Abstract— Recent advances in wireless sensor network (WSN) and radio frequency identifier (RFID) have made it possible to extend current human-to-human communication to the future unified communication environment among human society, computer network and the external physical world. A ubiquitous hierarchical generalized-sensor network (UHGSN) is presented in this paper as a typical network model for such unified communication architecture. The sensor information processing unit and the hierarchical distributed agent server are introduced as the basic network elements in this architecture for effective information communication, and the topology, addressing and recovery problems are analyzed. As the basic applications, the combined “key word” based characterized searching mechanism and the performance monitoring processes are discussed, along with the corresponding protocol message definition and encoding format. Simulation results show the advancements of the presented hierarchical architecture and the job-list based characterized searching mechanism. Index Terms—ubiquitous hierarchical generalized-sensor network (UHGSN), WSN, characterized searching, RFID
I. INTRODUCTION The success of data communication and computer
technologies has brought us into the information society in which the thirst for both information itself and the information exchange becomes our basic and even instinctive requirements. Due to the rapid developments of mobile communication and the Internet, people have already been able to communicate with each other freely regardless of the location or the time. Furthermore, the emergence of various novel data services and intelligent terminals enables us to enjoy a comprehensive human-to-human communication. However, in order to get enough information required
for the daily lives, people need not only to frequently communicate with others but also to keep contacting with the external physical world, so the current human-to-human communication is not satisfying, but a much wider communication environment among human society,
computer network and the external physical world is required for the future unified communication [1-2]. This purpose becomes increasingly realizable since the recent advances in wireless sensor network (WSN) [3-4] and radio frequency identifier (RFID) [5] have made it possible to acquire huge amount of information from the external physical world. Many novel concepts emerge in such circumstance including M2M communication [6], ubiquitous network and internet of things (IOT) [7-8], which are all focusing on extending the communication scopes to the external physical world with more intelligent control and management.
The M2M communication stands for machine-to-machine communication, and may also be defined as a more general concept which includes man-to-man, man-to-machine and machine-to-machine communication. The fundamental technologies for M2M communication include hardware realization, middleware design, and communication protocols. The ubiquitous network is defined as a novel kind of network with full ability to realize the communication among anyone and anything at anytime and anywhere. The IOT makes us possible to interact with the environment around us and to receive information on its status that was previously not available to use [9]. Current researches on these subjects are mainly focusing on the mechanisms for smart node design, energy saving, and event representation. There are still many questions that should be addressed for the complete implementation of future unified communication environment in terms of network architecture, addressing and routing, and information aggregation, however, to the best of our knowledge, the research on these issues is not thorough enough. In [10], a wide area ubiquitous network is presented as the future sensor network architecture from the view of the network operators. In such network, ubiquitous small networked computers, including sensors and actuators, will enable us to be networked anywhere and anytime with anybody and anything to monitor and immediately manage the situation and status of individual objects and persons through the use of computerized information processing and storage. However, the paper only discussed the wireless terminal implementation and the system design without details on the top architecture. In [11], a hierarchical context monitoring and composition
Manuscript received April 28, 2011. This research was supported bythe National 973 Program (No. 2011CB302702), National 863 Program(No. 2009AA01A345), P. R. China. The corresponding author isZhitong Huang: [email protected].
576 JOURNAL OF NETWORKS, VOL. 7, NO. 3, MARCH 2012
framework is proposed to realize the next-generation context-aware services but the paper concentrates only on the hierarchical abstraction on network data and contexts and there is no specific analysis on the supporting communication techniques. In [12], we presented a performance monitoring mechanism which uses different sensors to monitor the external physical world for active fault alarm on the optical transport networks, but in that paper we have not presented a complete architecture for sensor network. Therefore, a general architecture which specifies the information processing operations in future unified communication network is still need to be studied.
