Jian An, Xiao-Lin Gui, Xin He

parsimoniousknotRéseaux et Communications

16 févr. 2014 (il y a 3 années et 5 mois)

62 vue(s)



Study on the Architecture and Key Technologies for Internet of Things
Jian An, Xiao-Lin Gui, Xin He
Department of Computer Science and Technology, Xi’an Jiaotong Uinversity, Xi’an 710049, China;
The Key Laboratory of Computer Network Xi’an 710049, China;
E-mail: an2025918@sina.com
Keywords: Internet of Things; Architecture; Key Technologies; Sensor technology
Abstract. Internet of Things (IoT) combines sensor technology, embedded computing, Internet and
wireless communication technology, distributed information processing technology, and has broad
application prospects. Therefore, the revolution of the Internet of Things (IoT) has attracted highly
attention of domestic and foreign government, academia and industry. There are still many open
issues in the definition, architecture and key technologies of IoT. In this paper, firstly, based on
analysis of existing literatures of IoT, the basic concept of IoT is discussed from the view point of
the architecture and key technology. Five essential factors which the IoT should have are
summarized. Secondly, the architecture is further proposed. its structure and functions are described.
Thirdly, the key technologies of IoT are analyzed, including recognition technology, sensing
technology, network technology and smart technology. Finally, this paper gives some suggestions
of IoT development.
1. Introduction
As the rapid development of sensor technology, embedded technology, wireless communication
technology, high-performance computing and related fields, Internet of Things (IoT) as a new
generation of intelligent network came into being. It takes the RFID and the wireless sensor network
as the sensation foundation, through the integration of Internet and database technology for data
transfer and sharing and using high-performance computing technology to achieve information
management and decision-making [1,2]. Internet of Things has obtained the national government and
research institutions broad support, IBM proposed the "Smarter Planet", Japan and South Korea
proposed the "U-Japan" and "U-Korea" strategy. In view of this, China proposed the “Sensing
China". Internet of Things is called the third wave of Internet information technology which after
computer, Internet and telecommunications networks.
The paper first analyzes the basic concepts of the Internet of Things, then has a specific research
on its architecture and key technologies and summarized the future existing opportunities and
challenges finally.
2. What is the Internet of Things?
Internet of Things has a great value and development prospects which attracting a wide range of
industrial and academic attention. However, the concept, architecture and key technologies of
Internet of Things are still in the initial chaotic phase. From now on, researchers are still unable to
give a clear concept and structure.
International Telecommunications Union (ITU) believed that IoT is mainly realized the
Interconnection which between Thing to Thing, Human to Thing and Human to Human [2], some
2012 International Conference on Electrical and Computer Engineering
Advances in Biomedical Engineering, Vol.11
978-1-61275-029-3 /10/$25.00 ©2012 IERI ICECE 2012

329


scholars also proposed other concepts, such as Machine to Machine [3], Cyber-Physical
Systems[4],Pervasive Network, Next Generation Internet.
The paper summarized that IoT should have following five features which based on the analysis
and research on the IoT concepts.
Ser
vice
Security
Dev
ice
Techn
olo
gy

Figure 1. The Internet of Things
First is the things, in IoT, the so-called "Things" will no longer be confined to the traditional
range of physical devices, people and other mobile entities will greatly expand the scope of
application of IoT, They existing and moving in time and space, we can discovery and identification
them by relevant attributes which attach on themselves; Next is sensing, IoT take the outside
physical information as a sensing foundation, using RFID and other technologies to realize
identification of things, using sensor nodes to achieve dynamic sense of environmental information,
then making things with sensing ability and building network by using various wired and wireless
network communication technology to achieve information transmission; Third is connection, it will
realize the distributed data sharing by
integrating sensing subnets with existing networks. Fourth is intelligence, IoT using high-
performance computing technology to achieve intelligent data management, decision-making; Fifth
is control, IoT feeding back the decision information to the nodes, then realize the control of things
and environment.
As shown in Figure 1, in IoT, anyone and anything can seamless integration with any network in
anytime and anyplace, to form a real new generation of intelligent information network.
3. Architecture for the Internet of Things


