Computer
COVER FEATURE OUTLOOK
qM qMqM
THE WORLD’S NEWSSTAND®
Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page qMqM Qmags
Enabling the Internet of Things
Roy Want, Bill N. Schilit, and Scott Jenson, Google
Merging the virtual World Wide Web with nearby physical devices that are part of the Internet of Things gives anyone with a mobile device and the appropriate authorization the power to monitor or control anything.
28
The Internet of Things (IoT) paradigm enables interconnectedness among devices—anytime, anywhere on the planet—providing the Inter- net’s advantages in all aspects of daily life. An-
alysts predict that the IoT will comprise up to 26 billion interconnected devices by 2020, a 30-fold increase from 2009(w__w_w_.g_a_r_tn_e_r._co_m__/n_e_w_s_ro_o_m_/_id_/_2_63_6_0_7_3).
The conventional Internet has proved valuable in al- most all endeavors by giving people the ability to inter- act with global information and services. The majority of this interaction happens through the World Wide Web, with client computers running a browser and communi- cating with cloud-based servers. However, the Internet is not limited to the Web: a wide diversity of other pro- tocols are employed to make use of global Internet con- nectivity. The IoT is considered to be the next logical evo- lution, providing extensive services in manufacturing, smart grids, security, healthcare, automotive engineer- ing, education, and consumer electronics. Many of these systems already have a Web presence but use protocols that are largely Web independent.
Practical issues with the IoT vision must be addressed,
COMPUTER PUBLISHED BY THE IEEE COMPUTER SOCIETY
including how to handle dramatic increases in network scale and how to determine device proximity, sometimes referred to as localized scalability.1 In an IoT world, pref- erentially discovering things nearby and letting users interact with them is a powerful mechanism for over- coming a global network’s scale and complexity. Other important IoT enablers are peer-to-peer connections, low-latency real-time interaction, and integration of de- vices that have little or no processing capability.
THE IOT VISION
The Web provides an important interaction model for the IoT by letting users get device-related information and in some cases control their devices through the ubiquitous Web browser. The conventional Web is a convenience we enjoy as we search for information, respond to email, shop, and engage in social networking; the IoT would expand these capabilities to include interactions with a wide spectrum of appliances and electronic devices that are already ubiquitous in the early 21st century.2 We refer to devices that are part of the IoT and directly accessed, monitored, or controlled by Web technologies
0018-9162/15/$31.00 © 2015 IEEE
Computer
qM qMqM
Qmags
Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page qMqM
THE WORLD’S NEWSSTAND®
Computer
qM qMqM
THE WORLD’S NEWSSTAND®
Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page qMqM Qmags
as the Physical Web: Physical Web = Web technology + IoT.
Cloud service
(for example, fitbit.com)
Bluetooth
Internet cloud
Proxy Web service (for example, movieposters.com)
Reference: “Buy Movie Tickets”
URI
NFC tag
Bluetooth LE tag
Identifiers are the key to enabling any kind of interac-
tion among devices. From an
IoT perspective, IPv6’s 128-bit addresses serve as identifiers
for a global network of de-
vices. Alternatively, Uniform Resource Identifiers (URIs),
which include both locators
and names, provide a higher-
level concept that bridges
those devices to existing Web technology. The Uniform Re-
source Locator (URL) is used
in conjunction with a Distrib-
uted Name Service (DNS) to route and connect to services. Uniform Resource Names (URNs), such as globally unique IDs, are resolved by scheme-specific methods. A distinguishing aspect of the Physical Web is to consider URIs as the primary identifier.3
Many researchers and practitioners in this field, including the authors of this article, expand the IoT definition to include enabling an Internet presence for any person, place, or thing on the planet, thereby pushing our notion of the Physical Web beyond smart devices. Clearly, an Internet presence cannot occur without processing and network- ing, so instead of providing them di- rectly, an Internet service can provide information and perform actions via other nearby devices serving as a gate- way to that proxy service.4,5
Gateway devices will enable billions of people, places, and things to partic- ipate in the IoT—most people today already carry one. The smartphone, the most popular computing device of all time, with more than 1 billion users (www.idc.com/prodserv/smartphone -_o_s-_m_a_rk_e_t-_s_ha_r_e._js_p), is well equipped to serve as this pervasive portal.
