It’s a smart world A special report on smart systems November 6th 2010

SmartsystemCOV.indd 1 26/10/2010 12:07 The Economist November 6th 2010 A special report on smart systems 1

It’s a smart world Also in this section A sea of sensors Everything will become a sensor‹and humans may be the best of all. Page 3

Making every drop count Utilities are getting wise to smart meters and grids. Page 4

Living on a platform For cities to become truly smart, everything must be connected. Page 6

Augmented business Smart systems will disrupt lots of industries, and perhaps the entire economy. Page 8

The IT paydirt Who will clean up? Page 9 The real and the digital worlds are converging, bringing much greater eˆciency and lots of new opportunities, says Ludwig Siegele. But is it Your own private matrix what people want? Tracking your life on the web. Page 10 HAT if there were two worlds, the many places at once and is often not un• Wreal one and its digital re‡ection? The derstood for what it is. It is most advanced Sensors and sensibilities real one is strewn with sensors, picking up in controlled environments. For example, A smarter world faces many hurdles. Page 11 everything from movement to smell. The software developed by Siemens, a technol• digital one, an edi†ce built of software, ogy conglomerate, maintains virtual repli• takes in all that information and automati• cas of factories to monitor and recon†gure Horror worlds cally acts on it. If a door opens in the real them. But it is spreading everywhere and Concerns about smart systems are justi†ed world, so does its virtual equivalent. If the has developed a language all of its own. and must be dealt with. Page 13 temperature in the room with the open Glen Allmendinger of Harbor Research, a door falls below a certain level, the digital consultancy, calls it the Œvirtualisation of world automatically turns on the heat. the real world. Researchers at MIT’s Me• This was the vision that David Gelern• dia Lab who connect real•life objects with ter, a professor of computer science at Yale copies in Second Life, a virtual world, refer University, put forward in his book ŒMirror to the result as Œcross reality. Google’s Worlds in the early 1990s. ŒYou will look Earth and Street View services are the †rst, into a computer screen and see reality, he if static, replicas of the entire world; sen• predicted. ŒSome part of your world‹the sors placed in cows allow the tracking of town you live in, the company you work their every move from birth to abattoir; for, your school system, the city hospital‹ smart power meters tell utilities in real will hang there in a sharp colour image, ab• time how much electricity is used. stract but recognisable, moving subtly in a Acknowledgments Besides those cited in the text, the author would like to thousand places. The many uses of smartphones thank many others for their time, ideas and kindness Even two decades later that sounds like Yet it is the smartphone and its Œapps during the preparation of this report. They include Bruno science †ction. But this special report will (small downloadable applications that Berthon of Accenture, Eric Clementi of IBM, Tom Davenport of Babson College, Peter Elliott of PA argue that Mr Gelernter was surprisingly run on these devices) that is speeding up Consulting Group, Scott Gnau of Teradata, Jeanne Harris prescient: mankind is indeed building the convergence of the physical and the of Accenture, Chris King of eMeter, Michael Liard of ABI more and more Œmirror worlds, or Œsmart digital worlds. Smartphones are packed Research, Michael LoCascio of Lux Research, Edzard Overbeek of Cisco, Joachim Schaper of SAP and Elisabeth systems, as they are often called. The real with sensors, measuring everything from Zornes of Cisco. and the digital worlds are converging, the user’s location to the ambient light. thanks to a proliferation of connected sen• Much of that information is then pumped A list of sources is at sors and cameras, ubiquitous wireless net• back into the network. Apps, for their part, Economist.com/specialreports works, communications standards and the are miniature versions of smart systems activities of humans themselves. that allow users to do a great variety of An audio interview with the author is at This convergence may not be instantly things, from tracking their friends to con• Economist.com/audiovideo/specialreports discernible, because it is happening in trolling appliances in their homes. 1 2 A special report on smart systems The Economist November 6th 2010

2 Smartphones are also where the virtual platforms for sensor data. and the real meet most directly and merge Governments, too, have jumped on the into something with yet another fancy bandwagon. Many countries have been name: Œaugmented reality. Download an spending large chunks of their stimulus app called ŒLayar onto your smartphone, packages on smart•infrastructure projects, turn on its video camera, point at a street, and some have made smart systems a pri• and the software will overlay the picture ority of industrial policy. The Œinternet of on the screen with all kinds of digital infor• things, another label for these systems, is mation, such as the names of the business• central to the European Union’s ŒDigital es on the street or if a house is for sale. Agenda. The main contenders in this mar• These and other services are bound to ket are countries that are strong in manu• grow together into what Jan Rabaey, a com• facturing, above all Germany and China. puter scientist at the University of Califor• But the bandwagon is not just rolling for nia at Berkeley, grandly calls Œsocietal in• the bene†t of technology companies and formation•technology systems, or SIS. ambitious politicians. It has gained mo• Technological progress is sure to supply mentum because there is a real need for the necessary components. Moore’s law, such systems. In many countries the physi• which holds that the processing power of a cal infrastructure is ageing, health•care alone were just 5% more eˆcient, it would single computer chip roughly doubles ev• costs are exploding and money is tight. Us• save greenhouse emissions equivalent to ery 18 months, applies to sensors, too. ing resources more intelligently can make 53m cars, calculates IBM. In 2007 its con• More processing power and better con• taxpayers’ money go further. Monitoring gested roads cost the country 4.2 billion nectivity also allow the construction of patients remotely can be much cheaper working hours and 10.6 billion litres of computing systems capable of storing and and safer than keeping them in hospital. A wasted petrol, according to the Texas Tran• crunching the huge amounts of data that bridge equipped with the right sensors can sportation Institute. And utilities around will be produced by these sensors and oth• tell engineers when it needs to be serviced. the world lose between 25% and 50% of er devices. All over the world companies China is a good example. It is becoming treated water to leaks, according to Lux Re• are putting together networks of data cen• urbanised on a scale unprecedented in his• search, a market•research †rm. tres packed with thousands of servers, tory. By 2025 an additional 350m Chinese‹ With so much to gain, what is there to known as Œcomputing clouds. These not more than the current population of the lose? Privacy and the risk of abuse by a ma• only store data but sift through them, for United States‹will have moved to cities, levolent government spring to mind †rst. instance to allow a smart system to react according to a study by McKinsey, a consul• Indeed, compared with some smart sys• instantly to changes in its environment. tancy. Without an infrastructure enhanced tems, the ubiquitous telescreen monitor• by digital technology it will be very hard to ing device in George Orwell’s novel Œ1984 The next big thing provide the country’s newly urbanised seems a plaything. The book’s hero, Win• Information•technology (IT) †rms have population with enough food, transport, ston Smith, would soon have a much hard• identi†ed smart systems as the next big electricity and water. er time †nding a corner in his room to hide thing. Predictably, the most ambitious de• Most important, smart systems may from Big Brother. signs have been produced by the indus• well be humankind’s best hope for dealing Second, critics fear that smart systems try’s giants, particularly IBM, where Sam with its pressing environmental problems, could gang up on their creators, in the way Palmisano, the †rm’s boss, made smart notably global warming. Today power they did in ŒThe Matrix, a 1999 †lm in systems a priority. A couple of years ago grids, transport systems and water•distri• which human beings are plugged into ma• the company launched a campaign called bution systems are essentially networks of chines that simulate reality to control hu• ŒSmarter Planet, touting digital technol• dumb pipes. If the power grid in America mans and harvest their bodies’ heat and ogy that would make energy, transport, cit• electrical activity. Fortunately, such a sce• ies and many other areas more intelligent. nario is likely to remain science †ction. But Other †rms have followed suit, each with The way to go 1 smart systems might be vulnerable to mal• a di erent take re‡ecting its particular Total worldwide internet-enabled-device revenues functioning or attacks by hackers. strengths. $bn Third, some people fret that those with Cisco, the world’s biggest maker of net• Industrial Energy Buildings access to smart systems will be vastly bet• Transport Consumer/ Other working gear, is trumpeting ŒSmart+Con• professional ter informed than those without, giving nected Communities. Hewlett•Packard, 12 them an unfair advantage. Mr Gelernter FORECAST highlighted this risk in ŒMirror Worlds. number one in hardware, intends to spin a 10 ŒCentral Nervous System for the Earth. There are plenty of other concerns, and Siemens and its competitor General Elec• 8 unless they are dealt with they could pro• tric, which are more at home in the physi• voke a neo•Luddite reaction. The world 6 cal world, plan to put together lots of smart has already seen one extreme example: the systems in which they can deploy their 4 Unabomber, a disa ected American who deep knowledge of certain industries, targeted, among others, computer scien• 2 such as health care and manufacturing. tists with mail bombs. Two years after the And there is a growing wave of Œsmart 0 publication of ŒMirror Worlds he sent one start•ups, o ering everything from ser• 2009 10 11 12 13 14 to Mr Gelernter, who was seriously in• Source: Harbor Research vices to pinpoint a devices’s location to jured‹though fortunately he survived. 7 The Economist November 6th 2010 A special report on smart systems 3