The architecture of IoT should be an open architecture, using open protocols to support a variety of
existing network applications; In addition, it should also include some scalability, security and
semantic representation middleware to promote data world integration with Internet. Therefore, by
summing up some literature's research work [5~10] and combing with our proposed features of
Internet of Things, we try to design this architecture (Figure 2) to guide theoretical research. Model
focuses on the qualitative description rather than specific protocol definition, Therefore, IoT should
be a five-layer structure, it contains sensing control layer, networking layer, resource management
layer, information processing layer and application layer.
Sensing and control layer: It is the foundation of the development and application of IoT,
including RFID readers, smart sensor nodes and access gateways, etc. A variety of sensor nodes
sensing the relevant information of the target environment and pass it to the nearest gateway, then
gateway submit the data which collected via the Internet to background processing platform.
Networking layer: It is mainly responsible for the different types of networks integration, such as
Internet, Mobile Communications Network, and Broadcast Television Network. In addition, it will
also provide routing, format conversion, address conversion, etc.
Resource management layer: It will provide the initialization of resources, monitoring the
operation status of resources, coordination of work between various resources and achieve cross-
domain interactions between resources.
Information processing layer: This layer realized reasoning and semantic understanding of
sensing data, it also provide data query, storage, analysis, mining, etc. Cloud computing could
provide a good platform for sensing data storage and analysis. It is an important component of
information processing.
Application layer: After analyzing and processing the sensing data, application layer using these
data to provide users with a variety of different types of services. IoT application can be divided into
network monitoring (logistics, pollution control), control type (intelligent transportation, intelligent
household), scanning type (mobile purse, highway no parking fees), etc.
In addition, the IoT should also include some support technologies such as network security,
fault-tolerant mechanism and quality control which throughout all levels to provide application
support.
4. Key technologies for the Internet of Things
Key technologies for the Internet of Things as shown in Figure 3 ,such as RFID, IPv6 which
responsible for things identification, sensor technology which responsible for the dynamic
information sensing, communication technology and network integration technology which realize
the information transmission and intelligent information processing technology, they were called
four key technologies of IoT.
4.1 Identification technology
Internet of Things is a vast network that contains millions of things, various intelligent equipment
forms a network through interconnection ways. Therefore, what we first need to solve in the
application of IoT is recognition of things. Identification technology is associated with the things,
and it is a globally unique value which used to unambiguously identify the object. The essence of
identity is encoded and digitize to all things. There are many coding rules, such as EPC coding which
used RFID technology, IPv4 and IPv6 which based on TCP/IP. Mapping and compatibility between
different coding rules, mapping between coding and service, all these problems need to solve in
future.

