The paper presents the architecture and experimentation of a 10-Gb/s QoS-enabled almost-all-optical packet switching system (QOPS) for metro WDM networks. By applying cluster-based wavelength sharing and downsized single-staged optical buffers, QOPS is featured by its highly scalable and cost-effective design. In this paper, we first introduce the switch architecture, system operation, and the key techniques. We describe the in-band header/payload modulation and optical label swapping that is suitable for high-speed optical packet switching. We also present the design of the highly efficient Four-Wave Mixing wavelength converters for packet preemption. We then present an adaptive bifurcated routing (ABR) that directs same-connection packets to different switch clusters according to optimal bifurcation probabilities. Experimental and simulation results demonstrate that QOPS can achieve superior packet-loss performance, QoS differentiation, and minimize traffic blocking probability.

An optical header extraction and recognition structure was presented in the paper. The optical header is extracted by the improved iTOAD structure with DPSK modulation, while the header recognition is realized by discrete time stretch technology with super-continuum pulse to slow down the optical header speed for the high-speed asynchronous optical packet switch. Analysis and Experimental results shows that the iTOAD with time stretch technique can enlarge the output header 20dB and improve the switching ratio to 15dB when the input pulse is only 0.1pJ. By adjusting the tunable FDL length, the two pulses after the coupler switching have the π phase shift in the iTOAD SOA. The structure can slow down the high-speed optical header speed. Furthermore, the low speed header has the low latency time decreases (latency time less than 100μs) and overcomes the drawbacks of time bandwidth limitation associated with dispersion in long lengths of fiber. The system is highly scalable, lower power consumption and photonic integration.

A scheme using ahead-label to record packets' reaching time is proposed. The Principle of ahead-label and scheduling strategy for FDL is explained. With this scheme, the performance of slotted optical packet switching networks is significantly improved.

The suitability and efficiency of network architectures strongly depend on the service requirements and the mixture of services at different network levels. A case study has been performed in which five different service scenarios have been defined considering the characteristics of network operators. Starting with a service scenario which assumes only Layer 3 IP services, the amount of Layer 2 (Ethernet) and Layer 2 services (leased line) have been increased subsequently in the remaining four scenarios. Moreover, Layer 1 services have been differentiated with respect to broadband leased lines and narrowband services. The service scenarios have been applied to assess network architectures based on IP/MPLS, Ethernet and flexible and static optical systems, respectively. The impact of control plane implementations, i.e GMPLS and ASON, have been analyzed. Ethernet is making its way into metro and core networks. The development of Ethernet services and technology and their integration into MPLS/GMPLS based networks is still hampered by open issues. Evolutions in this area will be a key differentiator for network architecture evolutions.

A distributed flexible architecture of adaptive control and service provisioning founded on automatic switched optical network is proposed in this paper. The testbed named Adaptive Multi-Service Optical Network (AMSON) has been developed for proof of the fundamental of adaptive control for intelligentized service realization and lightpath transmission.

GMPLS-based Bandwidth on Demand (BoD) service experiment using the GMPLS network testbed of JGN II was conducted, and GMPLS connections with GbE bandwidth were successfully established by our developed user-friendly web-based GUI on the demand basis. The combination of the GMPLS-based network and the user-supporting system was verified to enable the BoD service in the actually operational environment. The setup time of the GMPLS-based lambda-path in this BoD service was very immediate and became available within only one minute from the time when the users want to use it.

Grid technology which provide a uniform interface to geographically distributed resources has been extensively developed in recently years. It is imperative to support QoS for Grid application, the main problems providing QoS in OBS network is exact understanding different Grid application's requirements and selecting the right policy to quantitatively offer them. In this paper we combine the Class Based Weight Fair Queuing and Preemption Latest Available Unused Channel with Void Filling (PLAUC-VF) and propose a absolute QoS architecture in OBS network for Grid application.

To reduce burst drop probability is a key issue in optical burst switching networks for it was connected strongly with the network's overall performance. In this paper, we propose a novel O/E/O conversion and electronic storage based strategy to resolve burst contention. Unlike traditional FDLs based contention resolution which is only able to provide discrete buffering time, O/E/O storage based strategy can offer unlimited buffering time. We propose an intermediate node architecture and corresponding scheduling algorithm called Nearest Available Void (NAV). Simulations show that our novel burst contention resolution enjoys high contention resolution capability and only introduce extremely low extra delay time compared to FDLs based mechanism. The maximum electronic RAM demand is also studied, and the result shows that the demand is comparatively small, which implies the cost of this strategy can be economic.

