Upload
leque
View
214
Download
2
Embed Size (px)
Citation preview
It’s a smart worldA special report on smart systemsNovember 6th 2010
SmartsystemCOV.indd 1 26/10/2010 12:07
The real and the digital worlds are converging, bringing much greatere�ciency and lots of new opportunities, says Ludwig Siegele. But is itwhat people want?
many places at once and is often not understood for what it is. It is most advancedin controlled environments. For example,software developed by Siemens, a technology conglomerate, maintains virtual replicas of factories to monitor and recon�gurethem. But it is spreading everywhere andhas developed a language all of its own.Glen Allmendinger of Harbor Research, aconsultancy, calls it the �virtualisation ofthe real world�. Researchers at MIT’s Media Lab who connect reallife objects withcopies in Second Life, a virtual world, referto the result as �cross reality�. Google’sEarth and Street View services are the �rst,if static, replicas of the entire world; sensors placed in cows allow the tracking oftheir every move from birth to abattoir;smart power meters tell utilities in realtime how much electricity is used.
The many uses of smartphonesYet it is the smartphone and its �apps�(small downloadable applications thatrun on these devices) that is speeding upthe convergence of the physical and thedigital worlds. Smartphones are packedwith sensors, measuring everything fromthe user’s location to the ambient light.Much of that information is then pumpedback into the network. Apps, for their part,are miniature versions of smart systemsthat allow users to do a great variety ofthings, from tracking their friends to controlling appliances in their homes.
The Economist November 6th 2010 A special report on smart systems 1
It’s a smart world
WHAT if there were two worlds, thereal one and its digital re�ection? The
real one is strewn with sensors, picking upeverything from movement to smell. Thedigital one, an edi�ce built of software,takes in all that information and automatically acts on it. If a door opens in the realworld, so does its virtual equivalent. If thetemperature in the room with the opendoor falls below a certain level, the digitalworld automatically turns on the heat.
This was the vision that David Gelernter, a professor of computer science at YaleUniversity, put forward in his book �MirrorWorlds� in the early 1990s. �You will lookinto a computer screen and see reality,� hepredicted. �Some part of your world�thetown you live in, the company you workfor, your school system, the city hospital�will hang there in a sharp colour image, abstract but recognisable, moving subtly in athousand places.�
Even two decades later that sounds likescience �ction. But this special report willargue that Mr Gelernter was surprisinglyprescient: mankind is indeed buildingmore and more �mirror worlds�, or �smartsystems�, as they are often called. The realand the digital worlds are converging,thanks to a proliferation of connected sensors and cameras, ubiquitous wireless networks, communications standards and theactivities of humans themselves.
This convergence may not be instantlydiscernible, because it is happening in
An audio interview with the author is at
Economist.com/audiovideo/specialreports
A list of sources is at
Economist.com/specialreports
A sea of sensorsEverything will become a sensor�and humansmay be the best of all. Page 3
Making every drop countUtilities are getting wise to smart meters andgrids. Page 4
Living on a platformFor cities to become truly smart, everythingmust be connected. Page 6
Augmented businessSmart systems will disrupt lots of industries,and perhaps the entire economy. Page 8
The IT paydirtWho will clean up? Page 9
Your own private matrixTracking your life on the web. Page 10
Sensors and sensibilitiesA smarter world faces many hurdles. Page 11
Horror worldsConcerns about smart systems are justi�edand must be dealt with. Page 13
Also in this section
AcknowledgmentsBesides those cited in the text, the author would like tothank many others for their time, ideas and kindnessduring the preparation of this report. They include BrunoBerthon of Accenture, Eric Clementi of IBM, TomDavenport of Babson College, Peter Elliott of PAConsulting Group, Scott Gnau of Teradata, Jeanne Harrisof Accenture, Chris King of eMeter, Michael Liard of ABIResearch, Michael LoCascio of Lux Research, EdzardOverbeek of Cisco, Joachim Schaper of SAP and ElisabethZornes of Cisco.
1
2 A special report on smart systems The Economist November 6th 2010
2 Smartphones are also where the virtualand the real meet most directly and mergeinto something with yet another fancyname: �augmented reality�. Download anapp called �Layar� onto your smartphone,turn on its video camera, point at a street,and the software will overlay the pictureon the screen with all kinds of digital information, such as the names of the businesses on the street or if a house is for sale.
These and other services are bound togrow together into what Jan Rabaey, a computer scientist at the University of California at Berkeley, grandly calls �societal informationtechnology systems�, or SIS.Technological progress is sure to supplythe necessary components. Moore’s law,which holds that the processing power of asingle computer chip roughly doubles every 18 months, applies to sensors, too.
More processing power and better connectivity also allow the construction ofcomputing systems capable of storing andcrunching the huge amounts of data thatwill be produced by these sensors and other devices. All over the world companiesare putting together networks of data centres packed with thousands of servers,known as �computing clouds�. These notonly store data but sift through them, forinstance to allow a smart system to reactinstantly to changes in its environment.
The next big thingInformationtechnology (IT) �rms haveidenti�ed smart systems as the next bigthing. Predictably, the most ambitious designs have been produced by the industry’s giants, particularly IBM, where SamPalmisano, the �rm’s boss, made smartsystems a priority. A couple of years agothe company launched a campaign called�Smarter Planet�, touting digital technology that would make energy, transport, cities and many other areas more intelligent.Other �rms have followed suit, each witha di�erent take re�ecting its particularstrengths.
Cisco, the world’s biggest maker of networking gear, is trumpeting �Smart+Connected Communities�. HewlettPackard,number one in hardware, intends to spin a�Central Nervous System for the Earth�.Siemens and its competitor General Electric, which are more at home in the physical world, plan to put together lots of smartsystems in which they can deploy theirdeep knowledge of certain industries,such as health care and manufacturing.And there is a growing wave of �smart�startups, o�ering everything from services to pinpoint a devices’s location to
platforms for sensor data.Governments, too, have jumped on the
bandwagon. Many countries have beenspending large chunks of their stimuluspackages on smartinfrastructure projects,and some have made smart systems a priority of industrial policy. The �internet ofthings�, another label for these systems, iscentral to the European Union’s �DigitalAgenda�. The main contenders in this market are countries that are strong in manufacturing, above all Germany and China.
But the bandwagon is not just rolling forthe bene�t of technology companies andambitious politicians. It has gained momentum because there is a real need forsuch systems. In many countries the physical infrastructure is ageing, healthcarecosts are exploding and money is tight. Using resources more intelligently can maketaxpayers’ money go further. Monitoringpatients remotely can be much cheaperand safer than keeping them in hospital. Abridge equipped with the right sensors cantell engineers when it needs to be serviced.
China is a good example. It is becomingurbanised on a scale unprecedented in history. By 2025 an additional 350m Chinese�more than the current population of theUnited States�will have moved to cities,according to a study by McKinsey, a consultancy. Without an infrastructure enhancedby digital technology it will be very hard toprovide the country’s newly urbanisedpopulation with enough food, transport,electricity and water.
Most important, smart systems maywell be humankind’s best hope for dealingwith its pressing environmental problems,notably global warming. Today powergrids, transport systems and waterdistribution systems are essentially networks ofdumb pipes. If the power grid in America
alone were just 5% more e�cient, it wouldsave greenhouse emissions equivalent to53m cars, calculates IBM. In 2007 its congested roads cost the country 4.2 billionworking hours and 10.6 billion litres ofwasted petrol, according to the Texas Transportation Institute. And utilities aroundthe world lose between 25% and 50% oftreated water to leaks, according to Lux Research, a marketresearch �rm.
With so much to gain, what is there tolose? Privacy and the risk of abuse by a malevolent government spring to mind �rst.Indeed, compared with some smart systems, the ubiquitous telescreen monitoring device in George Orwell’s novel �1984�seems a plaything. The book’s hero, Winston Smith, would soon have a much harder time �nding a corner in his room to hidefrom Big Brother.
Second, critics fear that smart systemscould gang up on their creators, in the waythey did in �The Matrix�, a 1999 �lm inwhich human beings are plugged into machines that simulate reality to control humans and harvest their bodies’ heat andelectrical activity. Fortunately, such a scenario is likely to remain science �ction. Butsmart systems might be vulnerable to malfunctioning or attacks by hackers.
Third, some people fret that those withaccess to smart systems will be vastly better informed than those without, givingthem an unfair advantage. Mr Gelernterhighlighted this risk in �Mirror Worlds�.