331



Figure 3. Key technologies of IoT
Radio Frequency Identification (RFID) [11] is a non-contact identification technology, it can
automatically identify targets and collecting data through its radio frequency signal which identify
process without human intervention. FID tag technology has the ability to uniquely identify the
object which can ensure object located and managed correct in communications and information
processing.
IPv6 technology which based on TCP / IP uses stateless address allocation scheme can solve the
massive address efficient allocation, it can meet the needs of the vast network address in IoT. What's
more, IPv6 protocols take some questions fully into account at the beginning of design, such as
mobility, security, quality of service. Although many details of the IPv6 needs further improvement,
its numerous characteristics which more suit for IoT will make it become the basis network
technology in the future.
4.2 Sensing technology
Data generation, access, transmission, processing and application are the important components of
IoT. The data acquisition is one of important links; otherwise, IoT is nothing but vain words. In the
IoT, based on sensing devices including RFID, sensors, infrared devices, global positioning system
devices, the real-time data are acquired to terminal.
Sensor can detect external environmental signals, including heat, power, light, electricity, sound
and others, they can provide the raw data which for data transmission, processing and application in
IoT. It achieved the dynamic information access of things, so that "things" have the ability to sense
the external world. Therefore, we can regard RFID, IPv6 technology as the "eyes" of people and
sensors can be considered "skin", RFID, IPv6 technology can solve the "Who", that is to achieve
recognition of things; Sensors can solve the "How", that is to achieve sensing of things.
The data collection process of IoT mainly through the control of various sensing devices to collect
information on the surrounding environment, and transfer the data to the nearby sink node through
the corresponding WSN network protocols, at last, delivered to the users of application layer through
the Internet. Therefore, the research of the WSN network protocol is a core technology in IoT. The
goal of design network protocol is to meet the application requirements while minimizing network
overhead, improving system throughput and improving the overall utilization of resources. At
present, most research of WSN network protocol focuses on the routing layer and MAC layer.
WSN is a multi-disciplinary overlapping domain [8], in addition to basic network protocols, there
should be a wide range of supporting technologies to achieve a variety of applications better in IoT.
WSN supporting technologies have the time synchronization, node localization, data fusion, quality
of service assurance and network management. In view of the different application scenes, we will
set the different request to these support technologies, for example, the multi-sensor collaborative
sensing, data compression and integration is established on the basis of time synchronization,
localization is mainly used in distance positioning, materials tracking and other fields.
332


Because of the broadcast nature of WSN in the wireless channel and the characteristics of self-
organizing network, node easy to suffer each kind of active threat or the passive attack, then obtain
the privacy information which stored in the node. In the future, different applications will have
different security needs, for example, in the remote medical environment, users usually don’t hope
that their own case information is seen by the third party, in the open country environmental
monitoring, must guarantee that the node will not come under the deceit attack and collect the false
information. Security objectives at this stage mainly depend on the key management, authentication
and data encryption.
Sensing technology will be a hot spot in the future research of IoT. Because the IoT has its own
characteristic which different with traditional network, correlation technique's research must aim at
the specific application scene, meets the different design need. In addition, human-based will
become a major characteristic in the data sensing of IoT, the human-based sensing will take human
participate in the process of information sensing. The human is not only the information gainer
(consumer), but also the information participant (provider). Its role's change will cause the new
transformation in the future networking deployment and sensing technology.
4.3 Communication and network integration technology
The transmission of sensing data in IoT mainly rely on network and communication technology, and
the communication may divide into the short distance wireless communication and wide-area
network according to the transmission type.
Network integration technology makes full use of different network communication resources,
according to different application environment and local conditions, providing a richer network
services for users through flexible and efficient networking ways. The future network level of IoT
will not be limited to traditional and single network structure. We will eventually realize the
seamless and transparent integration for different types of network, such as the Internet, 2G/3G
mobile communications network and Broadcasting and Television network. It involves wired,
wireless, mobile and other means of access, the unification of heterogeneous network address,
conversion, packet format, routing options and other issues. We believe that network convergence
should follow the principle of functional separation and hierarchical network, making different
terminals in different access modes can share the same network platform, isolating the upper
application and underlying control, shielding the complexity of heterogeneous network. The initial
stage of network convergence is build a unified service platform which provides the corresponding
interface for each network, gradually forms the network architecture which take the IP Internet as the
core, mobile communication network and cable television network coexistent. The ultimate goal of
networking is to build a truly unified and open platform which different network boundaries will no
longer exists, it will provide broadband, narrowband, mobile, wireless and other access methods, to
achieve any time, any place and any network interconnection.
4.4 Intelligent information processing technology
In the Internet of Things, in order to sense an event, it requires the deployment of many types of
different sensing devices to monitor different attributes of the event, and then determine whether the
incident occurred through the integration of sensing data. The key technology is how to transform
the physical sensing data into the logic data which easy to understand by man and machine.
Intelligent information processing technology integrates the intelligent computation, data mining,
optimized algorithm, machine learning and so on, we can process and analyze the sensing data
through smart technology and deliver the results to the user finally. For example, when we pick up a
product in the supermarket, it can back to us some interest information such as the origin of products,
structural components which help us better understand the product through intelligent information
processing technology. The revolution of IoT is thought injected into the things, so they can
communicate with people directly and form an intelligent network finally. How to make "things"
have thought, we believe that the key is the introduction of a variety of smart technology.
333