A main issue for the realization of Optical Packet Switched (OPS) networks is the resolution of packet contention due to the lack of RAM-like optical buffer. Although an optical buffer using Fiber Delay Line (FDL) has been proposed as alternative, its capacity is extremely limited. For this reason, several studies have been challenged on this problem. One approach is a deflection routing. It is widely used in electronic packet switched networks or optical burst switched (OBS) networks. However, in the OPS networks, packet length is short, thus, speed limitation for route look-up is significantly stringent. If the network topology is geometric such as a Manhattan Street Network (MSN), a hop-by-hop routing can be implemented by simple optical logic devices without an electronic routing table, but if the topology is non-geometric, it is hard to implement the deflection routing electronically or optically. Another approach is a reflection routing. It is easy to implement, but its packet loss probability is inferior to that of the deflection routing. In this paper, we propose a packet contention resolution scheme that is based on the reflection routing and furthermore enables to detour the reflected packet to alternate path if its primary path is under congestion. Our method enables to alleviate the time limitation for alternate path setting by using the packet reflection latency and to improve the performance of packet loss probability. We evaluate the performance of the proposed method to confirm its effectiveness.

This paper describes photonic network technologies from the viewpoint of optical switching functionality and photonic network transparency, and presents an analysis of the various types of current photonic nodes. The analysis results show that the most important choice is whether or not the photonic network node system is constructed in an optically transparent manner. This paper also describes various types of photonic node architecture. Regarding the optical add drop multiplexing (OADM) system, the merits of the reconfigurable OADM (ROADM) are precisely analyzed.

IP based traffic is seen as the convergence protocol for all future services and is already dominating today's networks. From the current situation with circuit switched SDH/SONET and packet switched IP backbones, an evolution is described towards long term optical transport network architectures featuring multi-layer transport with a novel Layer 2 transport service based on optical and opto-electronic burst/packet techniques as convergence layer providing the flexibility, scalability, reliability and Quality-of-Service to support all future higher layer services optimizing both CAPEX and OPEX. The operation of such a network will be mostly automatic supported by an integrated Control Plane on the basis of GMPLS protocols covering Layer 1, Layer 2 and Layer 3 (vertical integration of WDM, SDH, Ethernet, Bursts, IP etc.) and providing end-to-end control over various administrative and vendor domains (horizontal integration). The introduction of a new L2 transport service and its integration into a multi-layer Control Plane will be a long term evolutionary process also involving the necessary steps in the evolution of current standards. A few examples of network architectures and solutions are described together with dimensioning and control aspects and their relation to a roadmap for the introduction of optical burst/packet networks.

Recently, IP over WDM has been envisioned as dominant network architecture to meet the ever-increasing bandwidth requirement in the IP networks and to eliminate the intermediate layers and able to make better use of advanced optical technologies. Among existing several optical switching paradigms for IP over WDM, optical burst switching (OBS) is the more promising one, which combines the advantages of optical circuit switching (OCS) and optical packet switching (OPS), and can also loose the implementation complexity even though the optical random-access buffer is unavailable, and optical logic processing is not mature up to now. Therefore, OBS will play a very important role in next generation IP networks. However as much more trends shown, OBS-based IP optical network requires an appropriate control plane able to control and manage resource/service between IP and WDM. Currently, Generalized Multi-Protocol Label Switching (GMPLS) has been regarded as one of excellent candidate control planes for most of network scenarios because it enhances some issues of MPLS (including routing and signaling, link management, and fault protection and restoration), and also supports various granularity switching (such as packet, TDM timeslot, wavelength, waveband, and fibre-switched). Moreover, some practical deployments have shown GMPLS able to achieve automatic path provisioning and online network management. Therefore naturally, the integration and cooperation between GMPLS and OBS attracts more attentions from many research institutes and organizations. Of course, the similar concepts are originally given in the labeled OBS (LOBS) framework proposed by Dr. Qiao. Most of importance, LOBS provides the basic ideas for the collaboration of MPLS and OBS. However, how to efficiently integrate GMPLS and OBS to make the best of their advantages requires much more detailed discussions. To our knowledge, there are very few literatures and material to expatiate on the related issues. Here, the paper will present our some research results and viewpoints related to the integration and extension of GMPLS into OBS. Firstly, the paper proposes a GMPLS-based OBS network model for the integration of IP and WDM. Secondly, it defines the packets format for control and data information. Third, it also gives the node's functional architecture and the collaborative control operations between GMPLS and OBS. Finally, it proposes a unified software system to integrate the functions of GMPLS and OBS. Besides, the paper also presents some GMPLS enhancements to support OBS, such as traffic engineering, fault recovery, and link management, etc.