There are plenty of other concerns, andunless they are dealt with they could provoke a neoLuddite reaction. The worldhas already seen one extreme example: theUnabomber, a disa�ected American whotargeted, among others, computer scientists with mail bombs. Two years after thepublication of �Mirror Worlds� he sent oneto Mr Gelernter, who was seriously injured�though fortunately he survived. 7
1The way to go
Source: Harbor Research
Total worldwide internet-enabled-device revenues$bn
0
2
4
6
8
10
12
2009 10 11 12
F O R E C A S T
13 14
Industrial Energy Buildings
Transport Consumer/professional
Other
The Economist November 6th 2010 A special report on smart systems 3
1
GERMANS are known to separate theirrubbish diligently. Some even have
dedicated containers for di�erent kinds ofmetal. But they may soon need yet anotherbin: for electronic labels, also known as radiofrequency identi�cation (RFID) tags. Ifnot kept and treated separately, these couldbe very di�cult to recycle, Germany’s Federal Environment Agency said last year.The number being thrown away each yearcould rise from about 86m now to 23 billion by 2020, according to the agency.
RFID tags, which have been used toidentify everything from cattle to tombstones, will not be the only type of sensorcrowding the planet. Anything and anyone�machines, devices, everyday thingsand particularly humans�can become asensor, gathering and transmitting information about the real world.
The concept of the �internet of things�dates back to the late 1980s, when researchers at Palo Alto Research Centre (PARC) inSilicon Valley imagined a future in whichthe virtual and the real world would beconnected. In the following years much ofthe academic work concentrated on bringing this about with RFID tags, which are reliable, inexpensive and do not require apower supply. When exposed to a radiosignal, they use its energy to send back theinformation they contain, mostly a longnumber identifying an object.
Now it is �active� tags (which have theirown power source) and, even more, wireless sensors that are attracting most of theinterest. As with all things electronic, theseare becoming ever smaller and more versatile. �Tell me what you need, and we canbuild it for you,� says Reinhold Achatz,head of corporate research at Siemens.Startups, too, are producing devices thatsense everything from the rarest chemicalto the most exotic bacteria. Optiqua, a Singaporean �rm, has come up with a chipthat measures how fast light passesthrough a water sample to detect contaminants. And a biosensor developed at theLawrence Livermore National Laboratoryin Berkeley, California, can identify about2,000 viruses and 900 bacteria.
Researchers are also on the way to solving two big problems that have held backthe deployment of sensors. One is power.
Having to run wires or regularly replacebatteries would be too di�cult. But sensors have started to power themselves byscavenging for energy in their environment, for instance in the form of light andmotion. Similarly, some sensors alreadymake more e�cient use of another scarceresource: radio spectrum. Smart powermeters form �mesh networks� to relaytheir readings.
Engineers working on sensors thinkthis will eventually lead to �smart dust��sensors as small as dust particles that canbe dispersed on a battle�eld, say, to spy onthe enemy’s movements. Such devices arestill far o�, but at HewlettPackard (HP) inSilicon Valley a taste�or more precisely, afeel�of things to come is on o�er evennow. To demonstrate the �rm’s newaccelerometer, a device the size of a cigarette box that measures the acceleration ofan object, Peter Hartwell, a researcher, putsit on his chest, and a graph of his heartbeatappears on a screen beside him. �This sensor�, he proudly explains, �is one thousand times more sensitive than those inyour smartphone.�
One day, Mr Hartwell and his colleagues hope, a network of one trillionsensors will cover the world and deliverdata to anybody who needs them, fromcarmakers to municipal governments. Fornow, the �rm has teamed up with an oilcompany, Royal Dutch Shell. The computermaker plans to dot a prospecting areawith thousands of wireless sensors. They
are designed to pick up the echo of the seismic vibrations created by contraptionscalled �thumper trucks� pounding theground. The data thus gathered allowthem accurately to pinpoint pockets of oiland gas.
Yet RFID tags, wireless sensors and, forthat matter, digital cameras (so far the mostwidely deployed sensor thanks to the popularity of mobile phones) are only half thestory. Many objects no longer need an electronic tag or even a barcode to be automatically identi�ed. For example, Goggles, aservice o�ered by Google, can recognisethings like book covers, landmarks andpaintings. Users simply take a picture andsend it to Google’s computers, which willsend back search results for the object.
Many of the innumerable machinesand devices that populate the physical environment also already come with somedatagenerating digital technology. Moreand more are getting connected so thatthey can communicate the informationthey contain to the outside world. Examples range from co�ee machines and refrigerators to aircraft engines and body scanners in hospitals. These can all now phonehome, so to speak, and provide their makers with fountains of data.
People powerMost important, however, humans themselves have turned out to be excellent sensors. Many provide information withoutany extra e�ort, just by carrying around amobile phone. TomTom, a maker of navigation devices, uses connection data frommobile networks to update directions ifthere are delays. Others are connecting additional sensors to smartphones. Such devices and smartphones are gradually turning people into the sensory organs of theinternet, say John Battelle, boss of Federated Media, an online advertising agency,and Tim O’Reilly, who heads O’Reilly Media, a publishing house. �Our cameras, ourmicrophones, are becoming the eyes andears of the web,� they write in a paper entitled �Web Squared�.
More surprising than such �crowdsensing�, as it has come to be known, is the willingness of many people actively to gatherand upload information. The best example
A sea of sensors
Everything will become a sensor�and humans may be the best of all
High note
Source: Twitter
Tweets per day, m
Jan
2006 07 08 09 10
Jul Jan Jul Jan Jul Jan Jul Jan Jul
0
20
40
60
80
100
2
4 A special report on smart systems The Economist November 6th 2010
2
1
is Twitter, the microblogging servicewhose 160m users send out nearly 100mtweets a day (see chart 2, previous page).When they see, hear or read something,they type it into their computer or smartphone, 140 characters at a time. �Twitter isthe canary in the news coalmine,� wroteJe� Jarvis, a newmedia savant, after theservice beat mainstream media to newsabout the earthquake that struck China’sSichuan province in May 2008.
But there are plenty of other examples.At OpenStreetMap, a wikistyle website,some 250,000 volunteers record theirwanderings, using their smartphones’ positioning functions. And SeeClickFix, astartup, has come up with a smartphoneapp that allows users to report such thingsas broken streetlights or rubbish that needsto be picked up.
Too much of a good thingIt does not take much imagination to seethat all these sensors will generate immense amounts of data. �They don’t makeenough disk space in the world to save allthe data if every household had a smartmeter,� says Jim Goodnight, the boss ofSAS, one of the pioneers of analytics software, programs that sift through data. �Infact the most di�cult question is to decidewhat to discard.�
The quantity of data likely to be produced is anybody’s guess. Estimates byIDC, a marketresearch �rm, need to be taken with a pinch of salt, because they are
sponsored by EMC, a maker of storage systems. But for what they are worth, theysuggest that the �digital universe��theamount of digital information created andreplicated in a year�will increase to 35 zettabytes, or 35 trillion gigabytes�enough to�ll a stack of DVDs reaching halfway toMars. Even that may prove a conservativeestimate if sensors and other datagenerating devices spread as predicted.
Fortunately, the tools to deal with thisdata deluge are getting better. Give a marker and a whiteboard to Bijan Davari, a researcher at IBM, and he will draw you apicture of the future of computing as heand his employer see it. On the left sidethere are small squares, representing allkinds of sensors. The data they produceare fed into something he calls the�throughput engine�, a rectangle on theright. This is a collection of specialisedchips, each tailormade to analyse datafrom a certain type of sensor. �A systemthat can’t deal with these streams separately would quickly become overloaded,�says Mr Davari.
IBM has already introduced a productbased on what it calls �stream computing�that can ingest thousands of �datastreams� and analyse them almost instantaneously. The natal care unit at the University of Ontario is testing such a systemto monitor babies born prematurely. Ittakes in streams of biomedical data, suchas heart rate and respiration, and alertsdoctors if the babies’ condition worsens.
Analytics software is improving, too. Ithas long been used to crunch through datathat are �structured�, or organised in a database, and develop models to predict, forexample, whether a creditcard transaction is fraudulent or what the demand for�ights will be around a public holiday.Now such programs can also interpret �unstructured� data, mainly freeform text.Earlier this year SAS launched a productcapable of analysing the �sentiment� ofchatter on social media, including Facebook and Twitter.