The basic function of the unified communication network is to collect valuable information from both the human society and physical world, which can be used in economic analysis, academic research and our daily lives. Currently, the information in the Internet is mainly represented by character and text, so the key-word based context searching is the fundamental technology for information acquisition. However, with the growing demand on various novel digital services, information will be represented not only by text, but increasingly by audio, image, video and even multi-media, which makes the traditional searching mechanisms inadequate. One transitional and compatible solution suggests attaching a content-describing text label to each piece of audio, image and video, but it is inefficient because the semantic description is always incomplete and inaccurate. The optimal solution lies on the research of novel information searching technology based on combined “key word” of text, audio, image and video, which is called characterized searching here. The searching efficiency not only associates with the performance of such searching algorithms themselves, but also depends on the consistency between these algorithms and the characteristics of the unified communication network, such as its architecture, routing, addressing. Another basic application of this communication network is to monitor the status on the external physical world to realize some intelligent control. How to realize the performance monitoring on the current transport network is an interesting topic that need to be studied.
In this paper, a ubiquitous hierarchical generalized-sensor network (UHGSN) is presented as a typical network model for the future unified communication environment. The remaining of the paper is described as follows. Section Ⅱ describes the recent advances in the sensor network technologies to introduce the concept of generalized-sensor, and the function of UHGSN for future unified communication is analyzed. Section Ⅲ discusses the network architecture, basic elements, topology, addressing, recovery and functional planes for the implementation of UHGSN. Section Ⅳ describes the characterized searching procedure in the UHGSN as the basic application, along with the searching message definition and encoding format. Section Ⅴ discusses the performance monitoring principle and processes in this architecture, along with the message definition and encoding format. In section Ⅵ, a UHGSN simulation platform is constructed for performance comparison, and
the simulation results show the advantages of the presented hierarchical architecture and the job-list based characterized searching mechanism. Section Ⅶ concludes the paper.
II. UHGSN AND FUTURE UNIFIED COMMUNICATION ENVIRONMENT
The priority of building unified communication environment is to introduce the information connectivity of the external physical world into current communication network, which can be accomplished by various sensing technologies. The WSN integrates traditional sensor techniques with the wireless communication, which has received much popularity nowadays because of its flexibility and simple structure. The fiber sensor network (FSN) has also received much attention as a wired sensor technology due to its sensitivity on pressure and temperature. Besides, the RFID network may be treated as a wireless passive sensor network which is mainly used for object tracking and transportation. In [13], the authors discussed that the conventional sensor network communication model assumed the deployment of low-cost, multifunctional sensor nodes operating on limited power capacity of their batteries, which cannot be recharged due to the dense and random WSN deployment, but the external radio frequency power stands as a promising source for WSN. The paper discussed the implementation of such technology. Moreover, the camera monitoring network that widely deployed in the traffic and security area can also be considered as a multi-media sensor network (MSN), since it captures and processes data in the forms of image and video. The combination of these multiform sensor networks builds a ubiquitous generalized-sensor network which can be used to capture comprehensive information from the external physical world. In the future, different kinds of terminals that can get any form of information from the external world can be included as a kind of generalized-sensor.
In such network, different types of sensors are distributed in the external physical world to capture valuable information. For efficient management, these sensors should be divided into different domains according to the category and position, and the sensors in each domain share the same data format, communication protocol and functional modules. On the other hand, the amount of the original data captured by these sensors will be quite huge, and it is improper to use one centralized server to realize the data analysis, aggregation and further applications, thus the data and the associated information should be managed in different layers based on the location and network scale. Therefore a hierarchical framework UHGSN is constructed, which can be divided into two parts, the ubiquitous generalized-sensors distributed in the physical world to capture valuable data, and a hierarchical information management network (HIMN) which accomplishes the information collection, translation, analysis, aggregation, and transmission. The integration of UHGSN with current communication
JOURNAL OF NETWORKS, VOL. 7, NO. 3, MARCH 2012 577
network forms a unified communication environment, whose architecture is described in Fig. 1.