(a)
(b)
FIGURE 1. Two methods for a smartphone to interact with the Internet of Things: (a) direct and
(b) proxy. Networked computers can participate in the IoT—passive objects can, too. LE, low energy; NFC, near-field communication; URI, Uniform Resource Identifier.
Client interaction with services
Figure 1 shows the two distinct inter- action modes that smartphones can enable in the IoT. Through direct in- teraction, a smartphone can query the state of an IoT device in its proximity and then provide a bridge between low- level peer-to-peer protocols, such as Bluetooth or Wi-Fi, and Internet proto- cols, such as HTTP and TCP. One exam- ple is the Fitbit fitness monitor, which uploads a user’s step count through his or her phone over a 4G network to the user’s account in the cloud. Through proxy interaction, mobile users who happen to be near an IoT-enabled ob- ject or device can look up associated information published by interested parties through a Web service using their smartphone, just as they would when performing a Web search. One example is a movie poster that enables nearby people to automatically access a webpage on their smartphone and buy electronic tickets online.
DOES THE IOT
ALREADY EXIST?
The IoT is a popular buzzword in the computing industry; it appears in the marketing campaigns of
major networking companies such as Cisco and microprocessor gi- ants such as Intel. It even serves as the title or theme of conferences, such as the “Internet of Things” World Forum (http://iotinternetof t_h_in_ g_s_c_on_f_e_re_n_ce_._co_m_ ).
However, the phrase represents ideas that have existed since the be- ginning of the Web or been written about in whitepapers from well-known research laboratories such as (Xerox) PARC and HP Labs. So why isn’t the IoT a standard part of the way we do busi- ness today? Why is it still the subject of speculation and vision statements in keynote addresses at well-known computer industry events such as the annual Consumer Electronics Show?
The answer appears to be that the IoT exists for a small number of tech- nologies that have the ingredients for a successful business case. In general, these early systems have tended to be closed ecosystems, using private APIs and locking up the data. This is counter to the spirit of open systems at the heart of the original Internet standards, reflecting instead the more recent commercial successes of
JANUARY 2015 29
qM qMqM
Qmags
Computer
Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page qMqM
THE WORLD’S NEWSSTAND®
Computer OUTLOOK
qM qMqM
THE WORLD’S NEWSSTAND®
Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page qMqM Qmags
(a) (b) (c)
FIGURE 2. Various forms of electronic tags support the Physical Web (all about the size of a quarter): (a) a near-field communication (NFC) tag; (b) a quick response (QR) code; and (c) a Bluetooth low energy (BLE) tag. However, it is not clear which technology—each with its own affordances and problems—will become the primary IoT enabler.
proprietary business entities such as Apple’s App Store and Facebook. You can actually buy home automation systems that connect to the Internet through your home’s Wi-Fi. These sys- tems are usually built with a bridge that controls the automation compo- nents through proprietary protocols on one side and communicates with open protocols to a proprietary Web service on the other. Users can then employ desktop computers or smart- phones as a client to control their home by interacting with the Internet service, effectively providing user in- terface hardware at no cost to the IoT device manufacturer.
A significant hurdle to fully re- alizing the IoT relates to scale— specifically, expanding the Internet to IoT scale means that the address space for the Internet will need to increase by several orders of magnitude. There- fore, another requirement for support- ing the IoT is a larger device address space than that provided by IPv4. To enable this kind of expansion, the In- ternet Engineering Task Force (IETF) has been working on the IPv6 stan- dard for some time. When the tran- sition is complete, the address space will be large enough to support every object on the planet, enabling embed- ded computers of all sizes to be easily integrated into the Internet. However, a large percentage of the objects in the IoT will not be suitable for direct
wired or wireless connection to the Internet, falling into the class of pas- sive devices. For these objects, a tag, smartphone, and proxy Web service is needed to provide users with the ob- ject’s Web presence. Of all the vision- ary ideas around the IoT, this one has made the least progress to date.
ENABLING TECHNOLOGIES FOR THE IOT
As Figure 2 indicates, tagging an ob- ject to reference a proxy Web service can be achieved through a variety of technologies, but the early primary contenders have had issues that hin- dered their adoption.
RFID and near-field
communication
In the early 2000s, RFID was considered
one of the most likely technologies to
accelerate the formation of the IoT.6 A
new UHF RFID tag standard was devel-
oped by EPC Global (http://epcglobal
.org), with a goal of further automat-
that interfered with the wireless iden- tification process.