The software is also able to �nd the people who post the most in�uential comments on speci�c companies on Twitter,who can then be sent special marketingmessages. Indeed, Twitter itself is a kind ofcollective �lter that continuously sorts thecontent published on the web. And Facebook users, by tagging friends in the pictures they upload, allow the service to recognise these people on other pictures.�Meaning is ‘taught’ to the computer,�write Messrs Battelle and O’Reilly.
But the main goal of smart systems is�to close the loop�, in the words of a reporton the internet of things published inMarch by the McKinsey Global Institute.This means using the knowledge gleanedfrom data to optimise and automate allkinds of processes. The number of potential applications is vast, ranging frommanufacturing to heading o� car collisions. Yet the most promising �eld for nowmay be physical infrastructures. 7
LONDON’S streets can be a bit of a maze,but below ground things are even more
complex. Water pipes crisscross the city inall directions. Some areas used to havecompeting water companies, each ofwhich built its own system. Not evenThames Water, the utility that operates theBritish capital’s watersupply network today, knows exactly where all the pipes run.
Moreover, the network is ageing. Only a
few years ago more than half of the 10,000miles (16,000km) of water pipes below thestreets of London were over a hundredyears old and often burst. It did not helpthat over many years Thames Water,which was privatised in 1989, failed to invest enough. By the mid2000s Londonhad one of the leakiest watersupply systems in the rich world. Every day nearly900m litres of treated water were lost and
240 leaks had to be �xed.Over the past �ve years, though,
Thames Water has replaced 1,300 miles ofcastiron Victorian mains, those most likely to break, with plastic ones, reducingleakage to 670m litres per day. And whenthe �rm puts in new pipes, it also installsadditional wireless sensors, giving it a better view of its network. �We can now tellwhere we have a broken main even before
Making every drop count
Utilities are getting wise to smart meters and grids
The Economist November 6th 2010 A special report on smart systems 5
2
1
customers call us,� says Bob Collington, itshead of asset management.
Thames Water not only needs to knowwhat is going on in its network, but to beable to act quickly on the information. Thesame is true of infrastructure operatorsaround the world. Whether in water, power, transport or buildings, all are trying toturn their dumb infrastructures into something more like a central nervous system.That makes them pioneers of the convergence of the physical and the digital world.
Putting sensors and actuators (devicesto control a mechanism) into physical infrastructures is not exactly new. Known as�supervisory control and data acquisition�, such systems have been around fordecades. But many still require human intervention: workers have to be sent out todownload sensor readings or to �x problems. And even if sensors and actuatorsare connected, di�erent types often feedinto incompatible systems, so they cannotbe easily combined to automate processes.
The operations centre of Thames Waterin Reading, to the west of London, is a goodplace to see both the old and the new�andsoon the future. A big video screen showsexpected precipitation over the next fewhours, and workers monitor the water level of reservoirs on their own screens. But ifone of the pumps fails, they may still haveto make a call: not all the valves can be remotely controlled.
Thames Water is investing £100m($158m) so it can take action remotely andautomate a lot of its processes. If the project works, the system will not only automatically deal with leaks but also schedulework crews and send text messages to affected customers. Employees in the operations centre, explains Jerry White, the utility’s head of operational control, will thenspend less time monitoring the networkand more on making the utility’s processesmore e�cient.
A big chunk of this work will be analysing the data collected by all the systemsand correlating them with other information. Not every unexpected spike in thewater �ow is the result of a leak, says MrWhite. For instance, water use leaps afterdark during Ramadan and at halftimeduring World Cup football matches.
One day soon Thames Water may evenbe able to send out work crews before amain actually breaks. In early 2010 the �rmbegan using a webbased service providedby TaKaDu, an Israeli company, that acts asthe network’s �eyes and ears�, in the wordsof Amir Peleg, its founder and boss. The�rm analyses historical and online data to
provide a basis for comparison, enablingits algorithms to detect things that areabout to go wrong.
Similar progress is being made all overthe world. The scope for preventing wasteis enormous, in the water industry andelsewhere. Power utilities are well ahead,not least because they can use the grid itself to collect sensor data and controlswitches. Transport systems are behind,particularly roads, which often use nothing more than tra�c cameras. Even buildings are getting more automated, with continuous checks on their energy use.
At the edgeFor infrastructures to become truly smart,however, it is not enough to put more intelligence into the core of a network. Theedge�the interface with users and devices�also has to become clever. This is theidea behind smart metering, which hasmade a good deal of progress in the powerindustry. According to Accenture, a consultancy, there are about 90 smartgrid projects around the world today. By the end oflast year more than 76m smart meters hadbeen installed worldwide. That numberwill almost treble by 2015, to 212m, estimates ABI Research.
Smart meters and other gear needed tomake grids more intelligent will not comecheap. Morgan Stanley, an investmentbank, predicts that the worldwide smartgrid market alone will grow from $20 billion last year to $100 billion in 2030. Yet thebene�ts also promise to be huge: powersavings, reduced investment in electricitygeneration and lower carbon emissions.
The place to go to see the technology inaction is Boulder, Colorado, home to whatis considered the world’s �rst fully �edged�smart grid�. The local utility, Xcel Energy,
did not skimp. It deployed equipment thatautomatically reports power cuts. It installed more than 20,000 smart meters,connected them via a �breoptic network,launched a website to track power use andhas started to o�er pricing plans that encourage shifting consumption to o�peakhours. It has even equipped some households with gear that tells airconditioningsystems to turn themselves o� when demand for electricity is high, a mechanismcalled �demand response�.
The results so far are mixed. The systemhas certainly helped Xcel to run its gridmore e�ciently. The utility now knowswhat is happening in its network and power cuts have become rare. Problems can bepinpointed and �xed much more quickly.But customers are not using much lesspower than they did before.
Yet it is early days. Some �rms are already beginning to show what can beachieved with demand response. EnerNOC, an American energy middleman, forinstance, pays other �rms for allowing it toshut down their nonessential gear attimes of peak demand, thus freeing up capacity. By midyear some 3,300 customers,from steel plants to grocery stores, hadsigned up. Their combined consumption,which can be made available to other users if needed, is 4,800MW, exceeding theoutput of America’s largest nuclear plant.
The ultimate point of smart grids, however, is to allow dynamic pricing, withelectricity charges �uctuating in responseto demand. This could cut power demandby 1015% during peak hours, estimates Ahmad Faruqui of the Brattle Group, a consultancy�more than twice the reduction likely to be achieved by just giving customersrealtime information about their usage.That number could easily double again, hesays, with a combination of dynamic pricing and demand response.
The main objective of smart power meters is to lower the peak load and thus enable utilities to keep down their peak generating capacity. In the water industry theeconomics are somewhat di�erent, explains Stefan Helmcke, a water expert atMcKinsey. Water can be easily stored andconsumers have less discretion over whenthey use it (for instance, people cannot defer going to the toilet, which uses more water than any other activity at home), so thecase for smart water meters is weaker.
Yet they are spreading all the same. Boston has long been the shining example. Asearly as 2004 the city’s Water and SewerCommission had equipped almost all itscustomers with wireless smart meters. But
3Splashing
Sources: WRI; The Economist
Daily domestic water consumption per person 2000, litres
0 200 400 600 800
Canada
United States
South Korea
Japan
Brazil
Germany
India
Britain
China
6 A special report on smart systems The Economist November 6th 2010
2
1
it will soon be outdone by New York,which plans to install more than 800,000of the devices at a cost of about $250m.Even Thames Water, most of whose customers have no meters of any sort, is nowplanning to install some of the smart kind.
Getting on boardIn transport the equivalent of a smart meter is a vehicle’s onboard unit. That used tobe a simple device, working like a radiofrequency identi�cation tag when it passesunder a gantry on a toll road, but it is alsogetting smarter. Germany’s Toll Collectsystem, which ensures that lorry driverspay for using the country’s crowded motorways, relies on gadgets that are in someways as clever as a smartphone. Amongother things, they keep track of their position with the help of GPS, the satellitebased global positioning system.
Such toll systems are multiplying, particularly in big congested cities, includingLondon and Stockholm. But it is Singaporethat leads the pack. The citystate not onlycharges drivers for using muchtravelledroads (driving on an expressway can beS$6, or $4.60); it also adjusts tra�c lights tosuit the �ow of vehicles, uses data collected by taxis to measure average speed and isdeveloping a parkingguidance system,noting that cars looking for somewhere topark are now a big cause of congestion.
Singapore may also become the �rst
city to introduce realtime dynamic pricingon its roads. In 2006 the Land TransportAuthority tested a tra�cprediction system built by IBM to set the tolls. And nextyear it plans to test a satellitebased systemthat does not require gantries and cancharge according to how congested a roadis at that particular time.
