The Internet of Things and Ubiquitous Intelligence(4)

Dongliang Xie Yu Wang

(State Key Laboratory of Networking and Switching Technology,Beijing University of Posts and Telecommunications,Beijing 100876,P.R.China)

8 Cloud Computing

A distributed CPU is the core component in cloud computing and plays a pivotal role in the development of the Internet of things.Combining cloud computing and the IoTcan boost the entire industry and value chain if capability and resource sharing,quick service deployment,expansion of new human-thing interactive services,and deep data mining are optimized.When the Internet of things reaches a certain level,its dependence on cloud computing will be greater.

8.1 Concept and Current Status

With the development of electronics,communications,computers,and network technologies,cloud computing is an inevitable stage in the evolution of turing computing to network computing.Based on the Internet,network computing provides applications with hardware,infrastructure,platform,software,and storage services.The term“cloud”vividly describes the characteristics of cloud computing:virtualization,transparency,scalability,and elasticity.It is an apt term for describing the abstraction of underlying infrastructure.

Cloud computing integrates distributed computing,parallel computing and grid computing.At its core,it virtualizes computing,resources of large data centers and provides these to users as a service.

8.2 Key Technologies

Cloud computing is a form of data-intensive super computing.It has unique data storage and management technologies and a unique programming model.

8.2.1 Virtualization

Virtualization is at the core of cloud computing,and is the basis for allcloud services applications.It expands hardware capacity,makes software reconfiguration reduces easier,overheads of software virtual machines is reduced,and supports multiple operating systems.With virtualization,use of the server is greatly enhanced.

8.2.2 Data Storage and Management

In cloud computing,data is highly reliable and highly available because of storage redundancy.Data storage technologies involve cluster computing,data redundancy,and distributed storage.

The frequency of data access is much higher than that of data update because of one characteristic of cloud computing—analyzing the stored and accessing massive data[1].Therefore,cloud data management focuses on optimizing the data access operation.

8.2.3 Parallel Programming

The cloud computing programming modelmust be very simple,and complicated parallel execution and task scheduling must be hidden in the background.The programming model must be transparent,enabling users to develop programs for specific purposes.

Most cloud computing systems adopt Google’s Map Reduce[2].This is a distributed parallelprogramming model and also an efficient task-scheduling model.

8.3 Combining with the Internet of Things

Cloud computing and IoTare complementary.Development of IoT requires the powerful processing and storage capabilities of cloud computing.If cloud computing infrastructure is used to mine,process,and analyze mass data collected on the ubiquitous sensing layer,IoTcan quickly,accurately,and intelligently manage and control the physical world and provide technicalsupport for ubiquitous services.The IoTwill also be the largest user of cloud computing services,which will,in turn,lay the foundation for more successful cloud computing.

9 Cyber-Physical System

Cyber-physical system(CPS)was first proposed by the USNational Science Foundation(NSF)in 2006.CPS is expected to become the third wave of information technology following the computer and the Internet.CPS ubiquitous sensing,control,and computing and the close integration of human,machines,and things are the ultimate goals of IoT.

9.1 CPS Overview

CPSis complicated a multidimensional,embedded system that is integrated with computing,network and physical environments.With the integration of 3C(computation,communication,and control)technologies,CPSdelivers real-time sensing,dynamic control,and information services for large engineering systems.Through interactive,cyclic feedback in computing and physical processes,CPSclosely integrates these processes and adds or expands new functions in real time.In this way,a physical entity can be monitored and controlled in a secure,reliable,efficient,and real-time manner[5].

9.2 Characteristics

CPShas the following characteristics:

?It integrates computing,communication,and controlbecause it embeds computing capability deeply into each physical subsystem.It aims for precise control over the physical processes in networks.

?It requires the integration of computing and control technologies.To connect the cyber world to the physical world,CPSshould integrate computing technologies(which are discrete event-relevant and indifferent to time and space)with the control technologies(which are continuous process-relevant and time space-focused),enabling the discrete computing process to interact and tightly couple with the continuous physicalprocess.

?It is adaptive to dynamic changes in the physical world;it has powerful reorganization and reconstruction capabilities;it can esaily;and it is easy to connect with other CPSsubsystems.

?Because it has multidimensional time-space complexity,CPSshould be an open,trustable,and predictable embedded system.The embedded computing system of CPS interconnects and interoperates with other cyber systems via the Internet.Furthermore,CPShas stepped into fields that are closely related to national infrastructure and people’s daily lives.It is sensitive to security,and its technologies and products must be precise and highly reliable.

9.3 Challenges

To realize the goals of CPS,there are some technical challenges to be addressed.A good architecture is criticalin maintaining scalability,durability,diversity and continuous technical innovation in CPS[3].It is also the key to customer investment.

To ensure sustainable development of CPS,the architecture should be stable,a clear roadmap should be devised that can lead to a targeted study based on core CPSarchitecture,and a development schedule should be in place so that technical achievements are planned and predicted.The natural development cycle should be balanced.

There are several challenges in the development of CPS.The theoretical basis differs from the application model.The key to the success of a technology is its deployment in actual environments.It is also necessary to maintain the integrity,reliability and security of the architecture.Because of uncertainty in the external environment and potential changes,CPSshould respond to a system failure or malicious attack automatically,autonomously,and quickly.According to the characterisics described in 9.2 the OS and architecture should be able to manage redundant resources on the interconnected device layers,monitor errors in user applications and physical components,and quickly recover CPS from a system failure.Electricity CPSis a critical infrastructure,and guaranteeing its security is an important issue.

10 Summary

The IoT,cloud computing,and CPS interact and couple with each other.IoT is the intuitive application of CPS,and cloud computing provides technical support for IoTinformation service.The demands of economy and society on the Internet of things and CPSgo far beyond existing applications.Therefore,from the perspective of economic development and technical innovation,IoTand CPSare a trend in the development of global information technologies and industry.They will have a profound effect on existing industrial structure,become the core competence in the new industrial structure,and cultivate new economic growth.