Figure 1: The future unified communication environment
In this symmetrical architecture, the human society and
the external physical world on the two sides are not only the sources of the information, but also the end “users” of the information. The computer network, along with the associated access devices, provides information connectivity within the human society, between the human society and the physical world, and even within the physical world in the future. The human society uses various human-oriented intelligent terminals to access the computer network, such as mobile phone, personal data assistant (PDA), or directly a personal computer with human-machine interfaces. The physical world uses ubiquitous generalized-sensors to access the computer network such as RFID, camera, recorder and traditional sensors. The developments of intelligent terminals and the generalized-sensors are both driven by the emergence of new data services, applications and new forms of information carriers, and both limited by the cost and simplicity. The Internet and the HIMN constitute the core computer network, and they both faces the same backbone network problems such as addressing, routing, scalability, and survivability. Current telecommunication network operators may extend to operate the whole computer network and the associated access devices in the future.
III. UHGSN ARCHITECTURE AND IMPLEMENTATION Current researches related to UHGSN mainly focus on
the sensor technologies such as lifetime control and energy saving, distribution and localization, information representation and design of novel sensor, whereas this letter concentrates on the research of the HIMN. The HIMN is composed of two kinds of network elements, the sensor information processing unit (SIPU) and the distributed agent server (DAS). Because of the limits on lifetime and node size, the original data captured in each sensor domain are not recommended to be operated by the sensors themselves but by one SIPU which collects the data, translates them into valuable information, and accomplishes further operations on such information including searching and analysis. Toward different sensor domains, the interfaces between SIPUs and the sensors are different, so as to the format of the original data, but the translated information will share the same format. With the large-scale deployment of WSNs, RFIDs, MSNs, and other types of generalized-sensor networks, the amount of original data and the corresponding
information will be quite huge, so it is important to design proper architecture for efficient information communication and management.
The peer-to-peer (P2P) technology has received increasingly popularity in local and even metro area network as an efficient communication structure since it matches the requirements of network load balancing and network resource equalization, and research has been presented on extending this idea to all of the future intelligent terminals to build ubiquitous P2P network [14]. However, the precondition of using P2P technology is the high intelligence of the terminal devices and the active requirements of information exchange, but currently, the basic purpose of building UHGSN is to provide the information of external physical world to the human society, so the P2P structure is improper. In HIMN, we introduce the conventional client/server based network structure. Multi-layer DASs are deployed for aggregating the information, handling the requests from the human users and transferring the corresponding information back to the users. The network scale, along with the locations and numbers of SIPUs, decides the layer number for DASs required in the HIMN. Fig. 2 presents a typical model of a simple three-layer UHGSN where layer 1(L1) to layer 3(L3) DAS is responsible for the information communication in access, metro, and backbone network respectively. The users of UHGSN are preferred (but not required) to send their requests to the nearest L1 DAS through the Internet, as people are always most interested in the things around themselves. If the users try to get some information that far from themselves, the layer 2 (L2) or even the L3 DAS may be involved in the communication. The information in the HIMN is being converged and aggregated, so the SIPUs are connected to the bottom layer L1 DASs in tree topology, so as to the connections between different layers of DASs. All of the DASs at the top layer connect in a mesh network for information exchange.