Another of the island’s infrastructuremanagement systems has become a model for the world: that for water. At the information centre at the southern tip of the island, next to the Marina Barrage, visitorscan literally get a taste of it by picking up abottle of �NEWater�, waste water that afterextensive treatment has become potableagain. But most of the treated water is fedback, via a separate distribution system, toSingapore’s factories and power plants�and then treated again.
This closed loop is part of a watersupply system in which �every drop counts,�in the words of Yap Kheng Guan, a directorat the island’s Public Utilities Board (PUB).The Marina Barrage is another case inpoint. It was inaugurated in 2008 and actsas a tidal barrier to keep seawater out, thusturning the island’s most populated district into a watercatchment area and theharbour into a reservoir. When two otherreservoirs are opened next year, more thantwothirds of Singapore’s territory will beused to catch rainwater.
The citystate’s desalination plants are
also among the world’s most e�cient. Allthis means that the island�smaller thanLuxembourg and home to nearly 5m people as well as an economy nearly as big asthat of Hong Kong�is able to meet morethan 60% of its water needs on its own. Butit wants to go even further: 50 years fromnow it hopes to be selfsu�cient.
Sensors play a relatively small part inSingapore’s water management becausethe infrastructure is so new. On averagethere is only one leak a day. The PUB putssensors only in a few key spots, for instance where water leaves the reservoirs.Should the system detect a dangerous contamination, that part of the network can beshut down immediately. And if heavyrainfall in central Singapore threatens to�ood the city during high tide, seven hugepumps next to the Marina Barrage start topush water into the sea at 40 cubic metresper second each.
So far Singapore has no smart watermeters, and at the moment there is nopressing need. Most Singaporeans live inmultistorey apartment buildings, whichmakes it easy to read meters. But if the PUB
wants to reach its target of cutting daily domestic water use per person from 155 litresin 2008 to 147 litres by 2020 (about thesame as in India, and a quarter of the �gurein America, see chart 3 on previous page),Singapore will have to become smarterstill�and set yet another example. 7
IN SINGAPORE conversations about water quickly turn political. The citystate
no longer wants to depend on water fromMalaysia when the current watersupplyagreement between the two countries expires in 2061. More than once the neighbour to the north, of which Singapore waspart before an acrimonious split in 1965,has threatened to increase prices or evencut o� supplies.
Yet politics is not the only reason forSingapore’s advanced water system. Theinformation centre at the Marina Barragefeatures pictures of �oods and droughts.�We have either too much water or too little,� explains Yap Kheng Guan, a directorof Singapore’s PUB. Even today, despite asophisticated system of ditches and tunnels, �oods can suddenly strike. In July
parts of the main shopping district wereunder water after heavy rainfalls.
The problems of scarcity and excess arein evidence on the citystate’s roads too.Singaporeans, who are among the world’srichest people, love to drive, but space forroads is severely limited. When in the early1970s the central area became too congested, the government introduced the world’s�rst manual urban roadpricing system. In1998 it became the �rst to be automated.�Singapore proves that necessity is themother of invention,� says Teo Lay Lim,who heads the local o�ce of Accenture.
Now the city wants to become a �livinglaboratory� for smart urban technologiesof all kinds�not just water and transportsystems but green buildings, clean energyand city management too. Both local and
foreign �rms in these sectors will be able todevelop and show o� their products onthe island before selling them elsewhere,explains Goh Chee Kiong, who is in chargeof the cleanenergy cluster at Singapore’sEconomic Development Board.
There is strong demand for making cities smarter, not just in China and other rapidly urbanising countries but throughoutthe Western world. Resources like water,space, energy and clean air are scarce in urban areas, which makes them the naturalplace to start saving, says Mark Spelman,Accenture’s global head of strategy.
�Smartcity� projects have been multiplying around the world. Some of themare not as new as their labels suggest, andin any case what exactly constitutes asmart city is hard to de�ne. But they all
Living on a platform
For cities to become truly smart, everything must be connected
The Economist November 6th 2010 A special report on smart systems 7
2
1
have one thing in common: they aim to integrate the recent e�orts to introduce smartfeatures in a variety of sectors and use this�system of systems�, as IBM calls it, tomanage the urban environment better.
The bestknown smart city is Masdar, abrandnew development in Abu Dhabithat recently welcomed its �rst residentsand will eventually become home to40,000 people. It is being built entirely ona raised platform, which makes maintenance and the installation of new gearmuch easier. Below the platform sits thesmart infrastructure, including water pipeswith sensors and a �breoptic network.Above it is to be a showcase for all kinds ofgreen technology: energye�cient buildings, small pods that will zoom around onpaths (no cars will be allowed) and systems that catch dew as well as rainwater.
Yet experts see Masdar mostly as aproperty project: hardware in search of apurpose. What really matters to a city’ssmartness, they argue, is the software thatruns on it and the network that connects itsparts. �It is the common infrastructure forall the smart systems,� says Wim Elfrink,who heads the �Smart+Connected Communities� initiative of Cisco, the networkingequipment maker.
Getting it togetherCisco is trying to demonstrate that point inSongdo City near Seoul, where the �rmprovides all the digital plumbing. Sittingon reclaimed land, Songdo is perhaps themost ambitious smartcity project so far. Itis expected to cost $35 billion and will behome to 65,000 people. Like Masdar, it willboast all the latest green technologies. Butits main claim to fame is that everything inthe city is wired up.
Residents of �First World�, the �rst completed apartment complex, already enjoythe bene�ts of this allembracing connectivity. Smartphones unlock front doors.Airconditioning, blinds and security systems are controlled by displays all over theapartment which can also be used to access all kinds of online services. With a fewclicks or touches users are able set up avideoconference with a doctor, do business with the local government or �nd outhow best get to work.
What Cisco sees as the most importantapplication for running a smart city wasshown at the World Expo in Shanghai thisyear. In its pavilion the �rm built a command centre to keep tabs on an imaginarysmart city. A huge video screen displayedeverything from tra�c maps and energyuse to weather information and pictures
from security cameras. Visitors were givena demonstration of how city managerswould react to an accident on a citycentrebridge: cameras zoom in, an ambulance isdispatched, tra�c is rerouted to otherbridges�all automatically, within seconds.
The world’s smartest city, however,may soon rise in an unexpected place: nearPorto in Portugal. PlanIT Valley, designedfor an eventual population of 150,000, isan ambitious attempt to �combine technology and urban development�, in thewords of Steve Lewis, cofounder of LivingPlanIT, the startup behind the project,who used to work at Microsoft.
His experience at the software giantproved an excellent preparation for thejob. Microsoft is the very model of a platform company, providing technology toconnect things (such as printers and PCs)and a foundation for the products made byothers (such as browsers and media players). When he was at Microsoft in the early2000s, Mr Lewis also oversaw the relaunch of a strategy called .Net�an earlyexample of what geeks like to call �serviceoriented architecture�. The idea is to buildprograms as a combination of looselycoupled electronic services that can be redeployed elsewhere.
After Mr Lewis left Microsoft in 2005,he tried to introduce the concept of reusable components to the construction industry, which seemed ripe for it. Designsare often used only once, most buildingsare not energye�cient, the industry produces a lot of waste, and many materialsare simply thrown away. All this amountsto around 30% of the cost of construction,according to a case study on Living PlanIT
by the Harvard Business School.But instead of selling products to con
struction companies, Mr Lewis ended upapplying his ideas to an entire city. Even before the �rst concrete is cast, PlanIT Valleyhas already been built�in a simulationprogram that also allows detailed planning of the construction. Much of the city,which is to cost about $10 billion, will relyon prefabricated parts; its foundation, forinstance, will be made of concrete blocksthat come with all the gear for smart infrastructures preinstalled. Eventually the entire city and its buildings will be run by an�urban operating system� that integratesall parts and combines them into all kindsof services, such as tra�c managementand better use of energy.
Living PlanIT has a clear idea of whowill live and work in its city: the employees of the companies that form its �ecosystem�, another concept taken from the software industry. The startup has enrolled anumber of partners, among them Cisco,Accenture and McLaren Electronic Systems, a sister company of the eponymousFormula One brand, which will providesensor technologies. The idea is that these�rms will operate research facilities inPlanIT Valley, jointly improve the concept,develop applications and build similarprojects elsewhere.