Many recent studies have convincingly demonstrated that network traffic exhibits a noticeable self-similar nature, and most studies of optical burst switching (OBS) networks are under a fundamental assumption that full wavelength conversion is available throughout the network. In practice, however, economic and technical considerations are likely to dictate a more limited and sparse deployment of wavelength converters in the optical network. Therefore, we present a novel scheme for OBS Networks, called logical cascaded private subnet with start wavelength assignment policy. We define the concept of canoe relative to cluster in self-similar traffic, and introduce a new device named payload segregator at the edge node as a gateway to the OBS core node. According to the changes in the edge node framework, we put forward the concept of cluster private subnet and canoe private subnet in OBS core node correspondingly. Due to the absence of wavelength conversion capabilities, we assign start wavelengths to both private subnets before network start operation. A new start wavelength assignment policy is proposed for the absence of wavelength conversion capabilities in the core node of OBS Networks. The analytical results clearly show that logical cascaded private subnet scheme can yield better performance in terms of block probability than traditional OBS Networks, and logical cascaded private subnet with start wavelength assignment policy has the potential to improve the block probability by more than one order of magnitude compared with traditional OBS Networks, helping reduce the performance gap with respect to full wavelength conversion.

In this paper, we study access protocol for multicasting in an optical burst switched (OBS) WDM ring environment. Ring nodes are equipped with a single fixed transmitter and tunable receiver (FTTR). Signaling is done via a dedicated control wavelength. The performance of the multicast access protocol is analyzed by simulation.

With fiber being cheaper and cheaper, the core fiber bandwidth utilization is not the key factor of the network, how to use the huge bandwidth flexibly and efficiently is one of the hot topics of optical networks. Optical burst switching (OBS) is the best choice to statistically use the optical resources, OBS based transparent optical network attracting attentions. We proposed an integrated policy by using dual fiber links server model along with traffic spacing mechanism (TSM) by address differentiating and transmitting those traffics between adjacent nodes by IP directly, the others by OBS mechanism. The theoretical results show that this method can improve network performance and be practically used in fields.

All-optical label switching (AOLS) is a potential technique which is introduced to fully use the bandwidth capacity in future optical networks and the packet operations are mostly carried out in the optical layer. The key issue in the AOLS approach is the method of coding the optical label onto the packet as it directly determines the structure and the performance of the optical core router as well as the channel bandwidth efficiency and the transmission quality of the packet and its label. With the bit rate increasing above 40Gb/s, new labeling scheme is needed to enhance the spectral efficiency. In this paper, recent advanced labeling schemes for future high speed AOLS network are reviewed, especially the new schemes demonstrated in Tsinghua University, China. Two novel labeling scheme based on vestigial sideband (VSB) modulation are mainly introduced both in principle and experiment. One is compact wavelength labeling scheme which achieves high spectral efficiency since the wasteful double sideband of 40Gb/s has been reduced to vestigial sideband. This scheme can be easily carried out by using optical filters. Another is polarization shift keying scheme with VSB payload which has high spectral efficiency near to 0.8bit/s/Hz. This scheme utilizes two polarization orientations to carry the label signal and causes little distortions between the label and payload. And this scheme is also independent of the payload bit rate which is a good character for future networks. Both the two schemes have little power penalties after long distance SMF transmission and are promising to be used in beyond 40Gb/s optical networks.

With the fast development of optical network technologies, for dealing with the shortcoming of the traditional grid based on the Internet, it is considered to introduce the optical network resource as one of grid resources in the control and management of grid. So the idea of optical grid was proposed. This paper analyzes the characteristics of optical grid differing from traditional grid. And then we propose a new transmission scheme applied in optical grid---multi-path transport scheme. We compare and analyze its performance in the optical grid. Eventually we draw the conclusions.

We propose a more practical resource model for optical grid applications by extending the classic list scheduling algorithm to achieve communication contention aware scheduling of optical grid applications. An effective adaptive routing scheme is also proposed to improve the performance of the extended list scheduling. The impacts of different routing schemes on the extended list scheduling are comparatively investigated by simulations.