The latest opportunity for RFID technology is in the form of near-field communication (NFC) as a support for electronic payments. Although only a small number of smartphone products include NFC transceivers, the potential for this capability to propagate to all fu- ture smartphones is high. It would also enable phones to read passive NFC tags that can store a URI, while still being cheap, small, thin, and attachable to al- most anything. In September 2014, Ap- ple announced the iPhone 6 would in- clude NFC support for ApplePay. With Apple’s significant smartphone market share in the US, this move could influ- ence other handset manufacturers to follow suit, pushing NFC into becom- ing a key IoT enabler.
Optical tags and
quick response codes
Another contender for low-cost tag- ging is the optical or printed tag—in particular, the most popular 2D op- tical standard, the quick response (QR) code.7 The success of the QR code standard is directly related to the ubiq- uity of its reader, an application of the high-resolution camera found in all modern smartphones. A QR code is ex- tracted and decoded from a scene using image-processing techniques yielding a number, text, or URI. QR codes are already printed on many products,
W_W_W_._CO_M_P_U_T_E R_.O_R_G_/C_O_M_P_U T_E_R
30 COMPUTER
___
ing retail transactions and replacing barcodes with a tag that was machine readable at a distance of up to 10 feet. But after several trials with leading vendors such as Walmart and Tesco, the EPC standard met with limited up- take principally because, in practice, a significant number of tags were unde- tectable due to factors in retail envi- ronments such as poor product/tag ori- entation and the presence of materials
Computer
qM qMqM
Qmags
Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page qMqM
THE WORLD’S NEWSSTAND®
Computer
qM qMqM
THE WORLD’S NEWSSTAND®
Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page qMqM Qmags
including newspapers, magazines, bill- boards, and coupons; they even appear on prime-time television ads.
However, in practice, many QR ad-
vertising campaigns result in a poor
customer response. The reasons re-
late to the requirement that a prein-
stalled application is required to read
QR codes—which can be a barrier for
some users—as well as difficulty in po-
sitioning the phone so that the camera
can focus and accurately decode the
image. Some advertisers also feel that
a visible QR code spoils the aesthet-
ics of their campaigns (h_t_tp_s_:/_/w_w__w
.techdirt.com/blog/wireless/articles
This new technology standard, along with the availability of inexpen- sive BLE tag hardware and tag readers already integrated with smartphone hardware, has been a considerable cat- alyst in this space. Many established computer companies, and a significant number of startups, are experiment- ing with products and business oppor- tunities associated with these tags. As with the other tagging technology, this is also an enabler for the IoT, but with more opportunities for ubiqui- tous deployment, higher-accuracy tag reads, and the ability to blend in invis- ibly with a product.
THE PHYSICAL WEB
In the Physical Web, people, places, and things have webpages to provide information and mechanisms for user interaction. The notion of open Web technologies as the bridge to the phys- ical is not new: access points, routers, solar panels, electricity meters, and coffee shops have Web landing pages, for example. However, it is the breadth and depth of the stack surrounding the Web that make this an appealing vision for the IoT’s evolution. To be sure, HTTP will not be an exclusive protocol for communication with things in the same way that it is not an exclusive protocol for the Internet— there are plenty of use cases where Web protocols do not have the desired properties, such as the Real-Time Streaming Protocol (RTSP).
Like the conventional Web search engine to which you submit a query and it returns text snippets and links to relevant “things,” the Physical Web would return search results. However, because the IoT is the world that we can see, hear, and touch, search results would not only be ordered by conven- tional ranking algorithms but also by
proximity, and thus results could be shown as lists, enhanced maps, or floor plans. Searching the physical Web at home would bring up thermostats, DVRs, TVs, home audio systems, rout- ers, and water and electrical meters as results. You might see a snippet and a link to the manual for a nearby micro- wave oven, along with other links that control or provide information about devices in your house. There would be a lot to show, but no more than the re- sults of any Web search today; search engines are good at ranking and pro- viding the most relevant items first.