Such grand designs are possible onlywhen building a city from scratch. But entrepreneurs like Mr Lewis and the plannersof similar projects have other advantages.For a start they generally have governmentbacking. Portugal granted the PlanIT Valleyproject �potential national importance�status, which among other things meanscheap land and generous tax breaks.Songdo was launched by the South Koreangovernment. And new cities are free fromthe constraints of having to deal with anestablished population, old infrastructureand bureaucracy.
Something old, something newMaking an existing city smart is a di�erentproblem altogether, as demonstrated byAmsterdam, the Netherlands’ biggest city.Amsterdam Innovation Motor (AIM), apublicprivate joint venture created for thispurpose, is not intended to come up withsome master plan but to identify interesting �smart� projects, work with local �rmsand other stakeholders and �nd ways tomake projects worthwhile for everybody.
So far AIM has launched a dozen projects, ranging from installing smart metersin some households to connecting ships tothe electricity grid so that they no longerhave to use diesel generators whenberthed in the city’s port. The most ambi
8 A special report on smart systems The Economist November 6th 2010
2
1
tious e�ort so far is something called �Climate Street�, which aims to reduce the energy use of an entire shopping street.
Most existing cities, at least in the West,will go for such a stepbystep approach,predicts Carlo Ratti, an architect and engineer who heads the SENSEable City Lab atthe Massachusetts Institute of Technology(MIT). He and his colleagues have come upwith a number of smart urban projects oftheir own. In one, called �TrashTrack�, theyasked volunteers to attach small electronictracking devices to hundreds of pieces ofrubbish to see where they would end up inorder to improve waste logistics. More re
cently they introduced the �CopenhagenWheel�, a bicycle wheel whose red hubcan not only give the rider a boost but alsomeasure environmental conditions, suchas pollution and noise levels.
Yet many such projects will need a common platform that streamlines data gathering and supports all kinds of applications, says Mr Ratti. That would alsoenable him to realise his ultimate vision:turning the city into a �control system� thatmakes use of data from a variety ofsources, from mobile phones to smart meters and sensors in buildings. The datacould be mined to improve public tran
sport and security.So far Mr Ratti and his collaborators
have mainly used data from mobilephones for their projects. In �WikiCity�,implemented in Rome, such data allowedpeople to see visualisations of how theymoved through the Italian capital. However, a new mayor elected in 2008 provedmuch less interested in Mr Ratti’s projectsthan his environmentalist predecessor, sothe team has gone o� to Singapore.
But it will be not just governments, cities and utilities that will make the worldsmart. Private companies will also playtheir part, particularly startups. 7
CALL it the democratisation of sensors.Pachube (pronounced �patchbay�), a
startup based in London, o�ers a servicethat lets anybody make sensor data available to anyone else so they can use them tobuild smart services. One tinkerer has Pachube’s computers control the fan in his of�ce, guided by temperature readingsuploaded from a thermometer on his desk.
Such experiments are free, but thosewho develop more serious applicationsand do not want them to be available toanyone else have to pay. Usman Haque, Pachube’s boss, hopes that more and more�rms will do so as sensors multiply.
Pachube’s business model is one of themore interesting attempts to make moneyfrom the convergence of the physical andthe digital worlds, but there are plenty ofother �rms trying to cash in on smart systems. Many will fail, but those that succeedwill disrupt more than one industry andperhaps the economy as a whole.
But what is most exciting about smartsystems is the plethora of new servicesand business models that they will makepossible. �The internet of things will allowfor an explosion in the diversity of business models,� says Roger Roberts, a principal at McKinsey and one of the authors ofa recent study on the industry.
It is not just utilities that will bene�tfrom smart systems but other sectors too.The chemical industry, for instance, has already installed legions of sensors and actuators to increase its e�ciency. Others arejust starting. In the paper industry, according to the McKinsey study, one company
achieved a 5% increase in its production byautomatically adjusting the shape and intensity of the �ames that heat the kilns forthe lime used to coat paper. FoodLogiQ, astartup, allows food suppliers to tag andtrace their wares all along the supplychain�and consumers to check wherethey come from. Sparked, another startup,implants sensors in the ears of cattle,which lets farmers monitor their health,track their movements and �nd out a lot ofother things about their animals that theynever realised they wanted to know. Onaverage, each cow generates about 200megabytes of information a year.
Thanks to detailed digital maps, maintaining such things as roads and equipment will also become much more e�cient. As�nag, an Austrian �rm, used
aeroplanes equipped with special cameras to map the country’s highways. Its employees can now �y over them digitallyand even see what is underground. SanFrancisco’s Public Utilities Commissionknows the exact coordinates of everywastewater pump, its maintenance history and the likelihood of it failing. �Firmscan now send out maintainance crews before things actually break,� says SteveMills, who heads IBM’s software business.�Making the old stu� run better will be themost important bene�t of such systems inthe short run.�
Moreover, smart systems make newforms of outsourcing possible, some of itto unexpected places. Paci�c Control is notexactly a household name, but the company, based in Dubai, claims (with somejustice) that it is the �world leader in automation solutions�. Its global commandcentre remotely monitors buildings, airports and hotels, keeping an eye on suchthings as energy use, security and equipment. For the moment most of the �rm’scustomers are in Dubai itself, but it should�nd more than a few abroad.
Signi�cantly, once devices are connected and their use can be metered, there is nolonger any need to buy them. Alreadysome makers of expensive and complexequipment no longer sell their wares butcharge for their use. RollsRoyce, for instance, which makes pricey aircraft engines, rents them out to airlines, billingthem for the time that they run. Makers ofbloodtesting equipment have taken tocharging only if the device actually pro
Augmented business
Smart systems will disrupt lots of industries, and perhaps the entire economy
The Economist November 6th 2010 A special report on smart systems 9
2
1
duces usable data. And Joy Mining Machinery, a maker of mining equipment,charges for support by the tonne.
Some �rms are using metering in innovative ways. Zipcar and other carsharing�rms, for instance, put wireless deviceswith sensors into their vehicles so that customers can hire them by the hour. And insurance �rms, among them Progressive inAmerica and Coverbox in Britain, ask customers to install equipment in their carsthat can measure for how long, how fastand even where a car is driven. Premiumscan then be based on individual drivers’behaviour rather than on such proxies asage and sex.
Michael Chui, a coauthor of McKinsey’s report on the internet of things, saysthat such applications will allow companies to �have a much more dynamic interaction with customers�. In Japan, for instance, vending machines can recognise acustomer’s age and sex and change themessage they display accordingly.
The more data that �rms collect in theircore business, the more they are able to offer new types of services. By continuouslyassessing the performance of its jet engines around the world, RollsRoyce is ableto predict when engines become morelikely to fail so that customers can scheduleengine changes. Heidelberger Druckma
schinen, whose huge presses come withmore than 1,000 sensors, has started o�ering services based on the data it collects, including a website that allows customers tocompare their productivity with others.�Many companies will suddenly discoverthat their main business is data,� says PaulSa�o, a Silicon Valley technology forecaster who wrote a widely noted essay on theimpact of ubiquitous sensors back in 1997.
HewlettPackard is a prime candidatefor such a Sa�o moment�if its plans toscatter millions of sensors around theworld come to fruition. It is doing this to increase demand for its hardware, but it alsohopes to o�er services based on networks
CONTRARY to what might be expected, the industry that will see the least
change is information technology. Formost IT �rms smart systems simply meanmore business. This will drive a new waveof technology investment, predicts Andrew Bartels of Forrester Research. By2017, he says, �smart computing technologies� will represent about half thespending on IT equipment and softwarein America (see chart 4).
Clearly some parts of the IT industrystand to bene�t more than others. HarborResearch estimates that internetenableddevices alone will net more than $10 billion worldwide in 2014, compared with$4.3 billion in 2009. And wireless sensorsare growing exponentially. Last year amere 10m radio chips for such sensorswere sold, according to ABI Research. If ithas its calculations right, that total willrise to 645m by 2015.
Since all the data gathered by sensorshave to be kept somewhere, storage is hot,too. That explains the recent bidding warbetween Dell and HP over 3Par, a datastorage company. HP won by o�ering$2.35 billion for a �rm with only 650 employees and $194m in revenues. Forecastsfrom IDC, another marketresearch �rm,throw light on why HP is willing to pay somuch. IDC expects the capacity shippedto increase by 50% plus this year.