A sea of sensors

Everything will become a sensor‹and humans may be the best of all

ERMANS are known to separate their are designed to pick up the echo of the seis• Grubbish diligently. Some even have High note 2 mic vibrations created by contraptions dedicated containers for di erent kinds of Tweets per day, m called Œthumper trucks pounding the metal. But they may soon need yet another ground. The data thus gathered allow bin: for electronic labels, also known as ra• 100 them accurately to pinpoint pockets of oil dio•frequency identi†cation (RFID) tags. If and gas. not kept and treated separately, these could 80 Yet RFID tags, wireless sensors and, for be very diˆcult to recycle, Germany’s Fed• 60 that matter, digital cameras (so far the most eral Environment Agency said last year. widely deployed sensor thanks to the pop• The number being thrown away each year 40 ularity of mobile phones) are only half the could rise from about 86m now to 23 bil• story. Many objects no longer need an elec• lion by 2020, according to the agency. 20 tronic tag or even a barcode to be automati• RFID tags, which have been used to cally identi†ed. For example, Goggles, a identify everything from cattle to tomb• 0 service o ered by Google, can recognise stones, will not be the only type of sensor Jan Jul Jan Jul Jan Jul Jan Jul Jan Jul things like book covers, landmarks and crowding the planet. Anything and any• 2006 07 08 09 10 paintings. Users simply take a picture and Source: Twitter one‹machines, devices, everyday things send it to Google’s computers, which will and particularly humans‹can become a send back search results for the object. sensor, gathering and transmitting infor• Having to run wires or regularly replace Many of the innumerable machines mation about the real world. batteries would be too diˆcult. But sen• and devices that populate the physical en• The concept of the Œinternet of things sors have started to power themselves by vironment also already come with some dates back to the late 1980s, when research• scavenging for energy in their environ• data•generating digital technology. More ers at Palo Alto Research Centre (PARC) in ment, for instance in the form of light and and more are getting connected so that Silicon Valley imagined a future in which motion. Similarly, some sensors already they can communicate the information the virtual and the real world would be make more eˆcient use of another scarce they contain to the outside world. Exam• connected. In the following years much of resource: radio spectrum. Smart power ples range from co ee machines and refrig• the academic work concentrated on bring• meters form Œmesh networks to relay erators to aircraft engines and body scan• ing this about with RFID tags, which are re• their readings. ners in hospitals. These can all now phone liable, inexpensive and do not require a Engineers working on sensors think home, so to speak, and provide their mak• power supply. When exposed to a radio this will eventually lead to Œsmart dust‹ ers with fountains of data. signal, they use its energy to send back the sensors as small as dust particles that can information they contain, mostly a long be dispersed on a battle†eld, say, to spy on People power number identifying an object. the enemy’s movements. Such devices are Most important, however, humans them• Now it is Œactive tags (which have their still far o , but at Hewlett•Packard (HP) in selves have turned out to be excellent sen• own power source) and, even more, wire• Silicon Valley a taste‹or more precisely, a sors. Many provide information without less sensors that are attracting most of the feel‹of things to come is on o er even any extra e ort, just by carrying around a interest. As with all things electronic, these now. To demonstrate the †rm’s new mobile phone. TomTom, a maker of navi• are becoming ever smaller and more ver• accelerometer, a device the size of a ciga• gation devices, uses connection data from satile. ŒTell me what you need, and we can rette box that measures the acceleration of mobile networks to update directions if build it for you, says Reinhold Achatz, an object, Peter Hartwell, a researcher, puts there are delays. Others are connecting ad• head of corporate research at Siemens. it on his chest, and a graph of his heartbeat ditional sensors to smartphones. Such de• Start•ups, too, are producing devices that appears on a screen beside him. ŒThis sen• vices and smartphones are gradually turn• sense everything from the rarest chemical sor, he proudly explains, Œis one thou• ing people into the sensory organs of the to the most exotic bacteria. Optiqua, a Sin• sand times more sensitive than those in internet, say John Battelle, boss of Federat• gaporean †rm, has come up with a chip your smartphone. ed Media, an online advertising agency, that measures how fast light passes One day, Mr Hartwell and his col• and Tim O’Reilly, who heads O’Reilly Me• through a water sample to detect contami• leagues hope, a network of one trillion dia, a publishing house. ŒOur cameras, our nants. And a biosensor developed at the sensors will cover the world and deliver microphones, are becoming the eyes and Lawrence Livermore National Laboratory data to anybody who needs them, from ears of the web, they write in a paper enti• in Berkeley, California, can identify about carmakers to municipal governments. For tled ŒWeb Squared. 2,000 viruses and 900 bacteria. now, the †rm has teamed up with an oil More surprising than such Œcrowdsens• Researchers are also on the way to solv• company, Royal Dutch Shell. The comput• ing, as it has come to be known, is the will• ing two big problems that have held back er•maker plans to dot a prospecting area ingness of many people actively to gather the deployment of sensors. One is power. with thousands of wireless sensors. They and upload information. The best example1 4 A special report on smart systems The Economist November 6th 2010

2 is Twitter, the micro•blogging service sponsored by EMC, a maker of storage sys• Analytics software is improving, too. It whose 160m users send out nearly 100m tems. But for what they are worth, they has long been used to crunch through data tweets a day (see chart 2, previous page). suggest that the Œdigital universe‹the that are Œstructured, or organised in a da• When they see, hear or read something, amount of digital information created and tabase, and develop models to predict, for they type it into their computer or smart• replicated in a year‹will increase to 35 zet• example, whether a credit•card transac• phone, 140 characters at a time. ŒTwitter is tabytes, or 35 trillion gigabytes‹enough to tion is fraudulent or what the demand for the canary in the news coalmine, wrote †ll a stack of DVDs reaching halfway to ‡ights will be around a public holiday. Je Jarvis, a new•media savant, after the Mars. Even that may prove a conservative Now such programs can also interpret Œun• service beat mainstream media to news estimate if sensors and other data•generat• structured data, mainly free•form text. about the earthquake that struck China’s ing devices spread as predicted. Earlier this year SAS launched a product Sichuan province in May 2008. Fortunately, the tools to deal with this capable of analysing the Œsentiment of But there are plenty of other examples. data deluge are getting better. Give a mark• chatter on social media, including Face• At OpenStreetMap, a wiki•style website, er and a whiteboard to Bijan Davari, a re• book and Twitter. some 250,000 volunteers record their searcher at IBM, and he will draw you a The software is also able to †nd the peo• wanderings, using their smartphones’ po• picture of the future of computing as he ple who post the most in‡uential com• sitioning functions. And SeeClickFix, a and his employer see it. On the left side ments on speci†c companies on Twitter, start•up, has come up with a smartphone there are small squares, representing all who can then be sent special marketing app that allows users to report such things kinds of sensors. The data they produce messages. Indeed, Twitter itself is a kind of as broken streetlights or rubbish that needs are fed into something he calls the collective †lter that continuously sorts the to be picked up. Œthroughput engine, a rectangle on the content published on the web. And Face• right. This is a collection of specialised book users, by tagging friends in the pic• Too much of a good thing chips, each tailor•made to analyse data tures they upload, allow the service to re• It does not take much imagination to see from a certain type of sensor. ŒA system cognise these people on other pictures. that all these sensors will generate im• that can’t deal with these streams separate• ŒMeaning is ‘taught’ to the computer, mense amounts of data. ŒThey don’t make ly would quickly become overloaded, write Messrs Battelle and O’Reilly. enough disk space in the world to save all says Mr Davari. But the main goal of smart systems is the data if every household had a smart IBM has already introduced a product Œto close the loop, in the words of a report meter, says Jim Goodnight, the boss of based on what it calls Œstream computing on the internet of things published in SAS, one of the pioneers of analytics soft• that can ingest thousands of Œdata March by the McKinsey Global Institute. ware, programs that sift through data. ŒIn streams and analyse them almost instan• This means using the knowledge gleaned fact the most diˆcult question is to decide taneously. The natal care unit at the Uni• from data to optimise and automate all what to discard. versity of Ontario is testing such a system kinds of processes. The number of poten• The quantity of data likely to be pro• to monitor babies born prematurely. It tial applications is vast, ranging from duced is anybody’s guess. Estimates by takes in streams of biomedical data, such manufacturing to heading o car colli• IDC, a market•research †rm, need to be tak• as heart rate and respiration, and alerts sions. Yet the most promising †eld for now en with a pinch of salt, because they are doctors if the babies’ condition worsens. may be physical infrastructures. 7 Making every drop count