The internet protocol (IP) based network addresses (i.e. IPv4/IPv6) should be allocated to the SIPUs and DASs to realize the communication between HIMN and the Internet. Such IP address only identifies who the host is without defining where the host is, however, the services in UHGSN is primarily based on the physical world location information of the SIPUs, so each SIPU needs another physical world localization address(e.g., GPS address) for user application. A network address mapping list (NAML) should be created which records and maintains the relationship between these two types of addresses. The user of UHGSN need only provide the physical world address in the request to the DAS, and the DAS will use the NAML stored in its database to translate it into the IP address for the following message transfer. This NAML-based scheme is a transitional approach and may be substituted by novel addressing mechanism for future computer network which contains the information of both terminal identifier and location [15]. The DASs in different layer may keep different records of NAMLs in their databases which contains all of the
578 JOURNAL OF NETWORKS, VOL. 7, NO. 3, MARCH 2012
The above analysis is focusing on the network issues in the HIMN regardless of the bottom network infrastructure and the top network management. Strictly speaking, the HIMN is only the service plane of UHGSN and a network management system (NMS) which realizes the configuration, performance, fault and safety management is required as the management plane. To realize the long-haul communication with certain quality of service (QoS) requirements, both the generalized multi-protocol label switching (GMPLS) control plane [12] and the optical fiber network transport plane are recommended to be deployed, whose functions are as same as those in current backbone networks such as rapid routing and provisioning, flexible protection and restoration and dynamic resource allocation. Fig. 2 shows these four functional planes in the UHGSN architecture.
The network survivability issue has already become one of the central requirements for current network, since a single link in the network can carry so huge amount of data that even a short period of link failure may lead to an excessive loss of information which might be on the order of terabits per second. The GMPLS has received much attention nowadays due to its dynamic provisioning of bandwidth guaranteed paths with flexible recovery capability. There are basically two categories of mechanisms for network recovery: protection, a proactive procedure in which spare capacity is reserved during the path establishment, and restoration, a reactive procedure in which the resource allocation will only be done, if any, after a fault’s occurrence for rerouting the disrupted paths. Protection is attractive for its exact availability of the pre-provisioned backup paths, leading to especially simple and fast automatic switching operations in both control plane and transport plane, but with the disadvantage of the requirement of redundant capacity for protection paths, thus the low resource efficiency. Restoration, on the other hand, can achieve good resource efficiency but may be slow because of its on-line rerouting calculation after the fault’s occurrence. In this UHGSN architecture, in order to guarantee the performance of the network, some backup SIPUs and DASs nodes should be deployed in the network for protecting the node failure situations, and the important links in the network should also be protected by using
1+1 protection or 1:N protection. On the other hand, the research of efficient restoration mechanisms in the UHGSN is our future work.
IV. CHARACTERIZED SEARCHING IN UHGSN The basic application for the users of UHGSN is to
search out valuable information from the huge amount of data captured by the sensors. Duo to the growing demands on various digital services, along with the large-scale deployments of MSNs, information in the UHGSN will be represented not only by text, but increasingly by audio, image, video and multi-media. Therefore it is necessary to develop novel characterized searching mechanisms and algorithms based on different kinds of “key word”. The searching requirements from the users can be quite different, leading to different realization procedures. Here we define a standard format of user searching request message in UHGSN.
< UHGSN user searching request message> ::= <Hierarchical physical world localization address> <Characterized searching “key word”> <User QoS requirements>
The <Hierarchical physical world localization
address> object provides the physical world localization addresses for the associated SIPUs involved in this request, which will be then translated to the network addresses using the NAML in DAS. The translated information also indicates that which layer of the DAS should handle this request and whether it is a uni-cast, multicast or broadcast request. Every SIPU and DAS has its own address sequence number. The DAS translates the address information in the <Hierarchical physical world localization address> object, and makes a logic and calculation to decide which layer of DAS should handle this request. The <Characterized searching “key word“> object shows the type, length, and value (TLV) of the searching “key word”. The type may indicate the following “key word” is a piece of text, image or video. The value of “key word” is provides by the user. Such information will then transfer to the SIPUs to finish the characterized searching algorithms. The <User QoS requirements>object presents the user QoS requirements, such as the operation delay, protection and restoration, the results format, etc. This object may also contain several sub-TLVs to describe different kinds of QoS requirements. For example, if a user requires all the real-time information the MSN captured in a certain position, he may generate a searching request with the “minimal operation delay” requirements. Fig. 3 presents a typical encoding format for the user searching request message in UHGSN.