Recently, some solutions (such as UCLP, GARA and GridJIT, etc.) have been proposed to manage optical network resources as other grid resources. However, most of them are focused on using grid technologies to encapsulate the management and control functions of optical control plane. And actually, they do not perform schedule optical network resources as grid resources. This paper proposes an economic-based scheme for lightpath resources scheduling and leveraging in grid-enabled optical networks. We also present the scheduling model and algorithm, and finally demonstrate their effectiveness by experiments on GridSim.

Optical Grid that provides a way to integrate the geographical distributed resources for large-scale grid applications. This paper mainly focuses on the task scheduling issue in Optical Grid and proposes two task scheduling algorithms to minimize the total time or scheduling length required for a given task. A simulated optical grid has been addressed to demonstrate the effectiveness of the scheduling approach. The performance tests and analysis of the two algorithms are investigated to show the effect of the routing policy on the scheduling length. A more flexible policy is able to further improve scheduling length.

Optical grid provides scheduled optical network resource to grid application. For supporting intensive data in optical grid, LPoD service architecture is presented to implement multiple light paths scheduling on demand under the driven of applications. In the proposed architecture, we introduce load division for intensive data or huge size file to cut bulk data into fractions, which will be transferred by different light paths only if the original sending node owns multiple communication interfaces or using multi-wavelength paths. About the type of the applications that runs on optical grid with LPoD on load division, we illustrate the parallel and distributed computing, and compare the total expensing time under different processing ability and transmitting distances. The result shows that for remote collaborating job in optical grid, LPoD as communication resource providing method can satisfy grid demand. When there are many available nodes, selecting the nodes with nearer distance and strong processing will obtain good transmitting quality from executing time.

In this paper, we propose OBS ring network structure and OBS access protocol for burst collision resolution. Our collision free scheme can save wavelengths for transmission and effectively transport bursty and asymmetric traffic. It shows better throughput performance than OBS ODD RR-R, RR-P, and RR-NP protocols.

An efficient fault detection method, employing cycle cover theory, is proposed. Three classic cycle finding algorithms are compared in terms of cost-saving and feasibility. We through computation and statistic find that our proposed method is economical and practical.

Optical Burst Switching (OBS) is supposed to be a promising switching technology in the future WDM networks, which is suitable to support burst traffic. A problem with OBS network is that bursts with longer routing lengths are more likely to be dropped before arriving at the destination router. In this paper, we propose two approaches using offset time differentiation (OTD) and burst length differentiation (BLD) to solve the burst loss differentiation (BLP) unfairness problem in OBS networks. Simulations show that BLP fairness in OBS is greatly enhanced with our approaches, especially with the OTD approach. Besides, we also investigate the BLP unfairness problem in offset time based QoS scheme.

In this paper, different kinds of optical performance monitoring (OPM) techniques, in particular the optical signal-to-noise ratio (OSNR) monitoring and chromatic dispersion (CD) monitoring, are reviewed. Then recent trend of multiple-parameter simultaneous monitoring is examined.

Among all kinds of metro-ring networks, resilient packet ring (RPR) technology possesses excellent market prospect which is integrated with many advantages, such as the intelligence of Internet Protocol (IP), economy of Ethernet and high bandwidth and reliability of SONET/SDH. A RPR network is a dual ring topology with two unidirectional counter-rotating ringlets. This ring network architecture and associated protocol has been standardized by the IEEE 802.17 working group in 2004. One of the main requirements for RPR is to provide fairness in bandwidth allocation for each node, which is ensured by the implementation of a fairness policy in a distributed manner. However, the existing fairness mechanisms suffer from severe oscillations under certain conditions, such as unbalanced traffic scenario and noticeable time delay. In this paper, an improvement to current aggressive fairness algorithm is proposed, especially for networks with noticeable time delay. And the simulation results show that the system performance can be sustained by adjusting the compensating factor accordingly with respect to the delay.

We study and test several P-Cycle generation algorithms from different strategies. The basic approach of P-Cycle algorithm is to identify an elementary cycle-set firstly, and then to search for improvements on this set through different strategies to create the final P-Cycle-set with higher efficiency. The P-Cycle generation algorithms which are presented in this paper are in terms of three aspects that are elementary cycle-set generation, expansion strategies and expansion times. At last simulation and analysis with these different algorithms will be presented.