Vendors of programs that sift throughdata are also likely to do well. In recentyears IBM has invested billions of dollars
in buying �rms that make business analytics software. In September, for instance,it gave $1.7 billion for Netezza, a vendor ofdata warehouses (specialised computersystems that quickly crunch through hugeamounts of data). It is the biggest bet IBM
has made in any area, says Ambuj Goyal,the �rm’s global head of developmentand manufacturing. If it comes good, IBM
should be richly rewarded. The globalmarket for analytics programs, for instance, will grow from $25.5 billion thisyear to $34 billion in 2014, according toIDC.
This technology wave will almost certainly produce new software championstoo. SAP came of age when mainframes
were dethroned by smaller computer systems that allowed companies to streamline many more of their business processes. Similarly, experts predict that newprograms will be needed to manage a�rm’s interactions with the physicalworld. Already, there is a plethora of startups o�ering to help businesses cut greenhouse gases.
Platforms to integrate the data streamsfrom all kinds of sensors are another newmarket. Examples include Pachube, astartup, and Palantiri Systems, best described as a �Facebook for sensors�. Theservice allows devices to have their own�page� on corporate social networks, sotheir readings are shown in the form ofnewsfeeds and people can ask questionsabout them. �Devices thus become part ofthe conversation,� says John Canosa, Palantiri’s boss.
As usual, most money will be madefrom IT services, in particular to set upsmart systems in cities. China alone willneed to add the equivalent of one NewYork every year to accommodate thenumber of people expected to migrate tourban areas by 2025, according to McKinsey. Big IT �rms are chasing all over thecountry to get a piece of this huge pie. Cisco is already helping to build a Chinesereplica of Songdo, the smartcity project inSouth Korea. IBM, for its part, hopes tohelp develop an entire network of smartcities and has started to collaborate withseveral of them.
Who will clean up?The IT paydirt4Into the unknown
Source: Forrester Research *Estimate
US government and business spendingon IT equipment and software, $bn
0
200
400
600
800
2008*09* 10 11 12 13 14 15 16 17
Current-generationtechnology
Smart-computingplatform technology
Smart-computingprocess applications
Smart-computingindustry solutions
F O R E C A S T
10 A special report on smart systems The Economist November 6th 2010
2
1
of sensors. For instance, a few thousand ofthem would make it possible to assess thestate of health of the Golden Gate bridge inSan Francisco, says Stanley Williams, wholeads the development of the sensors atHP. �Eventually�, he predicts, �everythingwill become a service.�
Apple, though it prides itself on its fancy hardware, is already well on its way towards transforming itself into a serviceand data business thanks to the success ofits iPhone. When the computermakerlaunched the device in June 2007, it did notexpect �apps�, the applications that run onsmartphones, to become such a big deal.The �App Store�, where users can download these pieces of software, waslaunched a year after the �rst iPhone wasshipped. But the App Store now sells morethan 250,000 apps that have been down
loaded over 6.5 billion times. And with itsnew platform for mobile advertising,called iAd, Apple has started to make money from all the data it collects.
Some �rms will make a living based entirely on mining �data exhaust�, the bitsand bytes produced by other activities.One example is Google’s PowerMeter,which not only lets users check their use ofelectricity online but gives Google accessto lots of data to analyse and, not least, selladvertisements against.
Conventional services, too, can be metered. Those supplied by governmentsmay not be the �rst to spring to mind but,as a study by the Dutch economics ministry asks, why not use sensors for taxingthings like pollution? That might be uncontroversial, but analytics software couldalso be put to more manipulative uses by
�netuning charges for public goods to getcitizens to behave in certain ways.
Much of the innovation in this �eldmay come not from incumbents but fromnewcomers, and it may happen fastest onsuch platforms as Pachube. In a way it is across of YouTube and Windows. Whatmade the videosharing site so popularwas the way it converted all videos to acommon format. Pachube is doing thesame for data feeds from sensors. And likeMicrosoft’s operating system for applications, it provides basic features for smartservices, such as alerts, data storage and visualisation tools.
This spring at Where 2.0, a technologyconference in Silicon Valley, the star of theshow was Skyhook Wireless, a �rm that offers geographicallocation information asa service. It recently launched a new o�er
�MIRROR WORLDS� is only one ofDavid Gelernter’s big ideas. An
other is �lifestreams��in essence, vastelectronic diaries. �Every document youcreate and every document other peoplesend you is stored in your lifestream,� hewrote in the mid1990s together with EricFreeman, who produced a doctoral thesison the topic. Putting electronic documentsin chronological order, they said, wouldmake it easier for people to manage alltheir digital output and experiences.
Lifestreams have not yet replaced thedesktop on personal computers, as Mr Gelernter had hoped. Indeed, a softwarestartup to implement the idea folded in2004. But today something quite similarcan be found all over the web in many different forms. Blogs are essentially electronic diaries. Personal newsfeeds are atthe heart of Facebook and other socialnetworks. A torrent of short text messagesappears on Twitter.
Certain individuals are going even further than Mr Gelernter expected. Someare digitising their entire o�ce, includingpictures, bills and correspondence. Others record their whole life. Gordon Bell, aresearcher at Microsoft, puts everythinghe has accumulated, written, photographed and presented in his �local cyberspace�. Yet others �log� every aspect of
their lives with wearable cameras.The latest trend is �lifetracking�. Prac
titioners keep meticulous digital recordsof things they do: how much co�ee theydrink, how much work they do each day,what books they are reading, and so on.Much of this is done manually by puttingthe data into a PC or, increasingly, a smartphone. But people are also using sensors,mainly to keep track of their vital signs, forinstance to see how well they sleep orhow fast they run.
The �rst selftrackers were mostlyübergeeks fascinated by numbers. But themore recent converts simply want to learn
more about themselves, says Gary Wolf, atechnology writer and cofounder of ablog called �The Quanti�ed Self�. Theywant to use technology to help them identify factors that make them depressed,keep them from sleeping or a�ect theircognitive performance. One selftrackerlearned, for instance, that eating a lot ofbutter allowed him to solve arithmeticproblems faster.
A market for selftracking devices is already emerging. Fitbit and Greengoose,two startups, are selling wireless accelerometers that can track a user’s physical activity. Zeo, another startup, has developed an alarm clock that comes with aheadband to measure people’s brainwaveactivity at night and chart their sleep onthe web.
As people create more such selftracking data, �rms will start to mine them ando�er services based on the result. Xobni,for example, analyses people’s inboxes(�xobni� spelled backwards) to help themmanage their email and contacts. It liststhem according to the intensity of theelectronic relationship rather than in alphabetical order. Users are sometimessurprised by the results, says Je� Bonforte,the �rm’s boss: �They think it’s creepywhen we list other people before theirgirlfriend or wife.�
Tracking your life on the webYour own private matrix
The Economist November 6th 2010 A special report on smart systems 11
ing called SpotRank. Drawing on all thedata it has collected in recent years fromapps using its services, the �rm can predictthe density of people in speci�c urban areas�anywhere, any day and at any hour.�This will give us great insights into humanbehaviour,� says Brady Forrest, the chairman of the conference.
Another hit at the conference was a service called Wikitude. Its �World Browser�,a smartphone app, checks the device’s location as well as the direction in which itscamera is pointing and then overlays virtual sticky notes other users have left aboutthings like local landmarks. So far Wikitude and similar services are mostly usedas travel guides. But in principle userscould collaboratively annotate the entirephysical world�and even other people.TAT, another startup, already experiments with something called �AugmentedID�, which uses facial recognition to display information about a person shownon a smartphone’s screen.
It is di�cult to say what e�ect all thiswill have on business and the economy.But three trends stand out. First, sincesmart systems provide better information,
they should lead to improved pricing andallocation of resources. Second, the integration of the virtual and the real willspeed up the shift from physical goods toservices that has been going on for sometime. This also means that more and morethings will be hired instead of bought.Third, economic value, having migratedfrom goods to services, will now increasingly move to data and the algorithmsused to analyse them. In fact, data, and theknowledge extracted from them, may even
be on their way to becoming a factor ofproduction in their own right, just likeland, labour and capital. That will makecompanies and governments increasinglyprotective of their data assets.
In short, we may be moving towards a�Weightless World�, the title of a 1997 bookby Diane Coyle about a future in whichbytes are the only currency and the thingsthat shape our lives have literally noweight. But for now, gravity has not quitebeen repealed yet. 7
2
1
THERE is not much to see in the city ofBakers�eld, north of Los Angeles, but
recent events have put it on the global electricityindustry map. As in many otherCalifornian communities, Paci�c Gas &Electric (PG&E), the local power utility, hadinstalled smart meters in most households. Soon afterwards customers startedcomplaining about rocketing power bills.For some people they almost trebled. Predictably, this caused a political storm. Localpoliticians and consumer groups jumpedon the issue. Enterprising lawyerslaunched a classaction suit. PG&E admitted that some of its meters had technicalproblems, but the higher bills were clearlydue to a combination of exceptionally hotweather, increased charges and changes inthe rate structure.
