Cyber-physical systems: Providing Quality of Service (QoS) in a

5th IEEE International Conference on Digital Ecosystems and Technologies (IEEE DEST 2011), 31 May -3 June 2011, Daejeon, Korea

Cyber-Physical Systems Providing Quality of Service (QoS) in a Heterogeneous Systems-of-Systems Environment Tharam Dillon, Vidyasagar Potdar, Jaipal Singh, Alex Talevski Digital Ecosystems and Business Intelligence Institute Curtin University Perth, Australia E-mail: {tharam.dillon, vidyasagar.potdar, j.singh, a.talevski}@cbs.curtin.edu.au Abstract—The very recent development of Cyber-Physical Systems (CPS) provides a smart infrastructure for connecting abstract computational artifacts with the physical world. As new CPS applications start to interact with the physical world using sensors and actuators, there is a great need for ensuring that the actions initiated by the CPS is timely. This will require new Quality of Service (QoS) functionality and mechanisms for CPS. This paper will investigate QoS requirements for CPS and discuss the state of the art in CPS QoS techniques and models. Research challenges to providing QoS for CPS will be discussed and a new CPS framework that provides end-to-end QoS in a complex system-of-systems CPS will be presented. Two case studies, in smart energy grids and intelligent vehicle systems, will be provided to clearly illustrate the QoS requirements in CPS and how it can be achieved. Keywords-cyber-physical systems; web of things; quality of service; ubiquitous computing

I.

INTRODUCTION

Wireless sensor networks (WSN) is one of the key technologies in bringing physical world information into the cyber world as sensors would monitor the physical environment and transmit the physical state information over a network to a database for further processing. These sensors would abstract the collected data so it can be understood within the cyber world frame of reference. However, there is a disconnect between the physical world and the cyber world which limits how the cyber world manipulates or works on the physical world. As society and national infrastructure becomes more instrumented, there is a need to move from a one-way passive process, namely transmitting data from the physical world to the cyber world, to two-way interaction (active and passive) between the cyber world and physical world. Cyberphysical systems will be the key enabler in achieving this. The very recent development of Cyber-Physical Systems (CPS) provides a vision of interconnectedness between the physical world with the cyber world, allowing for robust and flexible systems with multi-scale dynamics and integrated wired and wireless networking for managing the flows of mass, energy, and information in a coherent way through integration of computing and communication capabilities with the monitoring and/or control of entities in the physical world in a dependable, safe, secure, efficient and real-time fashion [1]. Physical systems are becoming more and more complex, with new functionalities provided through embedded

ISBN: 978-1-4577-0872-5 (c) 2011 IEEE

microchips. Most new physical devices have an electronic brain to provide more functionality and better manage the usage of the device. According to IBM, a new family car has more lines of code than a space shuttle and use microchips to monitor the engine’s temperature and adjust the fuel injected into the engine, providing the right amount of fuel required to go faster when the driver accelerates the car. Unfortunately, these embedded systems are built specifically for an application and in most cases use proprietary technology. The applications are usually custom built and provide very limited connectivity to external systems. CPS changes all this as it will provide seamless and synergetic integration between sensing, computation, and control with physical devices and processes for new CPS-based applications. Not surprisingly, CPS was listed as the No. 1 research priority by the U.S. President’s Council of Advisors on Science and Technology [2]. The US National Science Foundation (NSF) listed three key applications for CPS - future distributed energy systems, future transportation systems and future health care systems [3]. The potential benefits of CPS have led to governments and businesses around the world to actively develop CPS capability and technologies to meet these future applications needs. Very recently, the UK's Technology Strategy Board launched a £5m initiative for promoting development of CPS applications [4]. In the US, the Power Systems Energy Research Centre (PSERC) was awarded $5.5m to research innovations in future electricity grid using CPS [5]. The NSF, a peak research body in the US, has allocated funding of $30m over 5 years in CPS research programs. However, the complex nature of a system-of-systems CPS, with different degree of application criticality operating at different time and space scales, makes it difficult to provide end-to-end quality of service (QoS) for such systems. While QoS is a widely researched topic in computer networks [6, 7], the authors have yet to find any suitable QoS protocol or technology for CPS applications. A QoS protocol or technique for CPS is a very important requirement for CPS applications, more so than for most computer applications. In most cases, computer applications or users can handle QoS deficiencies like increased delay in viewing a video or accessing a webpage. As CPS interact with the physical world, QoS requirements can be the difference between life and death or financial loss. For example, future CPS implementations for