We discuss a phenomenon of upstream bandwidth waste in a conventional dynamic bandwidth allocation scheme and propose a novel upstream bandwidth allocation scheme to enhance upstream bandwidth efficiency in Ethernet PON, and evaluate performance of the proposed scheme in terms of upstream channel throughput, mean queueing delay, and mean queue length of ONUs.

This paper presents the architectural designs and enabling technologies to implement the next generation metro WDM networks to satisfy the needs of emerging broadband applications.

Ethernet is a permanent success story, extending its reach from LAN and metro areas now also into core networks. 100 Gbit/s Ethernet will be the key enabler for a new generation of true end-to-end carrier grade Ethernet networks. This paper first focuses on functionality and standards required to enable carrier-grade Ethernet-based core networks and possible Ethernet backbone network architectures will be discussed. The second part then evaluates the CAPEX and OPEX performance of Ethernet core networks and competitive network architectures. The results propose that Ethernet will not only soon be mature enough for deployment in backbone networks but also provide huge cost advantages to providers. A novel complete, cost-effective and service-oriented infrastructure layer in the area of core networks will arise. The industry-wide efforts to cover remaining challenges also confirm this outlook.

The robust of the optical networks is one of the most important issues the optical networks should take into account when they are put into practice. Two typical causes for the instability of the WR-ION are analyzed in the paper: One is associated with something abnormal in the network and resulting in some measures for the survivability of the networks, and the other results from the dynamic characteristics of the network, such as dynamic routing. And some algorithms were proposed by Tsinghua Univ. to enhance the robust of the ION, for example, a Low Priority Hello Reduction Algorithm to decrease the number of false alarms resulting from LSA storm, and an Active Choke of Inter-domain Routing algorithm to overcome inter-domain routing instability. And some assistant methods were also developed, for example a Soft Preempt Algorithm to modify congestion degree of the ION in case congestions occur, and a software solution to decrease the processing time of the RSVP-TE. Experiment and simulation results proved the algorithms effective.

This paper describes the MPLS/GMPLS multi-layer network architecture and its control and management technologies. The recently investigated results are presented including GMPLS network migration into existing IP/MPLS network, MPLS/GMPLS multi-layer network management system and related standardization work.

An international field demonstration on lambda-based Internet eXchange (IX) was conducted between China and Japan for the first time, by using Generalized Multiple Protocol Label Switching (GMPLS)-controlled Photonic Cross-connects (PXCs) and Layer 2 switches (L2SWs). The IX operation could be demonstrated to provide the resilient lambda interconnection as well as L2 interconnection to the Internet Service Providers (ISPs). Stable lambda LSP establishment, dynamic bandwidth management and fast restoration were successfully achieved in the field trial. Traffic transmission of video stream between two ISPs via the IX was also verified.

A notification mechanism of topology information based on speaker node is proposed, which is aimed to realize topology discovery, synchronization and routing computation in hierarchical ASON networks. The simulation shows that with the increase of routing layers, the size of routing load and link state database are rapidly decreased. So the mechanism can save the memory space of the node and improve the routing efficiency.

With the overwhelming success of Ethernet technology in local access networks (LANs), owing to its ease of management and its much lower line rate cost, there has been a continuous effort in migrating Ethernet technology into metro-area networks (MANs) and wide-area networks (WANs). However, current Ethernet technology has to overcome the scalability and reliability deficiencies in order to be recognized as a viable MAN/WAN technology. Recently these concerns are being addressed under the framework of Generalized-MPLS (GMPLS) and its traffic-engineering (TE) intelligence. In this paper we propose an optical Ethernet solution viable for MAN/WAN based on the concept of a reconfigurable Ethernet node, integrating reconfigurable optical cross-connects (OXCs) with traditional Ethernet switches. We subsequently proposed a dynamic TE scheme that utilizes both the simplicity of all-optical switching and the flexibility of optical-electrical-optical (OEO) traffic grooming realized by the proposed optical Ethernet node. We showed through simulations that with our proposed TE scheme, the number of OEO transceivers needed by such an optical Ethernet node is only about 20% of a fully-populated one in the extreme case (very high load) and about only 10% in normal load conditions, without degrading the performance. This actually implies a tremendous cost saving on Ethernet OEO transceivers, and on the other hand indicates that a highly efficient and scalable MAN/WAN Ethernet network system could be well achieved by using Ethernet switches of only modest switching capacity (as legacy Ethernet switches have) along with reconfigurable OXCs. To demonstrate and test our proposed ideas we have built a GMPLS testbed using MEMS-based optical cross-bar switches and legacy Gigabit Ethernet switches. The testbed consists of three core nodes and six edge nodes. Performance experiments have been conducted on the testbed as well.