Utilities are getting wise to smart meters and grids

ONDON’S streets can be a bit of a maze, few years ago more than half of the 10,000 240 leaks had to be †xed. Lbut below ground things are even more miles (16,000km) of water pipes below the Over the past †ve years, though, complex. Water pipes crisscross the city in streets of London were over a hundred Thames Water has replaced 1,300 miles of all directions. Some areas used to have years old and often burst. It did not help cast•iron Victorian mains, those most like• competing water companies, each of that over many years Thames Water, ly to break, with plastic ones, reducing which built its own system. Not even which was privatised in 1989, failed to in• leakage to 670m litres per day. And when Thames Water, the utility that operates the vest enough. By the mid•2000s London the †rm puts in new pipes, it also installs British capital’s water•supply network to• had one of the leakiest water•supply sys• additional wireless sensors, giving it a bet• day, knows exactly where all the pipes run. tems in the rich world. Every day nearly ter view of its network. ŒWe can now tell Moreover, the network is ageing. Only a 900m litres of treated water were lost and where we have a broken main even before 1 The Economist November 6th 2010 A special report on smart systems 5

2 customers call us, says Bob Collington, its provide a basis for comparison, enabling did not skimp. It deployed equipment that head of asset management. its algorithms to detect things that are automatically reports power cuts. It in• Thames Water not only needs to know about to go wrong. stalled more than 20,000 smart meters, what is going on in its network, but to be Similar progress is being made all over connected them via a †bre•optic network, able to act quickly on the information. The the world. The scope for preventing waste launched a website to track power use and same is true of infrastructure operators is enormous, in the water industry and has started to o er pricing plans that en• around the world. Whether in water, pow• elsewhere. Power utilities are well ahead, courage shifting consumption to o •peak er, transport or buildings, all are trying to not least because they can use the grid it• hours. It has even equipped some house• turn their dumb infrastructures into some• self to collect sensor data and control holds with gear that tells air•conditioning thing more like a central nervous system. switches. Transport systems are behind, systems to turn themselves o when de• That makes them pioneers of the conver• particularly roads, which often use noth• mand for electricity is high, a mechanism gence of the physical and the digital world. ing more than traˆc cameras. Even build• called Œdemand response. Putting sensors and actuators (devices ings are getting more automated, with con• The results so far are mixed. The system to control a mechanism) into physical in• tinuous checks on their energy use. has certainly helped Xcel to run its grid frastructures is not exactly new. Known as more eˆciently. The utility now knows Œsupervisory control and data acquisi• At the edge what is happening in its network and pow• tion, such systems have been around for For infrastructures to become truly smart, er cuts have become rare. Problems can be decades. But many still require human in• however, it is not enough to put more intel• pinpointed and †xed much more quickly. tervention: workers have to be sent out to ligence into the core of a network. The But customers are not using much less download sensor readings or to †x pro• edge‹the interface with users and de• power than they did before. blems. And even if sensors and actuators vices‹also has to become clever. This is the Yet it is early days. Some †rms are al• are connected, di erent types often feed idea behind smart metering, which has ready beginning to show what can be into incompatible systems, so they cannot made a good deal of progress in the power achieved with demand response. Ener• be easily combined to automate processes. industry. According to Accenture, a consul• NOC, an American energy middleman, for The operations centre of Thames Water tancy, there are about 90 smart•grid pro• instance, pays other †rms for allowing it to in Reading, to the west of London, is a good jects around the world today. By the end of shut down their non•essential gear at place to see both the old and the new‹and last year more than 76m smart meters had times of peak demand, thus freeing up ca• soon the future. A big video screen shows been installed worldwide. That number pacity. By mid•year some 3,300 customers, expected precipitation over the next few will almost treble by 2015, to 212m, esti• from steel plants to grocery stores, had hours, and workers monitor the water lev• mates ABI Research. signed up. Their combined consumption, el of reservoirs on their own screens. But if Smart meters and other gear needed to which can be made available to other us• one of the pumps fails, they may still have make grids more intelligent will not come ers if needed, is 4,800MW, exceeding the to make a call: not all the valves can be re• cheap. Morgan Stanley, an investment output of America’s largest nuclear plant. motely controlled. bank, predicts that the worldwide smart• The ultimate point of smart grids, how• Thames Water is investing £100m grid market alone will grow from $20 bil• ever, is to allow dynamic pricing, with ($158m) so it can take action remotely and lion last year to $100 billion in 2030. Yet the electricity charges ‡uctuating in response automate a lot of its processes. If the pro• bene†ts also promise to be huge: power to demand. This could cut power demand ject works, the system will not only auto• savings, reduced investment in electricity by 10•15% during peak hours, estimates Ah• matically deal with leaks but also schedule generation and lower carbon emissions. mad Faruqui of the Brattle Group, a consul• work crews and send text messages to af• The place to go to see the technology in tancy‹more than twice the reduction like• fected customers. Employees in the opera• action is Boulder, Colorado, home to what ly to be achieved by just giving customers tions centre, explains Jerry White, the utili• is considered the world’s †rst fully ‡edged real•time information about their usage. ty’s head of operational control, will then Œsmart grid. The local utility, Xcel Energy, That number could easily double again, he spend less time monitoring the network says, with a combination of dynamic pric• and more on making the utility’s processes ing and demand response. more eˆcient. Splashing 3 The main objective of smart power me• A big chunk of this work will be analys• Daily domestic water consumption per person ters is to lower the peak load and thus en• ing the data collected by all the systems 2000, litres able utilities to keep down their peak gen• and correlating them with other informa• erating capacity. In the water industry the 0 200 400 600 800 tion. Not every unexpected spike in the economics are somewhat di erent, ex• water ‡ow is the result of a leak, says Mr Canada plains Stefan Helmcke, a water expert at White. For instance, water use leaps after United States McKinsey. Water can be easily stored and dark during Ramadan and at half•time South Korea consumers have less discretion over when during World Cup football matches. Japan they use it (for instance, people cannot de• One day soon Thames Water may even fer going to the toilet, which uses more wa• Brazil be able to send out work crews before a ter than any other activity at home), so the main actually breaks. In early 2010 the †rm Germany case for smart water meters is weaker. began using a web•based service provided India Yet they are spreading all the same. Bos• by TaKaDu, an Israeli company, that acts as Britain ton has long been the shining example. As the network’s Œeyes and ears, in the words China early as 2004 the city’s Water and Sewer of Amir Peleg, its founder and boss. The Commission had equipped almost all its Sources: WRI; The Economist †rm analyses historical and online data to customers with wireless smart meters. But 1 6 A special report on smart systems The Economist November 6th 2010