330


road safety in automobiles will make use of proximity sensors to detect an approaching vehicle. This event will trigger the brake control unit to stop the vehicle before it hits the oncoming vehicle. However, long delays in any one CPS component or the aggregate of all components will not allow the vehicle to stop in time, thus crashing the vehicle and resulting in injuries, loss of life or financial loss. This article will investigate QoS requirements for CPS and techniques that can be used for providing QoS in CPS. Section II presents the state of the art in QoS for CPS. Section III discusses the open challenges in providing QoS for CPS while a proposed Web-of-Things (WoT) CPS framework is presented in section IV. In section V, We highlight how end-to-end QoS can be provided for CPS applications using two case studies. II.

STATE OF THE ART

In this section, we describe the related work in the area of quality of service in cyber physical systems from several dimensions including energy management, real time communication, security, privacy, trust, sensor failure, scalability, and resource management to identify the issues that may arise in the context of sensor actuator communication across multiple communication platforms. In the context of energy management within CPS, Parolini et al. [8] present a coordinated control strategy in data centers, aiming to optimize quality of computational services, while simultaneously reducing the energy costs for computation and cooling. They consider the data centre model as a computational (cyber) network, and a thermal (physical) network. The proposed control framework is implemented based on the time-scales. This framework can be further improved by introducing bidirectional flow of information to coordinated control and also considering external influencing factors such as fluctuations in energy prices in market. Compared to Parolini et al. [8], Ilic et al. [9] take a more CPS focused view for energy management. They propose a model for cyber-physical energy system (CPES) of key components, which is used to decide what needs to be sensed at what rate and does it require any associated data mining activity as this is quite crucial in autonomic CPS, that is practically desired. In this direction this is a very useful and innovative work. Additionally certain studies consider performance of mission critical cyber physical systems as in [10]. For such systems, energy consumption and real-time energy transfer are the two key factors that are considered when evaluating performance. The main focus of the study performed by Jiang et al. [10] is to set up an application model for mission-critical low-end CPS and propose a policy, termed as Optimal Lazy Schedule (OLS), to reduce energy cost without breaching deadline constraints. The authors claim that their lazy schedule policy is much more energy efficient than non-lazy schedule. Recently some work has also been undertaken to understand the challenges to improve the survivability of CPS [11]. For example, the introduction of the concept of trust analysis in CPS can result in proliferation of security threats like DoS attacks. On the other hand proactive architectures that are robust need to prove their performance bounds. Reactive algorithms for real-time detection and response may need detailed investigation in terms of scalability. In this space Xia et al. [12] presents a

ISBN: 978-1-4577-0872-5 (c) 2011 IEEE

feedback scheduling framework to manage QoS in CPS. The concept of dynamic scheduler is used to dynamically regulate specific scheduling parameters of relevant traffic thus ensuring desired level of QoS. The proposed framework aims to increase the predictability of CPS, hence driving the resulting QoS to the pre-defined level. However ensuring high scalability at the same time managing security and privacy is another challenge in itself. Han & McMillan [13] present a model of information security of CPS in the family of Flexible AC Transmission System (FACTS) power system. The system analyzes FACT’s execution sequences and is formalized using an automatic tool for establishing security properties. Timing is introduced to reveal classified information required in the decision making among the FACTS devices. Cardenas et al. [14] do a more indepth security investigation for CPS by considering potential consequences of attacks from CPS threats, they also compare security in CPS versus security in traditional information technology systems, and based on that discuss potential security solutions that can be applied to CPS. This provides very useful insights into security for CPS. Wang et al. [15] identify challenges and issues for CPS security and propose a context-aware security framework for general CPS but at the same time identify numerous research challenges, which need to be addressed. Other than security, trust within CPS is equally important to ensure that information from only trusted resources is utilized. This refers to the problem of filtering meaningful information from large volumes of CPS data since CPS data is naturally very noisy [16]. Tru-Alarm was developed by Tang et al. [16] to analyze trustworthiness in CPS by filtering out noises and false information efficiently and capturing meaningful alarms while avoiding false alarms. Tru-Alarm estimates the locations of objects causing alarms, performs trustworthiness inference and uses this information for future communication. Based on the literature investigation, we identified numerous research gaps that need urgent attention for QoS in CPS. We discuss these issues in the next section. III.