The Ethernet technoglogy rapidly gains importance as becoming a dominant solution for a converged transport network. Recently added features such as QoS handling and OAM functions in the framework of ITU-T NGN provides a means for performance improvement to meet carrier-class network requirements. In addition, integration with newly developed T-MPLS by ITU-T offers a potential to replace current SDH/SONET-based network infrasctructure.

IP+Optical networking technology is a key technology for realizing high-performance Internet backbone networks. Dynamic optical path cut through can create high-throughput as well as low power consumption networks. National Institute of Information and Communications Technology (NICT) Kei-han-na Info-Communication Open Lab, Interoperability Consortium, is shooting for the leading edge photonic Generalized Multi-protocol Label Switching (GMPLS) Inter-Carrier Interface that utilizes wide-bandwidth, cost-effective photonic technology to implement IPcentric managed networks. Kei-han-na Open Lab is a consortium for researching the GMPLS protocol and advancing a de facto standard in this area. Its experiments, new ideas, and protocols are submitted to standardization bodies such as International Telecommunications Union-Telecommunication standardization sector (ITU-T), Internet Engineering Task Force (IETF), and Optical Internetworking Forum (OIF). This paper details a nationwide GMPLS field trial that used multi-vender, multi-switching capable equipment.

Neighbor discovery in IP networks means the activity of identifying adjacent IP layers using the very Hello protocol. Neighbor discovery in optical networks differs from that in IP networks due to the GMPLS architecture that separates control and transport planes. The discovery of control plane is similar to that in IP networks from a perspective of basic mechanism although message flows are different each other. However, the discovery of transport plane has a different situation due to the physical characteristic of optical fibers that lead to identify adjacent peer-to-peer layers using OAM facility. This paper, based on the standardization technology of relevant bodies, defines functional requirements and its structure for neighbor discovery in optical networks. In addition, the paper presents a verification course applied to the proposed structure using a modeling tool, called SPIN.

Wireless sensor networks (WSN) provides a global view of the monitors area based on local observations measured by each sensor. In this paper a novel solution of WSN in GMPLS and EPON based integrated Network Access Architecture is proposed. We emphasize on the Routing of data aggregation for WSN in GMPLS (Generalized Multi-protocol Label Switching) which is developed from MPLS and is an inevitable result of MPLS extension to optical networks. By the use of control plane we can employ the GMPLS on the optical communication network and control the optical transport network. By combining the advantages provided by EPON and GMPLS, the proposed scheme can realize fast soft handoff, paging and expected QoS request of subscribers quickly and efficiently. The architecture that extends GMPLS and EPON-based access network to support wireless access is presented in this paper.

We present control plane architecture which is utilized to form a grid infrastructure enabled ASON and execute optical resource scheduling for grid community. ASON network is a good mode to solve the bandwidth deficient or unbalance problem in current grid service and implant large scale intelligent and automatic switching on optical layer, which make the communication resources evenly allocating and sharing in grid community for increasing resource utilizing efficiency. Grid services show the distributed and parallel attribution, which are being employed in many researching areas, such as parallel and distributed computing and processing. Thus, to construct a realizable and flexible ASON network infrastructure and support grid applications, the present control architecture will provide the parallel service, introducing automatic adjacent discovery and resource parallel reservation mechanism. At last, we test a grid file transfer by ASON mechanism in a test system by two parallel light paths for different file slices and validate the feasibility.

ASON is new type of optical network which can offer the flexibility for bandwidth allocation and fast provision of the new service. However, until now there still are many non-ASON optical networks. It is a challenging issue to reconstruct them into ASON smoothly. In this paper, we propose a new reconstructing technology. This technology performs the function of the control plane such as link management, path management and interface management by an embedded top-set-box (ETSB). The ETSB is located between EMS and the traditional NE such as DWDM, WDM, SDH, etc. By this technology the hardware module of the existing optical network element (NE) need not modification. Thus, compared with other reconstructing technologies, the cost and the risk of this reconstructing technology are smaller.