2 it will soon be outdone by New York, city to introduce real•time dynamic pricing also among the world’s most eˆcient. All which plans to install more than 800,000 on its roads. In 2006 the Land Transport this means that the island‹smaller than of the devices at a cost of about $250m. Authority tested a traˆc•prediction sys• Luxembourg and home to nearly 5m peo• Even Thames Water, most of whose cus• tem built by IBM to set the tolls. And next ple as well as an economy nearly as big as tomers have no meters of any sort, is now year it plans to test a satellite•based system that of Hong Kong‹is able to meet more planning to install some of the smart kind. that does not require gantries and can than 60% of its water needs on its own. But charge according to how congested a road it wants to go even further: 50 years from Getting on board is at that particular time. now it hopes to be self•suˆcient. In transport the equivalent of a smart me• Another of the island’s infrastructure• Sensors play a relatively small part in ter is a vehicle’s on•board unit. That used to management systems has become a mod• Singapore’s water management because be a simple device, working like a radio• el for the world: that for water. At the infor• the infrastructure is so new. On average frequency identi†cation tag when it passes mation centre at the southern tip of the is• there is only one leak a day. The PUB puts under a gantry on a toll road, but it is also land, next to the Marina Barrage, visitors sensors only in a few key spots, for in• getting smarter. Germany’s Toll Collect can literally get a taste of it by picking up a stance where water leaves the reservoirs. system, which ensures that lorry drivers bottle of ŒNEWater, waste water that after Should the system detect a dangerous con• pay for using the country’s crowded mo• extensive treatment has become potable tamination, that part of the network can be torways, relies on gadgets that are in some again. But most of the treated water is fed shut down immediately. And if heavy ways as clever as a smartphone. Among back, via a separate distribution system, to rainfall in central Singapore threatens to other things, they keep track of their posi• Singapore’s factories and power plants‹ ‡ood the city during high tide, seven huge tion with the help of GPS, the satellite• and then treated again. pumps next to the Marina Barrage start to based global positioning system. This closed loop is part of a water•sup• push water into the sea at 40 cubic metres Such toll systems are multiplying, par• ply system in which Œevery drop counts, per second each. ticularly in big congested cities, including in the words of Yap Kheng Guan, a director So far Singapore has no smart water London and Stockholm. But it is Singapore at the island’s Public Utilities Board (PUB). meters, and at the moment there is no that leads the pack. The city•state not only The Marina Barrage is another case in pressing need. Most Singaporeans live in charges drivers for using much•travelled point. It was inaugurated in 2008 and acts multi•storey apartment buildings, which roads (driving on an expressway can be as a tidal barrier to keep seawater out, thus makes it easy to read meters. But if the PUB S$6, or $4.60); it also adjusts traˆc lights to turning the island’s most populated dis• wants to reach its target of cutting daily do• suit the ‡ow of vehicles, uses data collect• trict into a water•catchment area and the mestic water use per person from 155 litres ed by taxis to measure average speed and is harbour into a reservoir. When two other in 2008 to 147 litres by 2020 (about the developing a parking•guidance system, reservoirs are opened next year, more than same as in India, and a quarter of the †gure noting that cars looking for somewhere to two•thirds of Singapore’s territory will be in America, see chart 3 on previous page), park are now a big cause of congestion. used to catch rainwater. Singapore will have to become smarter Singapore may also become the †rst The city•state’s desalination plants are still‹and set yet another example. 7 Living on a platform

For cities to become truly smart, everything must be connected

N SINGAPORE conversations about wa• parts of the main shopping district were foreign †rms in these sectors will be able to Iter quickly turn political. The city•state under water after heavy rainfalls. develop and show o their products on no longer wants to depend on water from The problems of scarcity and excess are the island before selling them elsewhere, Malaysia when the current water•supply in evidence on the city•state’s roads too. explains Goh Chee Kiong, who is in charge agreement between the two countries ex• Singaporeans, who are among the world’s of the clean•energy cluster at Singapore’s pires in 2061. More than once the neigh• richest people, love to drive, but space for Economic Development Board. bour to the north, of which Singapore was roads is severely limited. When in the early There is strong demand for making cit• part before an acrimonious split in 1965, 1970s the central area became too congest• ies smarter, not just in China and other rap• has threatened to increase prices or even ed, the government introduced the world’s idly urbanising countries but throughout cut o supplies. †rst manual urban road•pricing system. In the Western world. Resources like water, Yet politics is not the only reason for 1998 it became the †rst to be automated. space, energy and clean air are scarce in ur• Singapore’s advanced water system. The ŒSingapore proves that necessity is the ban areas, which makes them the natural information centre at the Marina Barrage mother of invention, says Teo Lay Lim, place to start saving, says Mark Spelman, features pictures of ‡oods and droughts. who heads the local oˆce of Accenture. Accenture’s global head of strategy. ŒWe have either too much water or too lit• Now the city wants to become a Œliving ŒSmart•city projects have been multi• tle, explains Yap Kheng Guan, a director laboratory for smart urban technologies plying around the world. Some of them of Singapore’s PUB. Even today, despite a of all kinds‹not just water and transport are not as new as their labels suggest, and sophisticated system of ditches and tun• systems but green buildings, clean energy in any case what exactly constitutes a nels, ‡oods can suddenly strike. In July and city management too. Both local and smart city is hard to de†ne. But they all 1 The Economist November 6th 2010 A special report on smart systems 7

2 have one thing in common: they aim to in• struction companies, Mr Lewis ended up tegrate the recent e orts to introduce smart applying his ideas to an entire city. Even be• features in a variety of sectors and use this fore the †rst concrete is cast, PlanIT Valley Œsystem of systems, as IBM calls it, to has already been built‹in a simulation manage the urban environment better. program that also allows detailed plan• The best•known smart city is Masdar, a ning of the construction. Much of the city, brand•new development in Abu Dhabi which is to cost about $10 billion, will rely that recently welcomed its †rst residents on prefabricated parts; its foundation, for and will eventually become home to instance, will be made of concrete blocks 40,000 people. It is being built entirely on that come with all the gear for smart infra• a raised platform, which makes mainte• structures pre•installed. Eventually the en• nance and the installation of new gear tire city and its buildings will be run by an much easier. Below the platform sits the Œurban operating system that integrates smart infrastructure, including water pipes all parts and combines them into all kinds with sensors and a †bre•optic network. of services, such as traˆc management Above it is to be a showcase for all kinds of and better use of energy. green technology: energy•eˆcient build• Living PlanIT has a clear idea of who ings, small pods that will zoom around on will live and work in its city: the employ• paths (no cars will be allowed) and sys• ees of the companies that form its Œecosys• tems that catch dew as well as rainwater. tem, another concept taken from the soft• Yet experts see Masdar mostly as a from security cameras. Visitors were given ware industry. The start•up has enrolled a property project: hardware in search of a a demonstration of how city managers number of partners, among them Cisco, purpose. What really matters to a city’s would react to an accident on a city•centre Accenture and McLaren Electronic Sys• smartness, they argue, is the software that bridge: cameras zoom in, an ambulance is tems, a sister company of the eponymous runs on it and the network that connects its dispatched, traˆc is rerouted to other Formula One brand, which will provide parts. ŒIt is the common infrastructure for bridges‹all automatically, within seconds. sensor technologies. The idea is that these all the smart systems, says Wim Elfrink, The world’s smartest city, however, †rms will operate research facilities in who heads the ŒSmart+Connected Com• may soon rise in an unexpected place: near PlanIT Valley, jointly improve the concept, munities initiative of Cisco, the network• Porto in Portugal. PlanIT Valley, designed develop applications and build similar ing•equipment maker. for an eventual population of 150,000, is projects elsewhere. an ambitious attempt to Œcombine tech• Such grand designs are possible only Getting it together nology and urban development, in the when building a city from scratch. But en• Cisco is trying to demonstrate that point in words of Steve Lewis, co•founder of Living trepreneurs like Mr Lewis and the planners Songdo City near Seoul, where the †rm PlanIT, the start•up behind the project, of similar projects have other advantages. provides all the digital plumbing. Sitting who used to work at Microsoft. For a start they generally have government on reclaimed land, Songdo is perhaps the His experience at the software giant backing. Portugal granted the PlanIT Valley most ambitious smart•city project so far. It proved an excellent preparation for the project Œpotential national importance is expected to cost $35 billion and will be job. Microsoft is the very model of a plat• status, which among other things means home to 65,000 people. Like Masdar, it will form company, providing technology to cheap land and generous tax breaks. boast all the latest green technologies. But connect things (such as printers and PCs) Songdo was launched by the South Korean its main claim to fame is that everything in and a foundation for the products made by government. And new cities are free from the city is wired up. others (such as browsers and media play• the constraints of having to deal with an Residents of ŒFirst World, the †rst com• ers). When he was at Microsoft in the early established population, old infrastructure pleted apartment complex, already enjoy 2000s, Mr Lewis also oversaw the re• and bureaucracy. the bene†ts of this all•embracing connec• launch of a strategy called .Net‹an early tivity. Smartphones unlock front doors. example of what geeks like to call Œservice• Something old, something new Air•conditioning, blinds and security sys• oriented architecture. The idea is to build Making an existing city smart is a di erent tems are controlled by displays all over the programs as a combination of loosely•cou• problem altogether, as demonstrated by apartment which can also be used to ac• pled electronic services that can be rede• Amsterdam, the Netherlands’ biggest city. cess all kinds of online services. With a few ployed elsewhere. Amsterdam Innovation Motor (AIM), a clicks or touches users are able set up a After Mr Lewis left Microsoft in 2005, public•private joint venture created for this videoconference with a doctor, do busi• he tried to introduce the concept of re•us• purpose, is not intended to come up with ness with the local government or †nd out able components to the construction in• some master plan but to identify interest• how best get to work. dustry, which seemed ripe for it. Designs ing Œsmart projects, work with local †rms What Cisco sees as the most important are often used only once, most buildings and other stakeholders and †nd ways to application for running a smart city was are not energy•eˆcient, the industry pro• make projects worthwhile for everybody. shown at the World Expo in Shanghai this duces a lot of waste, and many materials So far AIM has launched a dozen pro• year. In its pavilion the †rm built a com• are simply thrown away. All this amounts jects, ranging from installing smart meters mand centre to keep tabs on an imaginary to around 30% of the cost of construction, in some households to connecting ships to smart city. A huge video screen displayed according to a case study on Living PlanIT the electricity grid so that they no longer everything from traˆc maps and energy by the Harvard Business School. have to use diesel generators when use to weather information and pictures But instead of selling products to con• berthed in the city’s port. The most ambi•1 8 A special report on smart systems The Economist November 6th 2010