OPEN CHALLENGES

There are numerous problems with sensor-actuator communication networks in multi-protocol environment. Some of these problems are now discussed. Real time communication, real-time monitoring and decision-making are critical in most of the time-critical CPS. The latency issues with multi protocols, that exhibits different latencies and jitter variations can affect the real time communication especially during event detection and communication. Ensuring cross platform sensor-actuator communication is still an open challenge that requires detailed frameworks to address it properly. Sensor failure is the other most important issue; as if the sensor fails it would affect the entire CPS as sensors are the key inputs for sensing data. The failure of a sensor of one standard will affect to other sensors, and actuators on different standards. Thus considering redundancy in sensor deployment as well as other models for sensor robustness should be investigated. Excessive replication of sensors (operating on different protocol standards) for robustness or security can result in significant drop in energy efficiency, hence models that can ensure energy efficient in a multi-protocol

331


environment is very crucial. One of the key challenges in this context is developing a unified energy efficiency framework that is extensible across different standards as different standards exhibits different energy efficiencies. Ensuring energy efficiency and scalability at the same time is another unique challenge because different standards support different communication ranges and different number of smart connections, hence finding the overall scalability of the framework is difficult. Furthermore security, trust and data privacy are other emerging issues within CPS as ensuring high level of overall security and privacy will be a combined result of the security and privacy offered by different protocols. Hence analyzing security loopholes across different protocols as well as inter-protocol communication becomes a big challenge. Based on the challenges listed above, a new unified framework for Cyber-Physical Systems needs to be developed. This framework needs to provide end-to-end QoS for all components in the CPS – cyber components, physical components and communication components. We give a brief overview of this proposed CPS framework in the next section and explain how end-to-end QoS can be achieved, in-line with the challenges we have discussed here. IV.

CONCEPTUAL FRAMEWORK

Large CPS systems might have different QoS properties employing a different set of protocols and will need to be analyzed using different theories at each component. Due to the interaction between cyber entities and physical entities, existing QoS protocols may not be orthogonal and, sometimes, could have pathological interactions [17]. Therefore, new QoS protocols will need to be developed for both the physical object as well as the cyber object. We propose a Web-of-Things (WoT) framework for CPS [1] that provides mechanisms for end-to-end QoS for all CPS components, cyber components, communication components and physical components. This framework provides 5 layers – WoT API, WoT Context, WoT Overlay, WoT Kernel and WoT Device – for building CPS applications that seamlessly connects cyber systems to physical systems with the required QoS. Details of these layers are in [1]. Figure 1 shows the WoT CPS reference architecture, which consists of two components – a CPS Node and a CPS Fabric [1]. A CPS Node is an embedded computing unit that connects with various physical devices through the WoT Device layer. The CPS Node implements 3 layers of the proposed framework – WoT Overlay, WoT Kernel and WoT Device. The QoS requests are handled by the WoT Kernel layer at the CPS Node, useing an intelligent scheduler for prioritizing communication, event handling and tasks. The internal architecture for the CPS Node is shown in figure 2. The CPS Fabric facilitates communication amongst CPS Nodes, handles QoS for cyber components, derives passive context from processing event streams, performs intelligent control based on active context and integrates with existing decision making modules and other related computational functions. The CPS Fabric implements 3 layers of the proposed framework – WoT API, WoT Context and WoT Overlay. The internal architecture for the CPS Fabric is shown in figure 3.

ISBN: 978-1-4577-0872-5 (c) 2011 IEEE

CPS Node

CPS Node Actuators

WoT Overlay

WoT Overlay

WoT Kernel

WoT Kernel

WoT Device

WoT Device

CPS Node WoT Overlay WoT Kernel WoT Device

Physical Environment

CPS Fabric

CPS Event

WoT API

CPS Mashups

WoT Context WoT Overlay

Sensors CPS Desktops

CPS Node CPS Event

WoT Overlay WoT Kernel

CPS Users

WoT Device

CPS Developers

Figure 1. CPS reference architecture (reproduced from [1]).