Because any link failure on the multicast session may disrupt the signal transmission to several downstream nodes, it is imperative to effectively protect the multicast sessions against link failure. In this paper, we study schemes for establishing primary and backup multicast trees for multicast sessions in a optical mesh network against link failure, e.g., fiber or conduit cut, with taking the different leaf availability requirement and availability guarantee into account. We propose an Mixed Integer Linear Program (MILP) approach to solve the problem of provisioning survivable multicast sessions with differentiated leaf availability guaranteed. Different from the previous schemes, our scheme finds out the complete (the tree spans the root and all the leaf nodes) primary and complete or incomplete (the tree spans the root and part of the leaf nodes) backup trees for multicast sessions guarantees the availability requirement. Because of taking the different leaf availability requirements into ccount, the proposed algorithm finds out more efficient feasible solution requiring less network resources, and even find a feasible solution where the previous algorithms fail.

The survivability is one of the important issues in Resilient Packet Ring (RPR) networks. This article analyzes a new failure scenario and put forward a new restoration mechanism named as C-steering. The mathematical model has been established to investigate the survivability of the RPR network in three operation methods. The throughput-delay performance and link utilization are discussed theoretically and experimentally.

Segment protection is an intermediate protection scheme between path and link protection. The major advantage of segment protection comes from its limited rerouting scope. It then leads to faster recovery over path protection and greater resource utilization efficiency over link protection. Overlapped Segment Shared Protection allows a protection of all nodes and links in addition to bandwidth savings. Routing for Overlapped Segment Shared Protection consists of finding, for each user request, a working path and a set of backup segments to protect it while minimizing the total required bandwidth. The problem in multi-domain optical mesh networks has not received a lot of interest so far although it is more difficult than in single-domain network. The main difficulty lies in the lack of the complete and global knowledge of the network topology as well as the bandwidth allocation meanwhile this knowledge are assumed to be easily available in single-domain networks. To overcome this difficulty, we propose a two-step routing heuristic using aggregate routing information. In the first step, a rough routing solution is sketched out on a compact virtual network which relies on an aggregate topology of the multi-domain network. Then the routing is refined within the different domains using the original domain networks. The comparison of the proposed heuristic with the optimal single-domain solution1 shows that our heuristic leads to resource efficient solutions that are close to the optimal ones while it requires a quite small computational effort. We also find that long overlapping does not necessarily leads to waste of resources and more segments in comparison with short overlapping.

This paper presents traffic analysis of Frame Bursting (FB) in Single-Input Single-Output (SISO) IEEE 802.3z (Gigabit Ethernet, GbE) Networks. This analysis characterizes the Frame bursting traffic behavior and performance measurement, shows its implications for future research issues. GbE is one of the world's largest Enterprise LAN solutions in terms of the number of concurrent users. FB Mechanism improves the performance of GbE CSMA/CD networks when transmitting short frames by reducing the Extension Field (EF) overhead, and tries to accord the controversy over the cross-layer (Layer2/Layer3) scheme. But FB get some negative effect like Jumbo Frame, long burst will affect the equal opportunity of other transmission. We simulate about 1 Tera-packets by random generation, let the packet pass through form non-FB domain to FB domain, observe that FB behavior like the burst per 100 packets (BPHP), packet per burst(PPB) and its distribution, the burst length in different Input Rate, Input Channel, and Packet Length Distribution Input, The analysis of the burst performance reveals that the geometrical proportion between burst size and burst length, and the inverse proportion between the percentage of Burst and the output utilization after bursting. We find that there is a linear relationship between Input Rate and the burst behavior. As for the burst behavior is highly dynamic complex in variable Number of Inputs, and Packet Length Distribution types.

Our proposed soft preemptive scheme was demonstrated highly effective at reducing the blocking probability and meanwhile guaranteeing the existing traffics' quality of service in wavelength-routed networks. In this paper, we focus our study on combining Soft Preemption and Wavelength Conversion. The performance analysis shows that the introduction of Wavelength Conversion can help enhance the advantages of Soft Preemption in reducing blocking and improving other performance metrics.