2 tious e ort so far is something called ŒCli• cently they introduced the ŒCopenhagen sport and security. mate Street, which aims to reduce the en• Wheel, a bicycle wheel whose red hub So far Mr Ratti and his collaborators ergy use of an entire shopping street. can not only give the rider a boost but also have mainly used data from mobile Most existing cities, at least in the West, measure environmental conditions, such phones for their projects. In ŒWikiCity, will go for such a step•by•step approach, as pollution and noise levels. implemented in Rome, such data allowed predicts Carlo Ratti, an architect and engi• Yet many such projects will need a com• people to see visualisations of how they neer who heads the SENSEable City Lab at mon platform that streamlines data gath• moved through the Italian capital. How• the Massachusetts Institute of Technology ering and supports all kinds of applica• ever, a new mayor elected in 2008 proved (MIT). He and his colleagues have come up tions, says Mr Ratti. That would also much less interested in Mr Ratti’s projects with a number of smart urban projects of enable him to realise his ultimate vision: than his environmentalist predecessor, so their own. In one, called ŒTrashTrack, they turning the city into a Œcontrol system that the team has gone o to Singapore. asked volunteers to attach small electronic makes use of data from a variety of But it will be not just governments, cit• tracking devices to hundreds of pieces of sources, from mobile phones to smart me• ies and utilities that will make the world rubbish to see where they would end up in ters and sensors in buildings. The data smart. Private companies will also play order to improve waste logistics. More re• could be mined to improve public tran• their part, particularly start•ups. 7 Augmented business

Smart systems will disrupt lots of industries, and perhaps the entire economy

ALL it the democratisation of sensors. achieved a 5% increase in its production by aeroplanes equipped with special cam• C Pachube (pronounced Œpatch•bay), a automatically adjusting the shape and in• eras to map the country’s highways. Its em• start•up based in London, o ers a service tensity of the ‡ames that heat the kilns for ployees can now ‡y over them digitally that lets anybody make sensor data avail• the lime used to coat paper. FoodLogiQ, a and even see what is underground. San able to anyone else so they can use them to start•up, allows food suppliers to tag and Francisco’s Public Utilities Commission build smart services. One tinkerer has Pa• trace their wares all along the supply knows the exact co•ordinates of every chube’s computers control the fan in his of• chain‹and consumers to check where waste•water pump, its maintenance his• †ce, guided by temperature readings they come from. Sparked, another start•up, tory and the likelihood of it failing. ŒFirms uploaded from a thermometer on his desk. implants sensors in the ears of cattle, can now send out maintainance crews be• Such experiments are free, but those which lets farmers monitor their health, fore things actually break, says Steve who develop more serious applications track their movements and †nd out a lot of Mills, who heads IBM’s software business. and do not want them to be available to other things about their animals that they ŒMaking the old stu run better will be the anyone else have to pay. Usman Haque, Pa• never realised they wanted to know. On most important bene†t of such systems in chube’s boss, hopes that more and more average, each cow generates about 200 the short run. †rms will do so as sensors multiply. megabytes of information a year. Moreover, smart systems make new Pachube’s business model is one of the Thanks to detailed digital maps, main• forms of outsourcing possible, some of it more interesting attempts to make money taining such things as roads and equip• to unexpected places. Paci†c Control is not from the convergence of the physical and ment will also become much more eˆ• exactly a household name, but the com• the digital worlds, but there are plenty of cient. As†nag, an Austrian †rm, used pany, based in Dubai, claims (with some other †rms trying to cash in on smart sys• justice) that it is the Œworld leader in auto• tems. Many will fail, but those that succeed mation solutions. Its global command will disrupt more than one industry and centre remotely monitors buildings, air• perhaps the economy as a whole. ports and hotels, keeping an eye on such But what is most exciting about smart things as energy use, security and equip• systems is the plethora of new services ment. For the moment most of the †rm’s and business models that they will make customers are in Dubai itself, but it should possible. ŒThe internet of things will allow †nd more than a few abroad. for an explosion in the diversity of busi• Signi†cantly, once devices are connect• ness models, says Roger Roberts, a princi• ed and their use can be metered, there is no pal at McKinsey and one of the authors of longer any need to buy them. Already a recent study on the industry. some makers of expensive and complex It is not just utilities that will bene†t equipment no longer sell their wares but from smart systems but other sectors too. charge for their use. Rolls•Royce, for in• The chemical industry, for instance, has al• stance, which makes pricey aircraft en• ready installed legions of sensors and actu• gines, rents them out to airlines, billing ators to increase its eˆciency. Others are them for the time that they run. Makers of just starting. In the paper industry, accord• blood•testing equipment have taken to ing to the McKinsey study, one company charging only if the device actually pro•1 The Economist November 6th 2010 A special report on smart systems 9

The IT paydirt Who will clean up?

ONTRARY to what might be expect• were dethroned by smaller computer sys• C ed, the industry that will see the least Into the unknown 4 tems that allowed companies to stream• change is information technology. For US government and business spending line many more of their business process• most IT †rms smart systems simply mean on IT equipment and software, $bn es. Similarly, experts predict that new more business. This will drive a new wave Current-generation Smart-computing programs will be needed to manage a technology process applications of technology investment, predicts An• Smart-computing Smart-computing †rm’s interactions with the physical drew Bartels of Forrester Research. By platform technology industry solutions world. Already, there is a plethora of start• 2017, he says, Œsmart computing technol• 800 ups o ering to help businesses cut green• ogies will represent about half the FORECAST house gases. spending on IT equipment and software 600 Platforms to integrate the data streams in America (see chart 4). from all kinds of sensors are another new Clearly some parts of the IT industry 400 market. Examples include Pachube, a stand to bene†t more than others. Harbor start•up, and Palantiri Systems, best de• Research estimates that internet•enabled 200 scribed as a ŒFacebook for sensors. The devices alone will net more than $10 bil• service allows devices to have their own lion worldwide in 2014, compared with 0 Œpage on corporate social networks, so $4.3 billion in 2009. And wireless sensors 2008*09* 10 11 12 13 14 15 16 17 their readings are shown in the form of Source: Forrester Research *Estimate are growing exponentially. Last year a newsfeeds and people can ask questions mere 10m radio chips for such sensors about them. ŒDevices thus become part of were sold, according to ABI Research. If it in buying †rms that make business ana• the conversation, says John Canosa, Pa• has its calculations right, that total will lytics software. In September, for instance, lantiri’s boss. rise to 645m by 2015. it gave $1.7 billion for Netezza, a vendor of As usual, most money will be made Since all the data gathered by sensors data warehouses (specialised computer from IT services, in particular to set up have to be kept somewhere, storage is hot, systems that quickly crunch through huge smart systems in cities. China alone will too. That explains the recent bidding war amounts of data). It is the biggest bet IBM need to add the equivalent of one New between Dell and HP over 3Par, a data• has made in any area, says Ambuj Goyal, York every year to accommodate the storage company. HP won by o ering the †rm’s global head of development number of people expected to migrate to $2.35 billion for a †rm with only 650 em• and manufacturing. If it comes good, IBM urban areas by 2025, according to McKin• ployees and $194m in revenues. Forecasts should be richly rewarded. The global sey. Big IT †rms are chasing all over the from IDC, another market•research †rm, market for analytics programs, for in• country to get a piece of this huge pie. Cis• throw light on why HP is willing to pay so stance, will grow from $25.5 billion this co is already helping to build a Chinese much. IDC expects the capacity shipped year to $34 billion in 2014, according to replica of Songdo, the smart•city project in to increase by 50% plus this year. IDC. South Korea. IBM, for its part, hopes to Vendors of programs that sift through This technology wave will almost cer• help develop an entire network of smart data are also likely to do well. In recent tainly produce new software champions cities and has started to collaborate with years IBM has invested billions of dollars too. SAP came of age when mainframes several of them.