A. QoS Management at the CPS Node The CPS Node provides a low-level run-time for communication, scheduling, and WoT resources management through the WoT Kernel layer. Any resource request will be routed directly to this layer, which resolves URIs to assess resources and manages the mapping between the URI address spaces of different resources. It dynamically allocates network bandwidth, processing power and storage capacity for dealing with a large amount of data from the CPS device (sensor/actuator). The kernel scheduler will allocate computing resources to working threads (events) with different priorities to meet end-to-end QoS requirement. For scalable pre-emptive scheduling, any incomplete jobs will be migrated from one computing unit to another using a code mobility mechanism [18] for dynamically changing the bindings between code fragments and the place where they are executed. The WoT Kernel is also responsible for detecting and identifying newly connected/disconnected physical devices and their associated resources which is crucial for self-configuration and plug-andplay. It will expose these resource specifications to the WoT Overlay layer for manipulation by the cyber object. Inbound REST Requests

Outbound REST Requests

WoT Overlay RESTful Interface

WoT Kernel

Request Scheduler

Controller

WoT Overlay

REST Client

Decision Maker

REST Server

Event Analyzer Event Channel

Inbound Event

Publisher Publisher Subscriber Proxy Proxy Proxy

Outbound Event

Internal Event

Publisher Proxy

Controller Component

Proxy Component

WoT Kernel

Plug-n-Play Component Framework Device Component

Device Component

Device Component

WoT Device

Device Connect Protocol device driver

device driver

device driver

Figure 2. Intelligent Scheduler in CPS Node (reproduced from [1]).

332


B. QoS Management at the CPS Fabric The CPS Fabric, through the WoT Overlay layer, provides an application-driven, network-aware logical abstraction atop the current Internet infrastructure to cope with volatile network behavior such as latency, data loss, jitter, bandwidth and so on by allowing selection of network resources with better and predictable performance on their own behalves. The CPS Fabric handles QoS requirements at the application (cyber) level and communication level. As CPS requires QoS for every component in the system – physical, cyber and communication – new QoS functionalities for physical and cyber objects will require modifications or replacement of current TCP/IP at a large scale, which is unfeasible given its ubiquitous presence as the backbone for today's Internet. Therefore, the WoT Overlay layer builds an application-aware virtual network on top of the existing IP-network based on the QoS setting configured in CPS applications. The interaction between WoT Overlay (Cyber and Communication Systems QoS) and WoT Kernel (Physical Systems QoS) is through an interface mapping from application specific QoS requirements to network QoS parameters supported by WoT Overlay. While previous overlay networks research has focused on smarter 'best effort' routing [19] and data loss guarantee and recovery [20], the WoT Overlay aims to (1) provide dynamic models that isolate the end-to-end delay on the virtual path from the variations in queuing delays at each underlying IP router, (2) provide multirouting, admission control, and on-network caching mechanism based on these models. The CPS Fabric provides QoS using the following components: •

Overlay Router component that uses multiple path delivery to ensure on-time delivery for time-critical RESTful messages.

•

Delay Analyser component that uses admission control to ensure end-to-end delays meets the CPS QoS requirement.

•

Net Connect component that packs/unpacks RESTful messages to and from the underlying networks as well as provide network-specific implementations of the interfaces defined in the QoS Monitor.

•

QoS Monitor component that obtains updated conditions of the underlying network using probes.

•

QoS Config component that tags RESTful messages with priority labels given application-specific requirements.

•

QoS-based Scheduler component that schedules incoming RESTful messages through a priority queue based on priority tags. V.

CASE STUDY

In this section we describe two case studies, 1) Smart Grids 2) Intelligent Vehicle Systems, to showcase how the proposed framework can be implemented in these two applications.

ISBN: 978-1-4577-0872-5 (c) 2011 IEEE

WoT Context

RESTful Messages RESTful RESTfulMessages Messages

QoS Networks

RESTful Stream Time Service

Stream Processing Cluster

Context Miner

QoS Config QoS-based Scheduler

external sources

Overlay Router Passive Context

Delay Analyzer

WoT API

Admission Controller

Active ActiveContext Context QoS Monitor

task list Context Engine

Node Manager

WoT Overlay

TCP/IP

Net Connect

GPRS / 3G

Ethernet

Figure 3. QoS components in CPS Fabric (reproduced from [1]).