Webpages are a great technology for human-to-machine (H2M) inter- action, but many use cases for the IoT are machine to machine (M2M). One solution that has already had great success in combining human and machine-readable content in open Web technologies is the use of structured data embedded by webmasters into their pages. The data formats used by S_c_h_e_m_a_.o_r_g and others let user agents and cloud services act intelligently, parsing data for events, organizations, people, places, products, reviews, and so on and acting on them either in- teractively or proactively. Structured data could also lead to more uniform user interfaces across devices, so that when users learn one interface, they do not have to relearn it for another device—for example, setting the time on an appliance that has a clock.
Open Web technologies, including HTML, Ajax, HTTPS, and OpenID, and structured data apply equally well to the IoT. However the open Web does not have an effective mechanism for locating objects in the physical world. One approach that looks particularly appealing is the use of radio beacons that broadcast URLs at very low power and over a small geographic radius.
JANUARY 2015 31
qM qMqM
Qmags
IN THE PHYSICAL WEB, PEOPLE, PLACES, AND THINGS HAVE WEBPAGES TO PROVIDE INFORMATION AND MECHANISMS FOR USER INTERACTION.
/20120307/06130018010/qr-codes-ugly
__________________________
-overused-doomed.shtml).
________________
Bluetooth low energy
One of the more promising new tech- nologies in the device tagging space is Bluetooth low energy (BLE),8 part of the Bluetooth v4.0 standard (Blue- tooth Smart) and adopted by the Blue- tooth Special Interest Group in 2010. Consequently, Bluetooth silicon vendors have included BLE in their latest chipsets, and all smartphones released in the last few years have BLE hardware, with various levels of capability depending on operating system support.
Bluetooth silicon can be pared down to only include the BLE aspects of the standard, removing the need for compatibility with classic Bluetooth. This results in a small, low-cost silicon implementation that can be used as a low-power electronic tag. Tags based on BLE can signal their presence by transmitting an advertising packet once per second at a power budget that enables them to operate for up to one year on a lithium coin cell battery (about the size of a US quarter, with 240-mAh capacity).
Computer
Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page qMqM
THE WORLD’S NEWSSTAND®
Computer OUTLOOK
qM qMqM
THE WORLD’S NEWSSTAND®
(a)
(b )
security, privacy, and cost. If interac- tion latency and connectivity are not a problem, cloud computing is an at- tractive paradigm. In recent times, the detrimental effect of malware on home computers has made the cloud even more attractive.
Based on these observations, it seems like a good idea for the IoT archi- tecture to register every device with a cloud service and communicate with that service alone. Users or other com- puters would then interact with this service to determine the device’s sta- tus or control its behavior. However, there are other factors to consider.
Benefits of peer-to-peer
Although the cloud model is clean and straightforward, the requirements for full Internet communication might be too costly or burdensome for what in many cases are simple low-performance devices. In practice, any hardware that can connect directly to the Internet would require a physical Ethernet, Wi-Fi radio, or cellular modem, all of which elevate the device’s cost and power consumption. In practice, it might be better to have one bridging device that supports Wi-Fi and enables simple pe- ripheral IoT devices to talk to the bridge. Many low-performance communication standards such as ZigBee, 6LoWPAN, Bluetooth Classic, and BLE have evolved to fill that gap—with no clear winner. As with the opportunity for tagging de- vices, the new BLE standard could well dominate this low-end space, but it will take time for this to play out.
IoT devices, unlike the traditional Internet, benefit from the concept of proximity as we described when intro- ducing the Physical Web. However, it is not just the knowledge of nearby ob- jects that is useful: colocated devices have the opportunity to cooperate
W_W_W_._CO_M_P_U_T_E R_.O_R_G_/C_O_M_P_U T_E_R
IoT services
Internet cloud
IoT services
Internet cloud
Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page qMqM Qmags
FIGURE 3. Alternative approaches for the IoT. (a) Peer-to-peer, with gateway to an IoT service); and (b) centralized IoT service. Because the IoT will connect many devices with constrained networking capabilities, P2P networking may become more prevalent and as common as cloud-based solutions are now.
One of the first projects to promote this idea was HP’s Cooltown,3 which used infrared beacons to transmit URLs. More recently, BLE provides a similar low-power beacon mechanism that can also integrate a URL emitted in short periodic advertisement packets (w__w_w_.u_r_ib_e_a_co_n_.o_r_g).