2 duces usable data. And Joy Mining Ma• Michael Chui, a co•author of McKin• schinen, whose huge presses come with chinery, a maker of mining equipment, sey’s report on the internet of things, says more than 1,000 sensors, has started o er• charges for support by the tonne. that such applications will allow compa• ing services based on the data it collects, in• Some †rms are using metering in inno• nies to Œhave a much more dynamic inter• cluding a website that allows customers to vative ways. Zipcar and other car•sharing action with customers. In Japan, for in• compare their productivity with others. †rms, for instance, put wireless devices stance, vending machines can recognise a ŒMany companies will suddenly discover with sensors into their vehicles so that cus• customer’s age and sex and change the that their main business is data, says Paul tomers can hire them by the hour. And in• message they display accordingly. Sa o, a Silicon Valley technology forecast• surance †rms, among them Progressive in The more data that †rms collect in their er who wrote a widely noted essay on the America and Coverbox in Britain, ask cus• core business, the more they are able to of• impact of ubiquitous sensors back in 1997. tomers to install equipment in their cars fer new types of services. By continuously Hewlett•Packard is a prime candidate that can measure for how long, how fast assessing the performance of its jet en• for such a Sa o moment‹if its plans to and even where a car is driven. Premiums gines around the world, Rolls•Royce is able scatter millions of sensors around the can then be based on individual drivers’ to predict when engines become more world come to fruition. It is doing this to in• behaviour rather than on such proxies as likely to fail so that customers can schedule crease demand for its hardware, but it also age and sex. engine changes. Heidelberger Druckma• hopes to o er services based on networks 1 10 A special report on smart systems The Economist November 6th 2010

2 of sensors. For instance, a few thousand of loaded over 6.5 billion times. And with its †ne•tuning charges for public goods to get them would make it possible to assess the new platform for mobile advertising, citizens to behave in certain ways. state of health of the Golden Gate bridge in called iAd, Apple has started to make mon• Much of the innovation in this †eld San Francisco, says Stanley Williams, who ey from all the data it collects. may come not from incumbents but from leads the development of the sensors at Some †rms will make a living based en• newcomers, and it may happen fastest on HP. ŒEventually, he predicts, Œeverything tirely on mining Œdata exhaust, the bits such platforms as Pachube. In a way it is a will become a service. and bytes produced by other activities. cross of YouTube and Windows. What Apple, though it prides itself on its fan• One example is Google’s PowerMeter, made the video•sharing site so popular cy hardware, is already well on its way to• which not only lets users check their use of was the way it converted all videos to a wards transforming itself into a service electricity online but gives Google access common format. Pachube is doing the and data business thanks to the success of to lots of data to analyse and, not least, sell same for data feeds from sensors. And like its iPhone. When the computer•maker advertisements against. Microsoft’s operating system for applica• launched the device in June 2007, it did not Conventional services, too, can be me• tions, it provides basic features for smart expect Œapps, the applications that run on tered. Those supplied by governments services, such as alerts, data storage and vi• smartphones, to become such a big deal. may not be the †rst to spring to mind but, sualisation tools. The ŒApp Store, where users can down• as a study by the Dutch economics minis• This spring at Where 2.0, a technology load these pieces of software, was try asks, why not use sensors for taxing conference in Silicon Valley, the star of the launched a year after the †rst iPhone was things like pollution? That might be uncon• show was Skyhook Wireless, a †rm that of• shipped. But the App Store now sells more troversial, but analytics software could fers geographical•location information as than 250,000 apps that have been down• also be put to more manipulative uses by a service. It recently launched a new o er•1

Your own private matrix Tracking your life on the web

IRROR WORLDS is only one of more about themselves, says Gary Wolf, a ŒMDavid Gelernter’s big ideas. An• technology writer and co•founder of a other is Œlifestreams‹in essence, vast blog called ŒThe Quanti†ed Self. They electronic diaries. ŒEvery document you want to use technology to help them iden• create and every document other people tify factors that make them depressed, send you is stored in your lifestream, he keep them from sleeping or a ect their wrote in the mid•1990s together with Eric cognitive performance. One self•tracker Freeman, who produced a doctoral thesis learned, for instance, that eating a lot of on the topic. Putting electronic documents butter allowed him to solve arithmetic in chronological order, they said, would problems faster. make it easier for people to manage all A market for self•tracking devices is al• their digital output and experiences. ready emerging. Fitbit and Greengoose, Lifestreams have not yet replaced the two start•ups, are selling wireless acceler• desktop on personal computers, as Mr Ge• ometers that can track a user’s physical ac• lernter had hoped. Indeed, a software tivity. Zeo, another start•up, has devel• start•up to implement the idea folded in oped an alarm clock that comes with a 2004. But today something quite similar headband to measure people’s brainwave can be found all over the web in many dif• their lives with wearable cameras. activity at night and chart their sleep on ferent forms. Blogs are essentially elec• The latest trend is Œlife•tracking. Prac• the web. tronic diaries. Personal newsfeeds are at titioners keep meticulous digital records As people create more such self•track• the heart of Facebook and other social of things they do: how much co ee they ing data, †rms will start to mine them and networks. A torrent of short text messages drink, how much work they do each day, o er services based on the result. Xobni, appears on Twitter. what books they are reading, and so on. for example, analyses people’s inboxes Certain individuals are going even fur• Much of this is done manually by putting (Œxobni spelled backwards) to help them ther than Mr Gelernter expected. Some the data into a PC or, increasingly, a smart• manage their e•mail and contacts. It lists are digitising their entire oˆce, including phone. But people are also using sensors, them according to the intensity of the pictures, bills and correspondence. Oth• mainly to keep track of their vital signs, for electronic relationship rather than in al• ers record their whole life. Gordon Bell, a instance to see how well they sleep or phabetical order. Users are sometimes researcher at Microsoft, puts everything how fast they run. surprised by the results, says Je Bonforte, he has accumulated, written, photo• The †rst self•trackers were mostly the †rm’s boss: ŒThey think it’s creepy graphed and presented in his Œlocal cyber• über•geeks fascinated by numbers. But the when we list other people before their space. Yet others Œlog every aspect of more recent converts simply want to learn girlfriend or wife. The Economist November 6th 2010 A special report on smart systems 11

2 ing called SpotRank. Drawing on all the data it has collected in recent years from apps using its services, the †rm can predict the density of people in speci†c urban ar• eas‹anywhere, any day and at any hour. ŒThis will give us great insights into human behaviour, says Brady Forrest, the chair• man of the conference. Another hit at the conference was a ser• vice called Wikitude. Its ŒWorld Browser, a smartphone app, checks the device’s lo• cation as well as the direction in which its camera is pointing and then overlays virtu• al sticky notes other users have left about things like local landmarks. So far Wiki• tude and similar services are mostly used as travel guides. But in principle users they should lead to improved pricing and be on their way to becoming a factor of could collaboratively annotate the entire allocation of resources. Second, the inte• production in their own right, just like physical world‹and even other people. gration of the virtual and the real will land, labour and capital. That will make TAT, another start•up, already experi• speed up the shift from physical goods to companies and governments increasingly ments with something called ŒAugmented services that has been going on for some protective of their data assets. ID, which uses facial recognition to dis• time. This also means that more and more In short, we may be moving towards a play information about a person shown things will be hired instead of bought. ŒWeightless World, the title of a 1997 book on a smartphone’s screen. Third, economic value, having migrated by Diane Coyle about a future in which It is diˆcult to say what e ect all this from goods to services, will now increas• bytes are the only currency and the things will have on business and the economy. ingly move to data and the algorithms that shape our lives have literally no But three trends stand out. First, since used to analyse them. In fact, data, and the weight. But for now, gravity has not quite smart systems provide better information, knowledge extracted from them, may even been repealed yet. 7 Sensors and sensibilities