In practical applications, the DAS may receive repetitive searching requests from different users when there is something popular happens in the real world. To avoid the unnecessary repeated searching operations in SIPUs, the DAS should create a job-list which records the latest searching requests and the corresponding results it have processed. When a new request arrives, the DAS
JOURNAL OF NETWORKS, VOL. 7, NO. 3, MARCH 2012 579
searches it in its job-list. If it is a reduplicate request, the DAS will directly send the associated results to the users. This job-list based searching mechanism will enhance the searching efficiency and reduce the searching overhead in the repetitive requests situation. The length of the job-list may influence the searching efficiency, so it should be set according to the network scale, user number and the location of the DAS. Fig. 4 shows the implementation of characterized searching for such a standard searching request in the UHGSN.
Figure 3: Typical encoding format of the searching message
Figure 4: The implementation of characterized searching
V. PERFORMANCE MONITORING IN THE UHGSN Besides the passive response to searching requests,
another primary application of such unified communication network is the active monitoring on the external physical world. The monitoring technology has
been widely utilized in various areas including industry and agriculture, traffic and transportation, security and defense. According to the attributes of the monitoring object, monitoring can be divided into different types such as performance monitoring, fault monitoring, security monitoring. Conventional monitoring is accomplished by different types of monitors or apparatuses which focus on special internal performance parameters in the monitored system, but the large-scale deployment of the wireless sensor and RFID networks significantly extends the monitoring fields. Currently, much attention has been paid on how to realize the event detection and representation in different wireless sensor networks [16], yet the research on a general information monitoring mechanism coincided to the architecture of the unified communication network is also necessary. The users can send their monitoring requests to the UHGSN using the similar messages as the above searching message and realize intelligent control and performance monitoring. On the other hand, the wireless sensor and RFIDs
deployed in the external physical world can also be used to monitoring the performance of the transport network itself. Currently, the optical fiber transport network is the basic network infrastructure, so it is valuable to develop some mechanisms to monitor the performance of the optical network, which is called optical performance monitoring (OPM) [17]. It is essential for managing high capacity optical transmission and switching systems. Examples of functions that require OPM include amplifier control, channel identification and signal health assessment. Ultra-long haul and optically switched networks promise improved operations, reduced footprint and cost. However, these benefits come with the added complication required by managing transparent networks and have stimulated interest in OPM for enhanced fault management applications. The need for fault management capability in the physical layer is also driven by transmission requirements for very high bit rate systems. While clear advantages have been identified for increased transparency and bit rates, these trends place tighter constraints on the engineering rules and transmission margins. OPM is a potential mechanism for relieving this tension both through improved control of transmission and physical layer fault management. New optical layer functionality such as dynamic reconfiguration and link level restoration also introduce a level of complexity that may require advanced OPM capabilities. All of these issues bring focus to OPM as an enabling technology for next generation optical networks. An underlying theme in OPM is the notion that for
transparent networks, one might envisage service level agreements certified through OPM. As performance monitoring and QoS measures migrate into the physical layer, OPM might be used to realize new methods of managing traffic. Routing decisions based upon OPM is one possibility. High capacity and priority traffic can be dynamically tuned to high-performance optical channels. Interactions between OPM and higher-level element management systems (EMS) and network management
580 JOURNAL OF NETWORKS, VOL. 7, NO. 3, MARCH 2012