A. Smart Grids With the growth of globalization and world’s population, the energy demand from the users to meet their day-to-day needs is constantly increasing. The majority of such energy demand is met by non-renewable sources of energy, and any shortages in meeting the energy demand are met by generating more energy. But with the increasing applications of carbon caps by most countries about the world, there is a need to meet the energy demand from renewable sources. Moreover research in the literature has shown that more than 90% of energy coming from the ground is wasted [21]. So a framework is needed by which energy can be produced and consumed efficiently and smartly. To achieve this, the concept of Smart Grid has been proposed in the literature. In order to realize such initiatives of Smart Grids, significant changes have to done to the existing traditional utility infrastructure and CPS has shown to be leading the way in this innovative exercise. The framework described in the previous section, shows how this can be realized and we now explain that in this section. A standard smart home comprises of numerous sensors like light intensity sensors, door and window sensors, flood detection sensors along with smart devices like energy monitors, etc. A standard smart home comprises of such sensors that communicate with a smart home controller, which is normally a coordinator that controls all the sensors in the home. A user can normally configure such sensors to meet their comfort requirements. However configuring a smart home with different sensors raises its own set of challenges i.e. communication across different sensors working on different standards as well as communication with the smartgrid. The framework proposed in this paper introduces the concept of WoT kernel that is specifically designed to address the cross platform communication challenge. WoT Kernel is a plug and play component framework that communicates and acts like a bridge between different communication standards. This assists in sensor actuator communication in the following way. When a sensor operating on a Zigbee standard wants to access another sensor (or actuator) operating on Z-Wave standard, the WoT kernel accepts requests from the Zigbee sensor and maps it directly to the Z-Wave request, so that the action can be completed. The WoT kernel ensures that the QoS is maintained in terms of latency and jitter. The WoT kernel also ensures that

333


data privacy and security is adhered to during this communication. A similar idea is now explained using another case study i.e. Intelligent Vehicle Systems. B. Congestion and Accident Avoidance using Intelligent Vehicle Systems Based on the CPS framework in section IV, an intelligent vehicle would constitute as a CPS that is comprised of a number of in-vehicle CPS Nodes and an in-vehicle CPS Fabric. The in-vehicle CPS Node would manage several sensors and actuators through the WoT Device layer. The in-vehicle CPS Node can also connect to the proprietary in-vehicle Controller Area Network (CAN) so that the CPS can directly control various physical systems in the vehicle, such as lights, navigation and other critical electrical control modules. The invehicle CPS Fabric provides an WoT Overlay atop of the underlying in-vehicle network to provide QoS, context processing and an interface between the driver and in-vehicle CPS. The new WAVE protocol [22] is being developed as a new communication protocol for inter-vehicle and vehicleinfrastructure communication. One of its primarily uses is for improving driver safety by providing a warning system for drivers to be aware of changes in the road environment [23]. For example, WAVE was proposed to warn drivers behind trucks on changes such as a vehicle braking in front or a vehicle approaching around a blind corner. In an intelligent vehicle system, CPS QoS requirements go beyond communication QoS as it also includes computational QoS and hardware QoS. If any one of these CPS components fails in providing QoS, the whole system will fail to respond in the required manner. For example, if a car suddenly brakes in front of a driver, the driver would only have a few seconds to stop the vehicle. If the sensor data captures the stopped vehicle event quickly, but the in-car CPS takes a long time to decide on a course of action, or if the braking system is activated too late, the car would hit the stopped vehicle in front. Our proposed WoT CPS framework intelligently handles these event communications through the CPS Fabric, taking into consideration internal vehicle sensor data and external communication data from other CPS sources. For example, if a driver wants to change lanes, the in-vehicle CPS Fabric take into consideration readings directly from sensors on the vehicle (side and rear sensors) as well as information from external sources (vehicles accelerating from the side). High priority events by external sources can be transmitted quickly to other vehicles using the WAVE Short Message Protocol [22]. The in-vehicle CPS will handle the event messages as a high priority queue at the CPS Node. This will be sent to the CPS Fabric and a decision will be made quickly by the processing unit as it is a high priority event. This event might interrupt other processors if its priority is of more important than other tasks. Based on the sensor information, the CPS will initiate immediate action (an audio warning) if the results show there is a danger to the vehicle. If the situation is safe, the driver can complete the lane change without any warning from the CPS. The in-vehicle CPS needs to identify and prioritize all events to minimize disruption to the driver. The successful