A smarter world faces many hurdles

HERE is not much to see in the city of before embracing the technology. ŒBakers• things as smart meters need to be sorted TBakers†eld, north of Los Angeles, but †eld is likely to slow down the installation out. Setting them too early would hamper recent events have put it on the global elec• of smart meters‹not just in the United innovation, but in their absence utilities tricity•industry map. As in many other States but worldwide, says Ahmad Faru• will hold back from investing, worried that Californian communities, Paci†c Gas & qui of the Brattle Group, a consultancy. Ba• they might bet on the wrong technology. Electric (PG&E), the local power utility, had kers†eld also demonstrates that a smarter Standards could also become a weapon of installed smart meters in most house• world will meet with resistance. The rea• industrial policy, in particular in countries holds. Soon afterwards customers started sons are part technical, part institutional. that see clean technology as an engine of complaining about rocketing power bills. growth. When China’s State Grid Corpora• For some people they almost trebled. Pre• A list as long as your arm tion, which operates most of the country’s dictably, this caused a political storm. Local Technology is a good place to start. Sensors power network, announced its smart•grid politicians and consumer groups jumped are getting ever cheaper, but for many ap• plans in June, it also released the standards on the issue. Enterprising lawyers plications they are still much too expen• it intends to implement. Some say this was launched a class•action suit. PG&E admit• sive. HP, for instance, likes to point out that a move to protect Chinese †rms. ted that some of its meters had technical its super•sensitive accelerometers are The internet•address system is a worry problems, but the higher bills were clearly made in the same factories as its printer as well. For a computer or any other device due to a combination of exceptionally hot cartridges. But the †rm’s sensors are still to be part of the internet, it needs a unique weather, increased charges and changes in too pricey to use them for anything but identi†er‹currently a long number called the rate structure. high•value applications, such as oil explo• an internet•protocol (IP) address. Because An independent auditor found nothing ration. RFID tags are a cautionary tale. of the network’s rapid growth in recent wrong with the smart meters, and Califor• They were supposed to revolutionise re• years, these numbers could run out as ear• nia’s regulators did not stop PG&E from in• tailing but the readers, software and other ly as the middle of next year. If the IT in• stalling more of them. But utilities and reg• gear needed to make them usefulis still not dustry keeps dragging its feet on moving to ulators elsewhere, spooked by the cheap enough to be universally adopted. IPv6, a new address system that uses many incident, have become much more careful Equally important, standards for such more numbers, the growth of the internet 1 12 A special report on smart systems The Economist November 6th 2010

2 of things will be stymied. to have an advantage in building smart sys• cipal administration to become a smart Space is also bound to get tight in the tems because governments are often less city, it created Amsterdam Innovation Mo• ether. A few wireless sensors and devices democratic and administrations more tor (AIM), a public•private joint venture in do not make a di erence, but as their num• hierarchical. China’s State Grid Corpora• charge of coming up with projects and me• bers grow they will need an ever bigger tion intends to have its smart grid fully diating between the parties involved. ŒBe• chunk of the available radio spectrum. The built by 2020. The country’s government ing a translator and making sure that a pro• number of wireless subscriptions has now has also made the implementation of IPv6 ject is worthwhile for everybody are our reached 5 billion worldwide, earlier than a central part of its †ve•year plan to build main jobs, explains Ger Baron, AIM’s pro• expected, not least because so many SIM the ŒChina Next Generation Internet. It ject manager. cards now sit in machines that communi• showed o its e orts at the 2008 Olympic If these three examples are any guide, cate with other machines. games in Beijing, where everything that smart systems may well change the way And then there are security concerns, was connected‹cameras, taxis, security that local governments, in particular, are particularly after the Stuxnet worm made systems‹used IPv6. organised. Instead of being a collection of the rounds in September. The malicious In the West it will often take a crisis to departmental silos, they could come to re• program quickly found its way into com• get there. When Thames Water in 2006 semble computing platforms. Most ser• puters controlling industrial processes the failed to meet targets set by the regulator to vices, from payment systems to traˆc in• world over, demonstrating how vulner• reduce leaks and was subsequently sold, formation, would be provided in only one able control systems are to such attacks. the new leadership went on to organise version and used by all departments‹or But even before Stuxnet struck, security things di erently. Today, at the utility’s op• by private †rms that want to o er their consultants had shown how large num• erating centre in Reading, the workers who own urban applications. bers of smart power meters could be monitor the network, take calls from cus• Some cities, such as London, San Fran• hacked and shut down. tomers and schedule work crews all sit in cisco and Washington, DC, have already one open•plan oˆce, allowing them to opened their data vaults. IBM, among Turf, ego and power communicate much more easily across de• many other companies, has already built a Yet all these technical issues pale by com• partmental boundaries. web•based application called City For• parison with the institutional barriers. For Similarly, when Bill Ritter became Col• ward that takes in data from 50 cities. Yet a city to o er smart services and save mon• orado’s governor in 2007, he made the con• others are not that generous with their ey, its departments have to work closely to• solidation of the state’s chaotic IT systems data, which is the third barrier. More open• gether, share their data and use a common a priority and named a state chief informa• ness should be good for innovation, not IT infrastructure. London, for instance, has tion oˆcer who is also a member of his just in terms of the information itself but di erent payment systems for public tran• cabinet. Since then Colorado has made how it is handled. But †rms with lots of sport, bicycle hire and toll roads. Such frag• great progress in achieving one prerequi• data‹be they power utilities or makers of mentation is costly and makes it more diˆ• site for becoming a smart state: a common medical equipment‹will be loth to give cult to come up with new o ers (say, IT infrastructure capable of delivering new them up, says Glen Allmendinger, presi• reducing the congestion charge for those services. dent of Harbor Research. who often hire a bicycle). But getting a Amsterdam, being the capital of a high• At the same time, Mr Allmendinger pre• city’s islands of bureaucracy to work to• ly pluralistic country, had to take a di erent dicts, some †rms will be forced to give up gether tends to be diˆcult, says Mark Clev• approach. Instead of relying on the muni• control of their data. Hospitals, for in• erley of IBM, who helps governments and stance, will hardly put up with dozens of cities develop plans for smart systems. dashboards that monitor the activities of The problem is not just that depart• di erent types of equipment: they will ments often jealously protect their data, want a uni†ed view. And some clever start• something experts call TEP, as in Œturf, ego ups and IT †rms will †nd ways around and power. Oˆcials also lack a common data monopolies. Nobody can stop con• language or generally agreed criteria for a sumers from giving data about their power smart city‹which is a big issue, too, for the usage to non•utilities, for instance. Some many companies that are usually involved private meters, attached to a sensor in a project. ŒIt’s hard to build a business clamped around the main power line, can case if people don’t understand each oth• now send the data they have gathered to er, says Simon Giles, in charge of strategy web•based energy•monitoring services, for smart technologies at Accenture. such as Google’s PowerMeter and Micro• Things are easier in Singapore. Minis• soft’s Hohm. tries and agencies compete for reputation Data are a problem for governments and resources, but they also co•operate too. Li Yizhong, China’s minister of indus• closely on implementing master plans try and information technology, has ex• such as ŒA Lively and Liveable Singapore: pressed concerns about IBM’s Smarter Strategies for Sustainable Growth, the Planet initiative. ŒThe US tries to use its in• city•state’s roadmap to becoming smart. formation network technology for things That helps to explain why Singapore will as small as controlling one computer or probably be the †rst city to combine its va• one generator and as large as controlling a rious smart systems into a single one. whole industry, to control every country’s More generally, Asian countries seem economy, he is reported to have said. ŒWe 1 The Economist November 6th 2010 A special report on smart systems 13