systems (NMS) become a critical issue. Questions arise such as what information should be passed around the network in order to keep the network management scalable. In optical transport systems, the physical layer performance is more closely tied to fault management and control applications than digital QoS applications. Control of optical transmission is the first key application area for OPM. There is obviously a gray area between monitoring in physical layer control systems to maintain healthy operation and physical layer performance monitoring. Traditional electronic performance monitoring also finds use in control and fault management applications in addition to QoS, so it is not surprising that OPM is used across these application areas as well. The above performance monitoring mechanisms focus
on the internal performance parameter, and there is another way for monitoring which pays attention to the external factors. Most of the data plane faults are caused by the external physical factors such as natural factors, weather factors, and man-made factors. These factors can be mainly divided into two categories. The first category contains the factors related to the impact of force, such as flood, shaking, and man-made damages. The other category contains the factors related to the impact of heat such as fire, high-temperature, and frost. The former impact can be represented by the parameter of intensity of pressure and measured by different kinds of pressure sensors or distortion sensors, and the latter impact can be represented by the parameter of temperature and measured by the heat sensors or temperature sensors. Other complex physical impact can be considered, which would be similarly represented by different parameters and measured by special sensors. Many sensor technologies can be used for performance monitoring, not only the traditional sensors like mercurial sensor for heat or elastomeric sensor for pressure, but also some advanced ones such as fiber sensor and wireless sensor network (WSN). The fiber sensor technology is especially recommended to be used because the next-generation transport network is almost based on the fiber communication, which will be convenient to integrate with such sensors. In the UHGSN, lots of generalized-sensors are distributed in the external world that may capture different information on the transport network. There is always an interim from the normal state to the
failed state for a data plane component. In this period, the performance of the component becomes increasingly worse, but not failed. This is called low-quality. According to the information captured from these sensors, if a node finds a network component has already become low-quality and the corresponding parameter value is still changing toward the failure direction, it should start the protection algorithms in advance. In this way, the network robust can be improved and this is effective in the UHGSN architecture since it contains so much information. For the implementation of such performance monitoring
in the UHGSN, a novel performance notification mechanism is presented here which is based on the
GMPLS routing protocol. Corresponding to the types of the external physical factors and sensors mentioned above, four concepts should be defined, the node stress parameter (NSP), the node temperature parameter (NTP), the link stress parameter (LSP), and the link temperature parameter (LTP). The former two are used to represent the physical impact on equipment and the latter two are used to represent the physical impact on fiber links of the data plane. More parameters can be defined if other physical impact is considered. For the extension of the routing protocols, take the open shortest path first protocol (OSPF) here as an example. Four sub-TLV (Type-Length-Value) would be added into the Link TLV of the opaque link statement advertisement (LSA). A typical encoding format for the extension is defined in Fig. 5. The information of the data plane performance and faults would be put into such new TLV, and then flooded throughout the control plane by the OSPF routing protocol. Such mechanism can be further extended to describe more complex performance factors in the UHGSN in order to realize more comprehensive monitoring.
Figure 5: Typical encoding format of the OSPF extension
VI. SIMULATION RESULTS AND PERFORMANCE ANALYSIS
To evaluate the performance of the hierarchical DASs in the UHGSN architecture and the job-list based characterized searching mechanisms, we construct a UHGSN simulation environment which contains one NMS, 20 user clients, 5 DAS L3 nodes, 20 DAS L2 nodes, 60 DAS L1 nodes and 120 SIPU nodes. All the nodes are simulated in servers connected in a local area network by a switch. The communication among the clients, the different layers of DASs, the SIPUs and the NMS are based on our user-defined user datagram protocol (UDP) messages. The simulation environment is shown in Fig. 6.