One way to bridge the gap between the physical and virtual is to attach beacons to all our objects. These bea- cons would broadcast a URL along with other information to help with rang- ing. It sounds like a huge investment, but beacons currently cost less than US$5, and the price is likely to drop, thus future manufactured “smart” objects are likely to integrate this ca- pability. By utilizing proximity URL beacons, the potential for interacting with the Physical Web is not only more practical, but has greater utility than both NFC and QR codes.
CLOUD COMMUNICATION VERSUS PEER TO PEER
One of the long debates in computer science has been whether to build centralized systems or to make them fully distributed. Much of the Web
today takes a centralized approach for its services, but it is not clear that this makes sense for much of the IoT. Fig- ure 3 compares the approaches, with cloud-based services representing the centralized paradigm.
Benefits of cloud computing
The computing world has shifted par- adigms several times, from the cen- tralized mainframe computer to the decentralized PC running standalone applications back to today’s central- ized cloud services. The computing industry is gravitating toward more centralized cloud services primarily because they are easier to manage. Advantages include the economics of scale when building datacenters, auto- matic backup of all data, and enforced physical security. However, modern client devices are both capable and flex- ible. Laptops utilize high-performance multicore technology, and even smart- phones contain powerful processors. We have the option of running simple clients connected to powerful cloud services or powerful local apps that run on their own. The decision comes down to our tolerance for tradeoffs in latency,
32 COMPUTER
Computer
qM qMqM
Qmags
Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page qMqM
THE WORLD’S NEWSSTAND®
Computer
qM qMqM
THE WORLD’S NEWSSTAND®
Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page qMqM Qmags
with each other in real time and fulfill a task that would not be possible with any single device.9 A simple example is the sharing of peripherals. A com- puter that finds itself close to another device with a bigger and better screen could wirelessly share it, for example, to take advantage of its display when playing a movie. Likewise, a computer could utilize a component that it does not have if it is wirelessly discovered on another nearby device—examples include a sensor, camera, or mouse.
Proximate sharing will be an im- portant aspect of the IoT, but this will not be easy to accomplish. Devices need to discover each other, trust each other, and then make a connection. Further- more, sharing is achieved by protocols and data formats that need to be stan- dardized and supported by both ends of the connection. In a world in which there are many more commercial play- ers than in the early days of the Internet, this becomes a difficult proposition. The traditional Internet has evolved a core set of standardized protocols, but in the undefined IoT world, many stan- dards and proprietary solutions are still up in the air. To illustrate how difficult this can be, consider the Digital Living Network Alliance (DLNA; http://dlna _.o_r_g),whichwasestablishedtoenable multivendor consumer electronics to discover each other and share content and services. To date, DLNA has only been integrated with a small percent- age of networked products because, in practice, larger companies are driven by the financial rewards of dominating a market with their own proprietary ecosystem.
Hybrid IoT solutions
Innovative solutions to latency prob- lems associated with existing cloud- based networking attempt to combine
various ideas. One approach, called edge computing, moves some of the cloud processing closer to devices that require real-time interaction, thus re- ducing the number of network hops and hence latency. An example of edge computing is the cloudlet paradigm,10 which provides a means of rapidly en- abling real-time services in the fixed infrastructure to be used by mobile devices, such as smartphones and wearables, just one network hop away using Wi-Fi. This is achieved through a virtual machine (VM) running on a powerful nearby workstation, and dy- namically provisioned with software and services customized for that ap- plication. When the task is complete, the resources can then be freed up, allowing a new VM, or multiple VMs, to be instantiated on the cloudlet. A cloudlet lets IoT devices interact in real time with cloud-like services, even though they are far removed from a datacenter.
WHAT THE IOT
MEANS FOR APPS
Today, any new IoT product almost al- ways comes with a smartphone app to control it. This is a consequence of two dominant forces: native smartphone appsoffertheonlypracticalmeansto access and communicate with smart devices, and there has been so much recent financial success with mobile apps that they are now expected. Peo- ple assume that they will always use apps for every possible interactive ex- perience with IoT devices.
This approach unfortunately cre- ates problems as the IoT landscape grows and matures. Although it is easy to imagine a phone with a few dozen applications on it, things become more problematic with millions of IoT de- vices. In the very near future, you will
likely pass thousands of smart devices every day, each one capable of interac- tion. It does not scale to have to install an app before using each one. You will also likely want to delete apps as you install others, because there will be many devices that you will interact with only once.