2 must be enlightened and vigorously devel• challenge hidden assumptions in a bewil• nearly 270m of these devices will be sold op strategic emerging industries, but also dering array of doctrinal †elds, writes Ke• this year, 55% more than in 2009, says IDC. must raise our vigilance and cannot fall un• vin Werbach, of the Wharton School of the The di erence is that Apple and other der the control of anyone. University of Pennsylvania, in a paper en• smartphone makers have made it their China’s concerns with IBM’s Smarter titled ŒSensors and Sensibilities. business to †nd out what consumers Planet (called Œwise Earth in Mandarin) Lastly, consumers may not play ball. want‹traditionally not the forte of utilities point to the fourth set of barriers: govern• PG&E’s woes in Bakers†eld are not the and government agencies. For example, it ment regulation‹or the lack thereof. Pri• only example. In Boulder, Colorado, the took that ‡urry of protests to prompt PG&E vacy legislation tops the list. New laws will showcase for smart grids, customers of to open a dedicated call centre for ques• multiply in response to an increasingly Xcel Energy, the local utility, are becoming tions about smart meters. Yet communicat• smart world. Germany’s government increasingly cross because they will have ing with customers should be one of the plans to strengthen its rules to deal with to pay much of the project’s costs. In the †rst things for †rms to do when introduc• Google’s Street View, an online service that Netherlands the government backed ing smart meters, says Mr Giles. lets users pan through photos of streets. down from making smart meters manda• Another lesson is that utilities should Among other things, the new bill is likely tory because of concerns that the data col• not try to achieve too many things at once, to enshrine in law what Google has al• lected could be used to see whether prop• says Mr Giles, who recently surveyed ready agreed to do under pressure from erties are empty or expensive new gadgets smart•grid projects worldwide for a study data•protection oˆcials: giving people the have been purchased. Consumers also dis• on how to speed up their introduction. option to have their house blurred on like feeling that they are being squeezed Some utilities brought in smart meters and Street View. Nearly a quarter of a million dry, particularly in America, which makes new pricing schemes at the same time, have done so. dynamic pricing hard to bring in. Even sim• thus overwhelming consumers and ob• The regulation of smart grids is a murki• ple rate plans where the price of electricity scuring the reasons for higher bills. er area. Some countries push utilities to re• depends on the time of day have had to be To avoid such problems, Amsterdam is duce energy demand, but out•of•date rules abandoned after customer protests. trying Œco•creation, in the words of AIM’s encourage them to sell more, says Accen• Nor is such resistance limited to smart Mr Baron. ŒWe did not just put in smart me• ture’s Mr Giles. Elsewhere, ill•conceived meters. The smooth introduction of Stock• ters and ask consumers to pay for them, deregulation of the power market is hold• holm’s toll system was the exception rath• he says. Instead 500 households in the dis• ing things up. Since the grid now has more er than the rule. In many countries politi• trict of Geuzenveld were invited to make than one owner, it is often hard to know cians do not even try. Germany, for suggestions on how to save energy and who will bear the risks and who will gar• instance, charges lorries for using its mo• monitor consumption. ner the rewards. torways, but only a suicidal government Yet utilities may have to resort to social A host of new legal questions will also would attempt to extend the system to cars engineering to get customers more en• have to be answered. Who is liable if an in a country where even buckling up was gaged. Opower, a start•up, lets them see autonomous system, such as an autopilot long opposed by motoring clubs as inter• not only their own consumption †gures that governs the movement of cars on a fering with drivers’ freedom. but numbers for their neighbours too (ano• motorway, causes an accident? What if a nymised to preserve privacy), and o ers single company manages to dominate data Make it attractive them tips on saving power. Peer pressure, from certain types of sensors or control the It is odd, then, that everybody loves mo• the company claims, persuades people to information about where sensors can be bile devices, which are not that di erent reach for the switches much more often. found and which of them are active? And from smart meters or on•board units. In Given enough time, all these barriers to can private surveillance still be restricted particular, smartphones and the applica• building smart systems can probably be when cameras and other sensors make it tions that run on them generally keep a overcome. But how smart does the world ubiquitous? ŒThe sensor revolution will close watch on what users do. Even so, really want to be? 7 Horror worlds

Concerns about smart systems are justi†ed and must be dealt with

T IS not possible to make a lasting paign or somebody might recognise his argued, smart systems will improve eˆ• ŒIcompromise between technology style of writing and unmask him. ciency and could help solve many envi• and freedom, because technology is by far Mr Kaczynski’s methods were abhor• ronmental problems, in particular global the more powerful social force and contin• rent. His bombs killed three people and in• warming. Yet if those systems seriously ually encroaches on freedom through re• jured 23 over nearly 20 years. He was ar• impinge on people’s freedom, many peo• peated compromises. Thus wrote the Un• rested in 1996 and is currently serving a life ple will balk. The protests against smart abomber, also known as Ted Kaczynski, in sentence. But his concern that technology meters in Bakers†eld and elsewhere may his manifesto, published in 1995 by the will slowly but surely undermine human be only the start. New York Times and the Washington Post in freedom is shared by quite a few main• Smart systems are rekindling old fears. the hope that he might end his terror cam• stream thinkers. As this special report has Top of the list are loss of privacy and gov•1 14 A special report on smart systems The Economist November 6th 2010

A further worry is that smart technol• ogy will ultimately lead to greater inequal• ity‹and not just because it could create an Œinformation priesthood, in the words of Mr Gelernter. Paul Sa o, a noted Silicon Valley technology forecaster, expects ubiq• uitous sensors to give a huge boost to pro• ductivity‹at the expense of human moni• tors. ŒWe are likely to see more jobless recoveries, he says. Whether computers will indeed start to eliminate more jobs than they create re• mains to be seen. But smart systems cer• tainly represent a conceptual change. So far IT has been used to automate and opti• mise processes within †rms and other or• 2 ernment surveillance. Internet users have spin out of control and even take over the ganisations as well as the dealings be• only recently begun to realise that every world, as they did in the †lm ŒThe Matrix. tween them. Now it will increasingly be single thing they do online leaves a digital As the Stuxnet worm and the May Œ‡ash used to automate and optimise interac• trace. With smart systems the same thing crash on Wall Street have shown, the †rst tions with the physical environment. will increasingly apply to the o‰ine world; two are already real possibilities, even if Some of the concerns raised will be Google’s Street View is only the beginning. the third still seems somewhat remote‹al• hard to deal with. For instance, there Even the champions of a smarter plan• though well•known computer scientists, would be little point in passing laws that et admit as much. ŒSome citizens have ex• arti†cial•intelligence researchers and ro• would give individuals the right to decide pressed discomfort at living in not a safer boticists met in California a couple of whether their data can be used by smart society but a ‘surveillance society’, said years ago to discuss that risk. systems if cameras and other sensors are Sam Palmisano, the boss of IBM, in a And there is a more subtle danger too: already ubiquitous. And building in cir• speech earlier this year. He cited a newspa• that people will come to rely too much on cuit•breakers to keep automation from go• per article recounting that there are now 32 smart systems. Because humans cannot ing too far could defeat the purpose of closed•circuit cameras within 200 yards of cope with the huge amounts of data pro• smart systems and sti‡e innovation. the London ‡at in which George Orwell duced by machines, the machines them• Still, technological progress is not some wrote his book Œ1984. selves will increasingly make the deci• force of nature that cannot be guided. ŒWe sions, cautions Frank Schirrmacher of the can and we should exercise control‹by Hopes and fears German daily Frankfurter Allgemeine Zei• democratic consensus, says Mr Gelernter. Mr Palmisano would be in the wrong job, tung in his recent book ŒPayback. Similar• Yet for a consensus to be reached, there however, had he not gone on to say that ly, Nicholas Carr, an American commenta• must be openness. The biggest risk is that such concerns have to be rethought and to tor on the digital revolution, in his book smart systems become black boxes, closed stress the economic and social bene†ts of ŒThe Shallows claims that the internet, even to citizens who have the skills to un• smart systems. Others point out less obvi• the mother of all smart systems, is on its derstand them. Smart systems will make ous advantages. ŒAll this technology actu• way to smothering creativity and pro• the world more transparent only if they ally strengthens the human side of cities, found thinking. themselves are transparent. 7 says Carlo Ratti, director of MIT’s SENSEa• ble City Lab. People who are always con• O er to readers Future special reports nected, he argues, can work wherever they Reprints of this special report are available from like. And Mr Haque, the boss of Pachube, Japan November 20th 2010 the Rights and Syndication Department. Security in Asia December 4th 2010 claims that Œsensors empower people be• A minimum order of †ve copies is required. cause measuring the environment allows Global leaders January 22nd 2011 Corporate o er them to make decisions in real time. Food February 26th 2011 Customisation options on corporate orders of 100 Property March 5th 2011 On the other hand smart systems are or more are available. Please contact us to discuss also undeniably useful as an instrument of your requirements. control. Singapore has made an impressive For more information on how to order special job of smartening up its physical infra• reports, reprints or any queries you may have structure, but its network of security cam• please contact: eras could also be used for enforcing rules more objectionable than a ban on chewing The Rights and Syndication Department 26 Red Lion Square gum. Similarly, the operations centres and London WC1R 4HQ dashboards for local governments in Chi• Tel +44 (0)20 7576 8148 na being built by Cisco, IBM and others beg Fax +44 (0)20 7576 8492 the question whether their only purpose is e•mail: [email protected] Previous special reports and a list of to make these cities smarter. www.economist.com.rights forthcoming ones can be found online Other deep fears brought on by smart Economist.com/rights Economist.com/specialreports systems is that machines could be hacked,