In addition, cloud computing can take full advantage of network computing power to achieve the
sharing of resources and services, there are many advantages of cloud computing, for example,
virtualization, customization flexibility, high reliability, security, and powerful computing power and
storage capacity [12]. Cloud computing can be effective management of huge amounts of data to
improve resource utilization and quality of service.
In addition, there are many other key technologies, such as security privacy, service discovery and
search, fault-tolerant mechanisms need further study and discussion. Currently, the research for the
Internet of things is just beginning, as many problems yet to be resolved, so there is much room for
future research.
5. Conclusion
Internet of Things brings us a great opportunity, but there also exist many problems at same time, we
believe that hinder the development of future IoT will no longer be technical details, but how to
realize large-scale application.
First is the architecture, protocols and standards. How to coordinate various aspects to formulate a
set of standards is urgently.
Second is the industry plan, any good technology, if there is no reasonable planning and business
model to support the operation, it is difficult to sustainable development.
Finally is the trade integration, the application of future IoT will no longer be limited to a specific
area, it will be a cross-product which between different industry and different disciplines. Only
breaking down the regions barriers, and carrying on better communication and integration, Internet
of Things will finally succeed.
6. Acknowledgment
We would like to thank the anonymous reviewers for their valuable comments. This work is
supported by National High Technology Research and Development Program of China
(No.2008AA01Z410), National Natural Science Foundation of China (No. 60873071) and the
Fundamental Research Funds for the Central University.
References
[1] R. G. Gustavo, M. O. Mario, D. K. Carlos, “ Early infrastructure of an Internet of Things in
Space for Learning”, Eighth IEEE International Conference on Advance Learning
Technologies, 2008, pp.381-383.
[2] International Telecommunication Union UIT. ITU Internet Reports 2005, “ The Internet of
Things” , 2005.
[3] J. P. Conti, “ The Internet of Things” , Communications Engineer, 2006,4(6), pp.20-25.
[4] W.Wolf, “ Cyber-physical Systems”, Computer, 2009,42(3),pp.88-89.
[5] S. B. Shen, Q. L. Fan, P. Zhou, “ Study on the Architecture and Associated Technologies for
Internet of Things ”, Journal of Nanjin University of Posts and Telecommunications,2009,29(6),
pp.1-13.
[6] C. Amardeo, J. G. Sarma, “ Identities in the Future Internet of Things”, Wireless Peers
Communication, 2009, 49, pp.353-363.
[7] L.E. Akyildiz,et al, “Wireless sensor networks: A survey”, Computer Networks, 2002, 38,
pp.393-422.
334


[8] F. Y. Ren, H.N. Huang, C. Lin, “ Wireless sensor networks ”, Journal of Software, 2003, 14(7),
pp.1282-1291.
[9] B. Y. Wang, “ Review on internet of things ”, Journal of Electronic Measurement and
Instrument, 2009.23(12), pp.1-7.
[10] L. Sha, S. Gopalakrishnan, X. Liu, “ Cyber-Physical Systems:A New Frontier ”, 2008 IEEE
International Conference on Sensor Networks, Ubiquitous and Trustworthy Computing,
2008,pp.1-9.
[11] S. E. Sarma, S. A. Weis, D. W. Engels, “ RFID Systems and Security and Privacy
Implications”, Proceedings of the 4th International Workshop on Cryptographic Hardware and
Embedded Systems, 2003, pp.454-469.
[12] K. Chen, W. M. Zhen, “ Cloud Computing: System Instances and Current Research”, Journal
of Software, 2009,20(5), pp.1337-1348.
335