JOURNAL OF NETWORKS, VOL. 7, NO. 3, MARCH 2012 581
The first simulation is to analyze the performance of the hierarchical DASs in the UHGSN architecture. We compare the characterized searching time and the searching overhead in four kinds of network structures. In the first structure, no DAS is deployed, and both the clients and SIPUs are directly connected to the NMS. The NMS is responsible for the characterized searching operations. The 5 DAS L3 nodes are added in the second structure, and similarly, the 20 DAS L2 and 60 DAS L1 nodes are added in the third and fourth network structures respectively (as shown in Fig. 2). In the simulation, the clients send searching requests randomly to the 120 SIPUs. The searching time is measured as the time from the instant that the first client sends out its first searching request to the instant that the last client receives the last searching result message. The searching overhead is measured by the number of the communication messages required throughout the network in the whole searching procedure. Fig. 7 shows that the NMS-based centralized structure performs better when there are few searching requests, but the hierarchical DAS based distributed structures become increasingly efficient when there are more requests arrive. The one-layer, two-layer, and three-layer hierarchical DAS deployed structure performs faster than the NMS-based structure when the number of the searching requests reaches 700, 800, and 1100 respectively. When the number of the requests becomes large enough, the three-layer DASs deployed structure has the best performance. For example, when there are 1900 requests, the NMS-based structure needs 18750 ms to finish the searching operation, but the one-layer, two-layer, and three-layer hierarchical DAS deployed structure only needs 17030, 13840, and 12260 ms respectively. On the other hand, Fig. 8 shows the searching overhead in the hierarchical structure is larger than the centralized structure since it needs some more communication messages among different layers of DASs.
The second simulation is to analyze the performance of the presented job-list based searching mechanism in the repetitive requests situation among three kinds of mechanisms in the three-layer UHGSN. In the first mechanism, no job-list is created and one searching procedure should be generated for each searching request. In the second mechanism, DASs in different layers are
deployed with the unified job-list which maintains the searching requests and the corresponding results it have processed in the latest 3000ms. In the third mechanism, DASs in L1, L2, and L3 are deployed with different job-lists which maintain the searching requests and the corresponding results in the latest 3000ms, 5000ms, and 8000ms respectively. In the simulation, the clients send searching requests to the 120 SIPUs controlled by pseudo-random code generation software in NMS to create the repetitive requests cases. Fig. 9 and Fig. 10 show the results that the different job-list based mechanism performs best among these three mechanisms as it needs the shortest searching time and generates least searching overhead. For example, when there are 1800 searching requests in the network (only averagely 8% are the repetitive requests), the different job-list based mechanism is 3000ms faster than the standard mechanism, and needs less 2500 communication messages.
In conclusion, the simulation results show that, the centralized NMS-based structure is preferred to be used if there are many SIPUs but only little service. When both the number of SIPUs and the service quantity are large enough, the presented UHGSN architecture will have better performance. Therefore, it is recommended to be deployed in the future large-scale unified communication environment. For application, the job-list based characterized searching mechanism will perform better than the original standard searching procedures.
Figure 7: Comparison on searching time among different structures
Figure 8: Comparison on operation overhead among different structures
582 JOURNAL OF NETWORKS, VOL. 7, NO. 3, MARCH 2012
Figure 9: Comparison on searching time among different searching
Figure 10: Comparison on operation overhead among different
searching
VII. CONCLUSION In this paper, the UHGSN architecture is presented as a typical network model for the future unified communication environment. Two kinds of network elements, the SIPU and the hierarchical DAS are introduced in this architecture for the effective information communication. The combined “key word” based characterized searching mechanism and the performance monitoring processes are presented as the basic applications in UHGSN, along with the corresponding protocol message definition and encoding format for implementation. Simulation results show the advancements of the presented hierarchical architecture and the job-list based characterized searching mechanism.
REFERENCES
[1] J.Pan, S.Paul, R.Jain, “A survey of the research on future internet architecture”, IEEE Commun. Mag., Vol. 49, no 7, pp.26-36, July. 2011.
[2] P. Stuckmann, R. Zimmermann, “European research on future Internet design”, IEEE Wireless Communi. Mag., vol. 16, no 5,pp. 14-22, May. 2009.
[3] S.M.George, Zhou Wei, H. Chenji, et al, “DistressNet: a wireless ad hoc and sensor network architecture for situation management in disaster response”, IEEE Commun. Mag., Vol. 48, no 3, pp.128-136, Mar. 2010.