This becomes even more evident when you factor in how smart devices will have a significant long-tail ef- fect.11 Instead of a few big apps that you use every day for many tasks, you will have a huge range of small-device apps that offer the tiniest of interac- tions, such as just controlling an on/ off switch. In fact, many devices will eschew interactivity all together and only offer a snippet of data, for exam- ple, when the next bus is arriving. The long tail of the many potential uses of the IoT challenges the scope and breadth of today’s smartphone apps, requiring only micro-interactivity or micro-information.
To enable this new Physical Web for the IoT, we need a way for any user, with any smartphone or tablet, to walk up to any IoT device and interact with it (without a specialized app). We need a richer extension of today’s Web, al- lowing each smart device to wirelessly broadcastaURLtoitssurroundings. Proximity is the context that can be used to filter this to a tractable num- ber, allowing any smart device to list and then interact with other nearby smart devices. This is the basis of a discovery service that is the best of today’s native and Web apps, creating a new platform, a type of interactive lingua franca allowing devices of any type to offer data and interactivity to any other device.
The current model of native apps, while quite popular, is not up to the task of supporting the many and
JANUARY 2015 33
qM qMqM
Qmags
THE LONG TAIL OF THE MANY POTENTIAL USES OF THE IOT CHALLENGES THE SCOPE AND BREADTH OF TODAY’S SMARTPHONE APPS.
Computer
Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page qMqM
THE WORLD’S NEWSSTAND®
Computer OUTLOOK
qM qMqM
THE WORLD’S NEWSSTAND®
Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page qMqM Qmags
varied use cases that the IoT will bring about. We need to extend well-known systems, such as the Web, to allow a significantly easier and more light- weight approach for enabling devices to interact with one another.
Context sensing
One of the ways an application can perform more effectively is through context awareness. Sensing what is around a host device (and its user) and the context in which it is used allows
a display, actuator, or switch can be controlled from a browser or through a Web service, thus making it easy to integrate related information into control decisions. For example, emerg- ing Web-connected irrigation systems might provide an interface for specify- ing the plants in your garden and use Web services to determine expert wa- tering recommendations.
Privacy and security
One of the primary challenges for the future will be avoiding the darker con- sequences of a world with globally connected devices. The Physical Web could enable hackers to control our de- vices unless precautions are taken. The conventional Web already has security measures in place that can be applied to the Physical Web, but it is unclear if these will be suitable or even adequate for all IoT applications. In addition to hacking, social threats can result when knowl- edge is leaked in unexpected ways. For example, knowledge that a house is in an energy-saving mode could be a good indication that nobody is home and thus invite a burglar. These challenges will become more pressing as use of the Physical Web continues to grow.
M
smarter automation and services. As- sociating a Web URI with every person, place, or thing forms the basic mecha- nism for bridging these two technol- ogies. A key enabler is the ability to discover proximate objects, and BLE beacons containing URIs are a useful tool. Web browsers can readily display Web information along with descrip- tions of nearby IoT objects; they are a promising familiar on-ramp to the IoT.
W_W_W_._CO_M_P_U_T_E R_.O_R_G_/C_O_M_P_U T_E_R
34 COMPUTER
THERE IS A REASON PEOPLE ARE EXCITED ABOUT THE IOT: IT FEELS LIKE A BIG OPPORTUNITY TO IMPROVE HOW WE DESIGN AND BUILD PRODUCTS.
FUTURE OPPORTUNITIES AND CHALLENGES
While the IoT applications will be varied and difficult to predict, some clear opportunities will arise for ubiquitous information gathering, context sensing, and control, which will likely be key enabling building blocks. However, some of the real challenges will be in the areas of pri- vacy and security.
Ubiquitous information
Users often need help when they en- counter an unfamiliar device. Tra- ditional products were sold with a user manual—later, they might have included an informational CD. More recently, a customer support URL is provided that points to online doc- uments, and even in the event that this is lost, a simple Web search for the model number will usually locate the online manual. In the IoT future, this problem can be solved by the Physical Web enabling all products to transmit a wireless machine-read- able URL that can be received by a nearby smartphone or tablet with lit- tle effort from the user.
an app to adapt how data is presented and filtered; consider, for example, a local map application that automat- ically shows the user’s current posi- tion. The more context a device has, the more likely it can automatically provide a user with the required in- formation. However, the worst thing a context-aware application can do is in- fer something on behalf of a user that turns out to be wrong. The future IoT will allow information to be collected from nearby physical sensors and Web services and then shared between de- vices on a scale that has not been pos- sible before. As a result, the Physical Web will lower uncertainty and im- prove accuracy. A consequence is that future “smart devices” will become much smarter.
