Search

Optical data center networks are becoming an increasingly promising solution to solve the bottlenecks faced by electrical networks, such as low transmission bandwidth, high wiring complexity, and unaffordable power consumption. However, the optical circuit switching (OCS) network is not flexible enough to carry the traffic burst while the optical packet switching (OPS) network cannot solve the packet contention in an efficient way. To this end, an improved switching strategy named OPS with multi-hop Negative Acknowledgement (MPNACK) is proposed. This scheme uses a feedback mechanism, rather than the buffering structure, to handle the optical packet contention. The collided packet is treated as a NACK packet and sent back to the source server. When the sender receives this NACK packet, it knows a collision happens in the transmission path and a retransmission procedure is triggered. Overall, the OPS-NACK scheme enables a reliable transmission in the buffer-less optical network. Furthermore, with this scheme, the expensive and energy-hungry elements, optical or electrical buffers, can be removed from the optical interconnects, thus a more scalable and cost-efficient network can be constructed for cloud computing data centers.

Intra-datacenter traffic is growing more than 20% a year. In typical datacenters, many racks/pods including servers are interconnected via multi-tier electrical switches. The electrical switches necessitate power-consuming optical-to- electrical (OE) and electrical-to-optical (EO) conversion, the power consumption of which increases with traffic. To overcome this problem, optical switches that eliminate costly OE and EO conversion and enable low power consumption switching are being investigated. There are two major requirements for the optical switch. First, it must have a high port count to construct reduced tier intra-datacenter networks. Second, switching speed must be short enough that most of the traffic load can be offloaded from electrical switches. Among various optical switches, we focus on those based on arrayed-waveguide gratings (AWGs), since the AWG is a passive device with minimal power consumption. We previously proposed a high-port-count optical switch architecture that utilizes tunable lasers, route-and-combine switches, and wavelength-routing switches comprised of couplers, erbium-doped fiber amplifiers (EDFAs), and AWGs. We employed conventional external cavity lasers whose wavelength-tuning speed was slower than 100 ms. In this paper, we demonstrate a large-scale optical switch that offers fast wavelength routing. We construct a 720×720 optical switch using recently developed lasers whose wavelength-tuning period is below 460 μs. We evaluate the switching time via bit-error-ratio measurements and achieve 470-μs switching time (includes 10-μs guard time to handle EDFA surge). To best of our knowledge, this is the first demonstration of such a large-scale optical switch with practical switching time.

When built, quantum repeaters will allow the distribution of entangled quantum states across large distances, playing a vital part in many proposed quantum technologies. Enabling multiple users to connect through the same network will be key to their real-world deployment. Previous work on repeater technologies has focussed only on simple entanglment production, without considering the issues of resource scarcity and competition that necessarily arise in a network setting. In this paper we simulated a thirteen-node network with up to five flows sharing different parts of the network, measuring the total throughput and fairness for each case. Our results suggest that the Internet-like approach of statistical multiplexing use of a congested link gives the highest aggregate throughput. Time division multiplexing and buffer space multiplexing were slightly less effective, but all three schemes allow the sum of multiple flows to substantially exceed that of any one flow, improving over circuit switching by taking advantage of resources that are forced to remain idle in circuit switching. All three schemes proved to have excellent fairness. The high performance, fairness and simplicity of implementation support a recommendation of statistical multiplexing for shared quantum repeater networks.

We propose a dynamic RWA method for optical path networks that pre-computes route candidates to attain better load-balancing. Numerical experiments elucidate that the proposed method can accommodate 1.3–1.8 times more traffic than the conventional method that ignores load-balancing.

Layout sizing in the mask preparation step is commonly used on critical layers to improve printing process windows. The current flow of yield calculation based on critical areas of a layout typically does not take into account this sizing step. In this paper, we propose a new and simple flow that accounts for the sizing to improve the accuracy of yield calculation. We compare the calculated results using the current and sizing-aware flow to demonstrate the differences. We also show that results from the sizing-aware flow better match the calculated circuit switching power, which already takes into account this sizing step.

A novel adaptive assembly algorithm called Cross-counter Balance Adaptive Assembly Period (CBAAP) is proposed in this paper. The major difference between CBAAP and other adaptive assembly algorithms is that the threshold of CBAAP can be dynamically adjusted according to the cross counter and step length value. In terms of assembly period and the burst loss probability, we compare the performance of CBAAP with those of three typical algorithms FAP (Fixed Assembly Period), FBL (Fixed Burst Length) and MBMAP (Min-Burst length-Max-Assembly-Period) in the simulation part. The simulation results demonstrate the effectiveness of our algorithm.

