties of types. The devices themselves are not expected to memory and processing power rich and thus solutions for communication must use less power and avoid unnecessary communication.

Current systems have deployed RFID technologies, ZigBee systems which enable the low-cost device to device communication, and even MANets are roped in7. One version of Wi-Fi, popularly known as Wi-Fi HaLow (its technical name is IEEE 802.11ah), specifically designed to use less power8 as well as group computing for both indoor as well as outdoor users, may prove to be a better alternative.

IoT devices of different sizes and shapes needed for different smart city requirements like smart parking system, smart lighting system, patient health monitoring system, systems measuring the state of public buildings and bridges, etc. One needs a cloud computing infrastructure (which in turn demand networking and storage infrastructure) for supporting these devices as they do not have much local storage. All in all, smart city networks with IoT devices is a network to reckon with.

6 Additionally, the system should also generate MIS reports like in which area there is more likelihood of accidents so they can decide the location of stationing ambulances nearby, or number of incidents of typical medical emergencies like snake bites and heart attacks related to a typical area as well as typical atmospheric conditions. For example, heart attacks are more probable during winter. It is easier to predict and stock required medicines beforehand if a good predictive modeling is done. Interestingly, in Gujrat, we have a service known as 108, where a citizen is to dial a number 108 from a mobile or landline and he is given emergency service free from government, most of the things being discussed here is actually provided. The information also includes some parameters which indicate if an operator is lazy in handling calls or not. This can prove to be very critical in improving the effectiveness of the system.

7 It seems Wi-Max like technology can be an idea but Wi-Max consumes a lot of energy not suitable for small devices, so low-cost alternative solutions are sought and used.

8 According to one report, it enables communicating between thousands of devices in the same area with default transmission power 200mW to communicate over 1 Km.

Summary

In this module we have seen few other types of networks, i.e. sensor networks, home networks, IoT based networks and Smart City networks. We have seen the challenges that they pose and some of the solutions used to address those challenges.

References

1. Computer Networks by Bhushan Trivedi, Oxford University Press
2. Data Communication and Networking, Bhushan Trivedi, Oxford University Press
3. Communications of the ACM, special issue on smart cities