Actions and control
As a complement to sensing, the IoT offers us a way to control the physical world through displays, actuators, and switches. Many modern systems benefit from remote control because it simplifies physical interaction design and extends capabilities. In the Phys- ical Web paradigm, anything with
erging the virtual Web with the IoT will enable really smart devices, providing
Computer
qM qMqM
Qmags
Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page qMqM
THE WORLD’S NEWSSTAND®
Computer
qM qMqM
THE WORLD’S NEWSSTAND®
Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page qMqM Qmags
ABOUT THE AUTHORS
ROY WANT is a research scientist at Google. His research interests include mo- bile and ubiquitous computing. Want received a PhD in computer science from Cambridge University. He is an ACM and IEEE Fellow. Contact him at _ro_y_w_an_t@_ g_m_a_il_.c_om_ orviawww.roywant.com/cs.
BILL N. SCHILIT is a research scientist at Google. His research focuses on smart personal and mobile technologies supporting knowledge work. Schilit received a PhD in computer science from Columbia University. He is an IEEE Fellow, a member of the IEEE Computer Society and ACM, and associate EIC of Computer. Contact him at s_c_h_ili_t@_g_m_a_il._co_m_.
SCOTT JENSON is a user experience strategist at Google. He has been in the design business for over 30 years, working on projects such as Mac System 7, Apple’s User Interface Guidelines, the Newton, Google Mobile Maps, and now the Physical Web. Contact him at _sc_o_tt@_ j_en_s_o_n._or_g.
There is a reason people are excited about the IoT: it feels like a big oppor- tunity to improve how we design and build products. Making the IoT as ac- cessible and useful as the Web is likely an even bigger opportunity.
REFERENCES
1. M. Satyanarayanan, “Pervasive Computing: Vision and Challenges,” IEEE Personal Comm., vol. 8, no. 4, 2001, pp. 10–17.
2. M. Weiser, “The Computer for the 21st Century,” Scientific Am., vol. 265, no. 3, Sept 1991; www.scientific american.com/article/the-computer _-fo_r_-t_he_-_21_st_-c_e_nt_u_ry_/.
3. T. Kindberg et al., “People Places and Things: Web Presence for the Real World,” ACM J. Mobile Networks and Applications, vol. 7, no. 5, 2002,
pp. 365–376.
4. R. Want et al., “Bridging the Physical
and Virtual Worlds with Electronic Tags,” Proc. ACM SIGCHI, 1999,
pp. 370–377.
5. S. Jain et al., “Exploiting Mobility for Energy Efficient Data Collection in Wireless Sensor Networks,” Mobile Networks and Applications, vol. 11, no. 3, 2006, pp. 327–339.
6. R. Want, “RFID: The Key to Automat- ing Everything,” Scientific Am., Jan. 2004, pp. 56–65.
7. H. Kato and K.T. Tan, “Pervasive 2D Barcodes for Camera Phone Applica- tions,” IEEE Pervasive Computing, vol. 6, no. 4, 2007, pp. 76–85.
8. R. Heydon, Bluetooth Low Energy, Prentice Hall, 2013.
9. B.N. Schilit and U. Sengupta, “Device Ensembles,” Computer, vol. 37, no. 12, 2004, pp. 56–64.
10. M. Satyanarayanan et al., “The Case for VM-Based Cloudlets in Mobile Computing,” IEEE Pervasive Comput- ing, vol. 8, no. 4, 2009, pp. 14–23.
11.
C. Anderson, “The Long Tail,” Wired, vol. 12, no. 10, 2004; http://archive .wired.com/wired/archive/12.10/tail _.h_t_m_l.
Selected CS articles and columns are also available for free at http://ComputingNow. computer.org.
On Computing
podcast
www.computer.org/oncomputing
Computer
JANUARY 2015 35
qM qMqM
Qmags
Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page qMqM
THE WORLD’S NEWSSTAND®