[4] S. Hong, D. Kim, M. Ha, “SNAIL: an IP-based wireless sensor network approach to the internet of things”, IEEE
Wireless Communi. Mag., vol. 17, no 6, pp. 34-42, Jun. 2010.
[5] E.Welbourne, L.Battle, G.Cole, “Building the Internet of Things Using RFID: The RFID Ecosystem Experience”, IEEE Internet Computing Mag., vol. 13, no 3, pp. 48-55, Mar. 2009.
[6] R. Lu, X. Li, X. Liang, “GRS: The green, reliability, and security of emerging machine to machine communications”, IEEE Commun. Mag., Vol. 49, no 4, pp.28-35, Apri. 2011.
[7] N. Koshizuka, K. Sakamura, “Ubiquitous ID: Standards for Ubiquitous Computing and the Internet of Things”, IEEE Perva. Comput., Vol. 9, no 4, pp.98-101, Apr. 2010.
[8] H. Ning, Z. Wang, “Future Internet of Things Architecture: Like Mankind Neural System or Social Organization Framework?”, IEEE Communi. lett., Vol. 15, no 4, pp.461-463, Apr. 2011.
[9] M. Zorzi, A. Gluhak, S. Lange, A. Bassi, “From today’s intranet of things to a future internet of things: a wireless- and mobility-related view,” IEEE Magazine Wireless Communi. , vol. 17, pp. 44-51, Dec. 2010.
[10] H. Saito, O. Kagami, M. Umehira, et al, “Wide area ubiquitous network: the network operator’s view of a sensor network”, IEEE Commun. Mag., Vol. 46, no 12, pp. 112-120, Dec. 2008.
[11] K. Cho, I. Hwang, S. Kang, et al, “HiCon: A hierarchical context monitoring and composition framework for next-generation context- aware services”, IEEE Network Mag., vol. 22, no 4, pp. 34-42, July 2008.
[12] Z. Huang, Y. Ji, Y. Lu, et al, “Sensor-based performance monitoring mechanism in GMPLS-controlled networks”, IEEE Commun. Lett.., vol. 12, no 4, pp.325-327, April 2008.
[13] O. Akan, M. Isik, B. Baykal, “Wireless passive sensor networks”, IEEE Commun. Mag., vol. 47, no 8, pp. 92-99, Aug. 2009.
[14] D. Braun, J. Buford, R. fish, et al, “UP2P: A peer-to-peer overlay architecture for ubiquitous communications and networking”, IEEE Commun. Mag., Vol. 46, no 12, pp. 32-39, Dec. 2008.
[15] H. Li, C. Peng, B. Li, et al, “ User ID routing architecture”, IEEE Vehi. Tech. Mag., Vol. 5, no1, pp. 62-69, March 2010.
[16] M. Sifalakis, M. Fry, D. Hutchison, “Event detection and correlation for network environments”, IEEE J. Selected Areas in Commun., vol. 28, no. 1, pp.60-68, Jan. 2010.
[17] D.C.Kilper, R.Bach, J. Blumenthal, et al. “Optical Performance Monitoring”, IEEE J.Light.Tech. vol. 22, no. 1, pp. 294-304, Jan. 2004.
Zhitong Huang, received his Ph. D in 2008 in School of Electronic Engineering at Beijing University of Posts and Telecommunications, Beijing, P.R.China. He is now a lecturer of Beijing University of Posts and Telecommunications. His research interests are in the areas of future network and sensor network technologies.
Yuefeng Ji, Ph. D. He is now a professor and Ph.D supervisor of Beijing University of Posts and Telecommunications (BUPT). He is vice president of the academic committee of Key Laboratory of Information Photonics and Optical Communications (BUPT), Ministry of Education, P.R.China. He is an IEEE member, and his research interests are in the areas of optical fiber communication and broadband communication networks.
JOURNAL OF NETWORKS, VOL. 7, NO. 3, MARCH 2012 583