ISBN: 978-1-4577-0872-5 (c) 2011 IEEE

prioritization of events is crucial as it affects the QoS handling of all other CPS components. An intelligent vehicle, using the proposed CPS framework, can mitigate disruptions caused by internal factors by interactively and automatically using CPS node sensors, actuators, and other in-vehicle intelligent control systems. For example, the in-vehicle CPS Node sensors will continuously transmit vehicle data to the in-vehicle CPS Fabric. This data will be prioritized in the CPS Node and processes based on this priority queue at the CPS Fabric according to the level of importance of the event. If the vehicle is operated erratically and with slow driver response time, the CPS will infer that the driver is fatigued and incapable of driving the vehicle safely. The change in driver behavior will escalate the event priority, thus increasing the QoS of the event messaging, processing and action taken by the CPS. In this example, the in-vehicle CPS will query the navigation system for the nearest safe location to stop and recommend that the driver pull over and rest. The event data from CPS sensor nodes will be handled by the CPS kernel scheduler where urgent events will be processed first. An important event that could lead to an accident will be presented to the user for immediate action. The in-vehicle CPS will also assist the driver if possible, such as slowing down the vehicle where appropriate. As a vehicle in operation will receive many event notifications, displaying all this information to the driver can be distracting. Rather than warning the driver of all events the in-vehicle CPS is able to receive event messages, intelligently identify high priority events that can lead to immediate dangerous situations and automatically take action to mitigate the event. The proposed architecture has two schedulers, one for normal events (non-dangerous situations) and another for high priority events (dangerous situations that could lead to road accidents). High priority events that are generated invehicle, i.e. from sensors in the vehicle, will be handled first at the CPS node kernel layer putting other normal event processes on hold. These high priority events are then transmitted to the CPS fabric for further processing. The CPS fabric scheduler will interrupt regular processes when it receives a high priority event from external sources. The CPS fabric will also take into consideration user actions as this will further refine the decision-making process. Once the CPS Fabric had made its decision, it will instruct the in-vehicle CPS Nodes (using WoT Overlay) to handle the situation, either automatically (start slowing down the vehicle), manually (informing the driver of immediate danger), or semiautomatically where the CPS Fabric will inform the driver of the danger as well as take action to help mitigate the danger. If the vehicle could not evade a collision, the in-vehicle CPS sensors will detect the crumple zone (collision point) and the sudden reduction in vehicle speed. This event will be processed by the sensors as very high-priority, interrupting all other processes. Depending on the severity of the collision, the CPS Fabric will enter high priority mode and increase QoS capacity in the CPS to handle the tasks required to mitigate this situation. In this example, the CPS Fabric will automatically instruct other in-vehicle CPS Nodes to protect the vehicle occupants by deploying air-bags and calling for an ambulance if required.

334


VI.

CONCLUSION

The recent growth of Cyber-Physical Systems (CPS) provides a smart infrastructure connecting the abstract computational artifacts with the physical world. The solution to CPS must break the boundary between the cyber world and the physical world by providing a unified framework that permits integrated models addressing issues from both worlds simultaneously. However, building CPS is not a trivial task as existing Computer Science and Control Theory are independently developed based on overly-simplified assumptions of each other. In this paper, we discuss the requirements of quality of service (QoS) in CPS and the current state of the art in providing QoS for CPS applications. We proposed a Web-ofThings (WoT) CPS framework that will provide end-to-end QoS at the computational level (cyber), network level (communication) and hardware (physical) level. We then demonstrated how the WoT framework can be used to provide QoS for CPS application in electricity smart grids and intelligent vehicle systems. REFERENCES [1] [2]

[3] [4]

[5]

[6] [7] [8]

T. S. Dillon, H. Zhuge, C. Wu, J. Singh, and E. Chang, "Web-of-things framework for cyber–physical systems," Concurrency and Computation: Practice and Experience, Vol. (DOI: 10.1002/cpe.1629), No. 2010. President's Council of Advisors on Science and Technology (PCAST), "Leadership under challenge: Information technology r&d in a competitive world," http://www.nitrd.gov/pcast/reports/PCAST-NITFINAL.pdf August 2007. National Science Foundation, "Cyber-physical systems summit report," http://precise.seas.upenn.edu/events/iccps11/_doc/CPS_Summit_Report. pdf, Missouri, USA 24-25 April 2008. Technology Strategy Board, "Technology strategy board to invest £5m into the ’internet of things’," http://www.innovateuk.org/content/news/technology-strategy-board-toinvest-5m-into-the-in.ashx, Vol. No. 11 May 2011. Power Systems Energy Research Center, "Pserc awarded $5.5 million to map a new trajectory for the energy system grid," http://www.eurekalert.org/pub_releases/2011-04/asu-pa041911.php, Vol. No. 20 Apr. 2011. J. Singh, "Multicast qos routing using collaborative path exploration," in 24th IEEE International Conference on Advanced Information Networking and Applications (AINA), Australia, 2010, pp. 120-125. J. Singh, P. Veeraraghavan, and S. Singh, "Qos multicast routing using explore best path," Comput. Commun., Vol. 29, No. 15, 2006, pp. 28812894. L. Parolini, N. Tolia, B. Sinopoli, and B. H. Krogh, "A cyber-physical systems approach to energy management in data centers," in 1st