The backup paths just SRLG (Shared Risk Link Group)-disjointed with working paths can be concentrated on a specific link. Thus, the idle resource is not used efficiently so that the resource utilization of the network is not so good. We propose the resource share backup path provisioning mechanism to guarantee the survivability and to utilize the network resource. We consider SDLG (Sub-Domain Link Group) as the information to eliminate the link accommodated the unbalanced sharing backup paths. We compare our mechanism with the existing mechanism, which allocate the backup path just SRLG-disjointed with working path. We also apply the diversity network plans to the mechanism to select the backup paths. In addition, to control the backup paths we use GMPLS (Generalized Multi Protocol Label Switching) according to the consensus, which is emerging in the industry on utilizing an IP-centric control plane within optical networks to support dynamic provisioning.

Waveband switching (WBS) is proposed as a promising technique to reduce the number of switching ports, which could translate to the switching fabric size and cost of the nodes in the dense wavelength division multiplexing (DWDM) network. We propose a routing and wavelength/waveband (RWA) heuristic algorithm, called Maximize the Number of Waveband Route (MNWR) for the dynamic arriving traffic demands in the multi-granularity WBS optical network. We study the performance of various MG-OXC architectures including the serial three-stage MG-OXC and the interconnecting three-stage MG-OXC proposed by our group employing the on-line MNWR algorithm. Simulation results indicate that the interconnecting three-stage MG-OXC outperforms the serial three-stage architecture in terms of blocking performance for dynamic requests.

OBGP is a distributed protocol, which can provide an inter-domain routing /signaling solution and allow the edge network customer to establish and manage a lightpath though multi-heterogeneous domains. In this paper, we propose a new scheme called DiffServ-aware OBGP by extending the OBGP. It can resolve the open problems of customer access admission in domain and the conflicting resource requests by distinguishing the priority of customer requests. The paper depicted the basic extension requirements for current OBGP message and the end-to-end lightpath establishment process at length. The simulation shows that the new scheme can establish adaptive end-to-end lightpath according to the request of customer and the higher priority requests have more probability of establishing lightpath when resource competition happened.

We developed the optical switching technologies of 3-dimentional micro-electro mechanical system switch, acousto-optic tunable filter, PLZT electro-optics deflector switch, and optical parametrically amplified fiber switch. We evaluated their potentials for application to photonic networks based on their optical switch characteristics.

RSVP-TE Signaling involves transporting control messages between all entities communicating through an ASON's control plane and is used to create, maintain, restore, and release connections. Control plane restoration has not been studied in details in the literature except for using protection switching or the source-initiation mechanism. Such schemes take a big bandwidth waste or long restoration to the networks. We provide a fast restoration mechanism based on (extended) local restoration using RSVP-TE as the signaling. This mechanism can reduce the restoration time by the theoretical analyses.

Today, as networks grow rapidly in size, hierarchical network structures (e.g. PNNI) have been proposed to solve the problem of network scalability as well as security in large networks. One of the most important requirements of hierarchical networks is survivability. In this paper, we first present a survey on some restoration schemes, and then propose a hierarchical hybrid restoration scheme (HHRS) for hierarchical networks. The performance of these restoration schemes in hierarchical networks is analyzed and compared theoretically. Restoration success rate, restoration time, fault notification message overheads are the major performance measures. HHRS can reduce the number of fault notification signal messages by sending notification signal messages vertically up and down the network hierarchy instead of horizontally along the paths. It can also maximize the restoration success rate with a fast restoration speed by dynamically determining the restoration scope.

The restoration time is an important network survivability index in large-scale Metropolitan Area Network. The early study indicates that the hierarchical restoration mechanism have lower restoration time, comparing with global restoration. But the restoration time increases rapidly when the restoration scope expands. In this paper, Overlap-extend hierarchical restoration mechanism is proposed to reduce the restoration time, by restricting the restoration in a definite scope. A theory model is established to describe the average restoration time of Overlap-extend mechanism and existing hierarchical restoration mechanism, which can be used to analyse the influencing factors in our mechanism. Simulation result is also presented, which shows that overlap-extend mechanism can observably reduce restoration time, especially in light traffic load.

This paper focuses on the survival mechanisms against single link failure for optical QoS services in ASON (Automatically Switched Optical Network). The new protection and restoration methods and optical QoS services enabled by ASON also bring new requirements. Correspondingly, the component functions in ASON control plane, such as Connections Controller, Link Resource manager and Routing Controller are extended. According to the survival level of the QoS service requested, we divide the traffic into five types, viz. 1+1, 1:1, shared-mesh, rerouting and extra traffic. Optical QoS services and resource preemption are realized in our simulation platform. Extensive simulations are performed concerning failure localization time, failure recovery time, recovery success ratio and network resource utilization.