An independent auditor found nothingwrong with the smart meters, and California’s regulators did not stop PG&E from installing more of them. But utilities and regulators elsewhere, spooked by theincident, have become much more careful
before embracing the technology. �Bakers�eld is likely to slow down the installationof smart meters�not just in the UnitedStates but worldwide,� says Ahmad Faruqui of the Brattle Group, a consultancy. Bakers�eld also demonstrates that a smarterworld will meet with resistance. The reasons are part technical, part institutional.
A list as long as your armTechnology is a good place to start. Sensorsare getting ever cheaper, but for many applications they are still much too expensive. HP, for instance, likes to point out thatits supersensitive accelerometers aremade in the same factories as its printercartridges. But the �rm’s sensors are stilltoo pricey to use them for anything buthighvalue applications, such as oil exploration. RFID tags are a cautionary tale.They were supposed to revolutionise retailing but the readers, software and othergear needed to make them useful is still notcheap enough to be universally adopted.
Equally important, standards for such
things as smart meters need to be sortedout. Setting them too early would hamperinnovation, but in their absence utilitieswill hold back from investing, worried thatthey might bet on the wrong technology.Standards could also become a weapon ofindustrial policy, in particular in countriesthat see clean technology as an engine ofgrowth. When China’s State Grid Corporation, which operates most of the country’spower network, announced its smartgridplans in June, it also released the standardsit intends to implement. Some say this wasa move to protect Chinese �rms.
The internetaddress system is a worryas well. For a computer or any other deviceto be part of the internet, it needs a uniqueidenti�er�currently a long number calledan internetprotocol (IP) address. Becauseof the network’s rapid growth in recentyears, these numbers could run out as early as the middle of next year. If the IT industry keeps dragging its feet on moving toIPv6, a new address system that uses manymore numbers, the growth of the internet
Sensors and sensibilities
A smarter world faces many hurdles
12 A special report on smart systems The Economist November 6th 2010
2
1
of things will be stymied. Space is also bound to get tight in the
ether. A few wireless sensors and devicesdo not make a di�erence, but as their numbers grow they will need an ever biggerchunk of the available radio spectrum. Thenumber of wireless subscriptions has nowreached 5 billion worldwide, earlier thanexpected, not least because so many SIM
cards now sit in machines that communicate with other machines.
And then there are security concerns,particularly after the Stuxnet worm madethe rounds in September. The maliciousprogram quickly found its way into computers controlling industrial processes theworld over, demonstrating how vulnerable control systems are to such attacks.But even before Stuxnet struck, securityconsultants had shown how large numbers of smart power meters could behacked and shut down.
Turf, ego and powerYet all these technical issues pale by comparison with the institutional barriers. Fora city to o�er smart services and save money, its departments have to work closely together, share their data and use a commonIT infrastructure. London, for instance, hasdi�erent payment systems for public transport, bicycle hire and toll roads. Such fragmentation is costly and makes it more di�cult to come up with new o�ers (say,reducing the congestion charge for thosewho often hire a bicycle). But getting acity’s islands of bureaucracy to work together tends to be di�cult, says Mark Cleverley of IBM, who helps governments andcities develop plans for smart systems.
The problem is not just that departments often jealously protect their data,something experts call TEP, as in �turf, egoand power�. O�cials also lack a commonlanguage or generally agreed criteria for asmart city�which is a big issue, too, for themany companies that are usually involvedin a project. �It’s hard to build a businesscase if people don’t understand each other,� says Simon Giles, in charge of strategyfor smart technologies at Accenture.
Things are easier in Singapore. Ministries and agencies compete for reputationand resources, but they also cooperateclosely on implementing master planssuch as �A Lively and Liveable Singapore:Strategies for Sustainable Growth�, thecitystate’s roadmap to becoming smart.That helps to explain why Singapore willprobably be the �rst city to combine its various smart systems into a single one.
More generally, Asian countries seem
to have an advantage in building smart systems because governments are often lessdemocratic and administrations morehierarchical. China’s State Grid Corporation intends to have its smart grid fullybuilt by 2020. The country’s governmenthas also made the implementation of IPv6a central part of its �veyear plan to buildthe �China Next Generation Internet�. Itshowed o� its e�orts at the 2008 Olympicgames in Beijing, where everything thatwas connected�cameras, taxis, securitysystems�used IPv6.
In the West it will often take a crisis toget there. When Thames Water in 2006failed to meet targets set by the regulator toreduce leaks and was subsequently sold,the new leadership went on to organisethings di�erently. Today, at the utility’s operating centre in Reading, the workers whomonitor the network, take calls from customers and schedule work crews all sit inone openplan o�ce, allowing them tocommunicate much more easily across departmental boundaries.
Similarly, when Bill Ritter became Colorado’s governor in 2007, he made the consolidation of the state’s chaotic IT systemsa priority and named a state chief information o�cer who is also a member of hiscabinet. Since then Colorado has madegreat progress in achieving one prerequisite for becoming a smart state: a commonIT infrastructure capable of delivering newservices.
Amsterdam, being the capital of a highly pluralistic country, had to take a di�erentapproach. Instead of relying on the muni
cipal administration to become a smartcity, it created Amsterdam Innovation Motor (AIM), a publicprivate joint venture incharge of coming up with projects and mediating between the parties involved. �Being a translator and making sure that a project is worthwhile for everybody are ourmain jobs,� explains Ger Baron, AIM’s project manager.
If these three examples are any guide,smart systems may well change the waythat local governments, in particular, areorganised. Instead of being a collection ofdepartmental silos, they could come to resemble computing platforms. Most services, from payment systems to tra�c information, would be provided in only oneversion and used by all departments�orby private �rms that want to o�er theirown urban applications.
Some cities, such as London, San Francisco and Washington, DC, have alreadyopened their data vaults. IBM, amongmany other companies, has already built awebbased application called City Forward that takes in data from 50 cities. Yetothers are not that generous with theirdata, which is the third barrier. More openness should be good for innovation, notjust in terms of the information itself buthow it is handled. But �rms with lots ofdata�be they power utilities or makers ofmedical equipment�will be loth to givethem up, says Glen Allmendinger, president of Harbor Research.
At the same time, Mr Allmendinger predicts, some �rms will be forced to give upcontrol of their data. Hospitals, for instance, will hardly put up with dozens ofdashboards that monitor the activities ofdi�erent types of equipment: they willwant a uni�ed view. And some clever startups and IT �rms will �nd ways arounddata monopolies. Nobody can stop consumers from giving data about their powerusage to nonutilities, for instance. Someprivate meters, attached to a sensorclamped around the main power line, cannow send the data they have gathered towebbased energymonitoring services,such as Google’s PowerMeter and Microsoft’s Hohm.
Data are a problem for governmentstoo. Li Yizhong, China’s minister of industry and information technology, has expressed concerns about IBM’s SmarterPlanet initiative. �The US tries to use its information network technology for thingsas small as controlling one computer orone generator and as large as controlling awhole industry, to control every country’seconomy,� he is reported to have said. �We
The Economist November 6th 2010 A special report on smart systems 13
2
1
must be enlightened and vigorously develop strategic emerging industries, but alsomust raise our vigilance and cannot fall under the control of anyone.�
China’s concerns with IBM’s SmarterPlanet (called �wise Earth� in Mandarin)point to the fourth set of barriers: government regulation�or the lack thereof. Privacy legislation tops the list. New laws willmultiply in response to an increasinglysmart world. Germany’s governmentplans to strengthen its rules to deal withGoogle’s Street View, an online service thatlets users pan through photos of streets.Among other things, the new bill is likelyto enshrine in law what Google has already agreed to do under pressure fromdataprotection o�cials: giving people theoption to have their house blurred onStreet View. Nearly a quarter of a millionhave done so.
The regulation of smart grids is a murkier area. Some countries push utilities to reduce energy demand, but outofdate rulesencourage them to sell more, says Accenture’s Mr Giles. Elsewhere, illconceivedderegulation of the power market is holding things up. Since the grid now has morethan one owner, it is often hard to knowwho will bear the risks and who will garner the rewards.