ISBN: 978-1-4577-0872-5 (c) 2011 IEEE

[9]

[10] [11] [12] [13] [14] [15]

[16]

[17]

[18] [19] [20]

[21] [22] [23]

ACM/IEEE International Conference on Cyber-Physical Systems, Sweden, 2010, pp. 168-177. M. D. Ilic, X. Le, U. A. Khan, and J. M. F. Moura, "Modeling future cyber-physical energy systems," in IEEE Power and Energy Society General Meeting - Conversion and Delivery of Electrical Energy in the 21st Century, USA, 2008, pp. 1-9. W. Jiang, G. Xiong, and X. Ding, "Energy-saving service scheduling for low-end cyber-physical systems," in 9th International Conference for Young Computer Scientists, China, 2008, pp. 1064-1069. A. A. Cardenas, S. Amin, and S. Sastry, "Secure control: Towards survivable cyber-physical systems," in 28th International Conference on Distributed Computing Systems Workshops, China, 2008, pp. 495-500. F. Xia, L. Ma, J. Dong, and Y. Sun, "Network qos management in cyber-physical systems," in International Conference on Embedded Software and Systems (ICESS), China, 2008, pp. 302-307. H. Tang and B. M. McMillin, "Security property violation in CPS through timing," in 28th International Conference on Distributed Computing Systems Workshops, China, 2008, pp. 519-524. A. Cardenas, S. Amin, B. Sinopoli, A. Giani, A. Perrig, and S. S. Sastry, "Challenges for securing cyber physical systems" in Workshop on Future Directions in Cyber-physical Systems Security, USA, 2009. E. K. Wang, Y. Ye, X. Xu, S. M. Yiu, L. C. K. Hui, and K. P. Chow, "Security issues and challenges for cyber physical system," in IEEE/ACM International Conference on Green Computing and Communications & International Conference on Cyber, Physical and Social Computing, China, 2010, pp. 733-738. L.-A. Tang, X. Yu, S. Kim, J. Han, C.-C. Hung, and W.-C. Peng, "Trualarm: Trustworthiness analysis of sensor networks in cyber-physical systems," in IEEE International Conference on Data Mining, Australia, 2010, pp. 1079-1084. L. Sha, S. Gopalakrishnan, X. Liu, and Q. Wang, "Cyber-physical systems: A new frontier " in IEEE International Conference on Sensor Networks, Ubiquitous, and Trustworthy Computing, Taiwan, 2008, pp. 1-9. A. Fuggetta and G. P. Picco, "Understanding code mobility," IEEE Transactions on Software Engineering, Vol. 24, No. 5, 1998, pp. 342 361. D. Andersen, H. Balakrishnan, F. Kaashoek, and R. Morris, "Resilient overlay networks," ACM SIGOPS Operating Systems Review Vol. 35, No. 5, 2001, p. 145. L. Subramanian, I. Stoica, H. Balakrishnan, and R. Katz, "Overqos: An overlay based architecture for enhancing internet qos," in 1st Symposium on Networked System Design and Implementation, USA, 2004. G. David, "Big goals get big results," in World Business Council for Sustainable Development (WBCSD) Liaison Delegates Meeting, Beijing, China, 2006. Intelligent Transportation Systems Committee, "Ieee trial-use standard for wireless access in vehicular environments (wave)-networking services," IEEE Std 1609.3-2007, Vol. No. 2007, pp. c1-87. R. A. Uzcátegui and G. Acosta-Marum, "Wave: A tutorial," IEEE Communications Magazine, Vol. 47, No. 5, 2009, pp. 126-133.

335