A host of new legal questions will alsohave to be answered. Who is liable if anautonomous system, such as an autopilotthat governs the movement of cars on amotorway, causes an accident? What if asingle company manages to dominate datafrom certain types of sensors or control theinformation about where sensors can befound and which of them are active? Andcan private surveillance still be restrictedwhen cameras and other sensors make itubiquitous? �The sensor revolution will
challenge hidden assumptions in a bewildering array of doctrinal �elds,� writes Kevin Werbach, of the Wharton School of theUniversity of Pennsylvania, in a paper entitled �Sensors and Sensibilities�.
Lastly, consumers may not play ball.PG&E’s woes in Bakers�eld are not theonly example. In Boulder, Colorado, theshowcase for smart grids, customers ofXcel Energy, the local utility, are becomingincreasingly cross because they will haveto pay much of the project’s costs. In theNetherlands the government backeddown from making smart meters mandatory because of concerns that the data collected could be used to see whether properties are empty or expensive new gadgetshave been purchased. Consumers also dislike feeling that they are being squeezeddry, particularly in America, which makesdynamic pricing hard to bring in. Even simple rate plans where the price of electricitydepends on the time of day have had to beabandoned after customer protests.
Nor is such resistance limited to smartmeters. The smooth introduction of Stockholm’s toll system was the exception rather than the rule. In many countries politicians do not even try. Germany, forinstance, charges lorries for using its motorways, but only a suicidal governmentwould attempt to extend the system to carsin a country where even buckling up waslong opposed by motoring clubs as interfering with drivers’ freedom.
Make it attractiveIt is odd, then, that everybody loves mobile devices, which are not that di�erentfrom smart meters or onboard units. Inparticular, smartphones and the applications that run on them generally keep aclose watch on what users do. Even so,
nearly 270m of these devices will be soldthis year, 55% more than in 2009, says IDC.
The di�erence is that Apple and othersmartphone makers have made it theirbusiness to �nd out what consumerswant�traditionally not the forte of utilitiesand government agencies. For example, ittook that �urry of protests to prompt PG&E
to open a dedicated call centre for questions about smart meters. Yet communicating with customers should be one of the�rst things for �rms to do when introducing smart meters, says Mr Giles.
Another lesson is that utilities shouldnot try to achieve too many things at once,says Mr Giles, who recently surveyedsmartgrid projects worldwide for a studyon how to speed up their introduction.Some utilities brought in smart meters andnew pricing schemes at the same time,thus overwhelming consumers and obscuring the reasons for higher bills.
To avoid such problems, Amsterdam istrying �cocreation�, in the words of AIM’sMr Baron. �We did not just put in smart meters and ask consumers to pay for them,�he says. Instead 500 households in the district of Geuzenveld were invited to makesuggestions on how to save energy andmonitor consumption.
Yet utilities may have to resort to socialengineering to get customers more engaged. Opower, a startup, lets them seenot only their own consumption �guresbut numbers for their neighbours too (anonymised to preserve privacy), and o�ersthem tips on saving power. Peer pressure,the company claims, persuades people toreach for the switches much more often.Given enough time, all these barriers tobuilding smart systems can probably beovercome. But how smart does the worldreally want to be? 7
�IT IS not possible to make a lastingcompromise between technology
and freedom, because technology is by farthe more powerful social force and continually encroaches on freedom through repeated compromises.� Thus wrote the Unabomber, also known as Ted Kaczynski, inhis manifesto, published in 1995 by theNew York Times and the Washington Post inthe hope that he might end his terror cam
paign or somebody might recognise hisstyle of writing and unmask him.
Mr Kaczynski’s methods were abhorrent. His bombs killed three people and injured 23 over nearly 20 years. He was arrested in 1996 and is currently serving a lifesentence. But his concern that technologywill slowly but surely undermine humanfreedom is shared by quite a few mainstream thinkers. As this special report has
argued, smart systems will improve e�ciency and could help solve many environmental problems, in particular globalwarming. Yet if those systems seriouslyimpinge on people’s freedom, many people will balk. The protests against smartmeters in Bakers�eld and elsewhere maybe only the start.
Smart systems are rekindling old fears.Top of the list are loss of privacy and gov
Horror worlds
Concerns about smart systems are justi�ed and must be dealt with
14 A special report on smart systems The Economist November 6th 2010
2
Previous special reports and a list offorthcoming ones can be found online
Economist.com/specialreports
Future special reportsJapan November 20th 2010Security in Asia December 4th 2010Global leaders January 22nd 2011Food February 26th 2011Property March 5th 2011
Economist.com/rights
O�er to readersReprints of this special report are available fromthe Rights and Syndication Department. A minimum order of �ve copies is required.
Corporate o�erCustomisation options on corporate orders of 100or more are available. Please contact us to discussyour requirements.
For more information on how to order specialreports, reprints or any queries you may haveplease contact:
The Rights and Syndication Department26 Red Lion SquareLondon WC1R 4HQ
Tel +44 (0)20 7576 8148Fax +44 (0)20 7576 8492email: [email protected]
ernment surveillance. Internet users haveonly recently begun to realise that everysingle thing they do online leaves a digitaltrace. With smart systems the same thingwill increasingly apply to the o�ine world;Google’s Street View is only the beginning.
Even the champions of a smarter planet admit as much. �Some citizens have expressed discomfort at living in not a safersociety but a ‘surveillance society’,� saidSam Palmisano, the boss of IBM, in aspeech earlier this year. He cited a newspaper article recounting that there are now 32closedcircuit cameras within 200 yards ofthe London �at in which George Orwellwrote his book �1984�.
Hopes and fearsMr Palmisano would be in the wrong job,however, had he not gone on to say thatsuch concerns have to be rethought and tostress the economic and social bene�ts ofsmart systems. Others point out less obvious advantages. �All this technology actually strengthens the human side of cities,�says Carlo Ratti, director of MIT’s SENSEable City Lab. People who are always connected, he argues, can work wherever theylike. And Mr Haque, the boss of Pachube,claims that �sensors empower people because measuring the environment allowsthem to make decisions in real time.�
On the other hand smart systems arealso undeniably useful as an instrument ofcontrol. Singapore has made an impressivejob of smartening up its physical infrastructure, but its network of security cameras could also be used for enforcing rulesmore objectionable than a ban on chewinggum. Similarly, the operations centres anddashboards for local governments in China being built by Cisco, IBM and others begthe question whether their only purpose isto make these cities smarter.
Other deep fears brought on by smartsystems is that machines could be hacked,
spin out of control and even take over theworld, as they did in the �lm �The Matrix�.As the Stuxnet worm and the May ��ashcrash� on Wall Street have shown, the �rsttwo are already real possibilities, even ifthe third still seems somewhat remote�although wellknown computer scientists,arti�cialintelligence researchers and roboticists met in California a couple ofyears ago to discuss that risk.
And there is a more subtle danger too:that people will come to rely too much onsmart systems. Because humans cannotcope with the huge amounts of data produced by machines, the machines themselves will increasingly make the decisions, cautions Frank Schirrmacher of theGerman daily Frankfurter Allgemeine Zeitung in his recent book �Payback�. Similarly, Nicholas Carr, an American commentator on the digital revolution, in his book�The Shallows� claims that the internet,the mother of all smart systems, is on itsway to smothering creativity and profound thinking.
A further worry is that smart technology will ultimately lead to greater inequality�and not just because it could create an�information priesthood�, in the words ofMr Gelernter. Paul Sa�o, a noted SiliconValley technology forecaster, expects ubiquitous sensors to give a huge boost to productivity�at the expense of human monitors. �We are likely to see more joblessrecoveries,� he says.
Whether computers will indeed start toeliminate more jobs than they create remains to be seen. But smart systems certainly represent a conceptual change. Sofar IT has been used to automate and optimise processes within �rms and other organisations as well as the dealings between them. Now it will increasingly beused to automate and optimise interactions with the physical environment.
Some of the concerns raised will behard to deal with. For instance, therewould be little point in passing laws thatwould give individuals the right to decidewhether their data can be used by smartsystems if cameras and other sensors arealready ubiquitous. And building in circuitbreakers to keep automation from going too far could defeat the purpose ofsmart systems and sti�e innovation.
Still, technological progress is not someforce of nature that cannot be guided. �Wecan and we should exercise control�bydemocratic consensus,� says Mr Gelernter.Yet for a consensus to be reached, theremust be openness. The biggest risk is thatsmart systems become black boxes, closedeven to citizens who have the skills to understand them. Smart systems will makethe world more transparent only if theythemselves are transparent. 7