This work proposes a new wavelength converted broadcast-selective (WCBS) buffer structure. It can deliver each packet to different delays, which enables flexible management of the buffered packet flows. By rearranging the packets leaving queue, the highest priority packets can leave first and in preemption process without being affected by the former queue. Furthermore, we use the WCBS buffer structure to construct an OPS node in output-buffered pattern. The packet loss and delay performances are studied under different traffic loads at different WCBS dimensions. The result shows that the performance of high priority packets is improved effectively.

We introduce a compact, low-power electro-optical switch on a silicon-on-insulator circuit through heterogeneous integration. A 10µm diameter III-V microdisk cavity is employed as the switching element. Switching of a 10Gbps optical signal is demonstrated by sweeping the bias between −1.1V and +0.9V with 15dB extinction ratio and 1.2ns switching speed.

Grid is a resource management technology which provides infrastructures to execute applications. Grid manages resources and provides an infrastructure composed of computing, storage and other resources. While Grid assumes existence of network which connects resources, network used to be considered as a non-manageable resource. However, in these several years, an idea and implementations to use bandwidth of network as a manageable resource have been proposed and demonstrated. Using this kind of schemes together with optical circuit (path) switching technology, a high-performance, energy-efficient and low-cost IT infrastructure can be realized. In this paper, the G-lambda project and its experiments, in which lambdas (optical paths) are reserved and used by requests from applications and middleware, are introduced.

Hybrid optical switching (HOS) combines the advantages of optical circuit switching (OCS) and optical burst switching (OBS). But HOS can not dynamically change the optical switching paradigm according to the traffic fluctuation. it can not efficiently support burst traffic. And the utilization of some lightpath in HOS networks is low. To solve the problems, a self-adaptive hybrid optical switching (S-HOS) is proposed in the paper. In S-HOS networks, ingress edge nodes monitor the traffic between a pair of an ingress node and an egress node, and based on the monitor results edge nodes self-adaptively adjust optical switching paradigm. Based on minimum burst length maximum assembly period algorithm, a new assembly algorithm of S-HOS is proposed to improve the lightpath utilization and efficiently support burst IP traffic. And a data channel scheduling is proposed to assign data channel resources for the lightpaths and data bursts by regarding lightpath as a huge data burst. And it can decrease the complexity of hybrid optical switching and improve the network performance. Simulation results show that S-HOS outperforms OBS in term of IP packet dropping probability and end-to-end delay.

In this paper, the architecture of the optical networks based on Time-Space Label (TSL) Switching is described in detail, including Time-Space Label and Time-Space Routing algorithm. The switching mode is more flexible and scalable. The labels can be changed to fit for different kinds of optical switching technologies, such as Optical Circuit Switching (OCS), Optical Burst Switching (OBS) and Optical Packet Switching (OPS). The TSL switching has a very wide range of applications in the optical networks. In the existing signaling protocols of OBS, the routing and the signaling are separated. However, TSL switching combines both of them together. Moreover, the resource reservations are realized in a two-dimension Time-Space plane, and the blocking probability is reduced greatly. The mature electronic processing technologies and the high-speed optical transport are cooperated effectively in the TSL switching. On the other hand, Multi-Granularity (MG) Switching can groom the vast bypass traffic in the optical networks effectively, with which the core nodes are simplified, and the throughput is increased significantly. When MG switching is combined with TSL switching, the performance of the entire network will be improved greatly. A test system built for validating the MG switching based on TSL is described, and the results show the switching performance of different granularities correctly.© (2007) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

This paper introduces a novel notion of multi-granularity and multi-connectivity circuit switched network. Based on this notion, four routing schemes - Fixed Routing (FR), Maximum Remain (MR), Secured Maximum Remain (SMR) and Premium/Punishment Modification (PPM) are proposed. Numerical simulation results about the performance of these four schemes are also presented in this paper.

Wavelength-routed networks have received enormous attention due to the fact that they are relatively simple to implement and implicitly offer Quality of Service (QoS) guarantees. However, they suffer from a bandwidth inefficiency problem and require complex Routing and Wavelength Assignment (RWA). Most attempts to address the above issues exploit the joint use of WDM and TDM technologies. The resultant TDM-based wavelength-routed networks partition the wavelength bandwidth into fixed-length time slots organized as a fixed-length frame. Multiple connections can thus time-share a wavelength and the grooming of their traffic leads to better bandwidth utilization. The capability of switching in both wavelength and time domains in such networks also mitigates the RWA problem. However, TMD-based wavelength-routed networks work in synchronous mode and strict synchronization among all network nodes is required. Global synchronization for all-optical networks which operate at extremely high speed is technically challenging, and deploying an optical synchronizer for each wavelength involves considerable cost. An Optical Slotted Circuit Switching (OSCS) architecture is proposed in this paper. In an OSCS network, slotted circuits are created to better utilize the wavelength bandwidth than in classic wavelength-routed networks. The operation of the protocol is such as to avoid the need for global synchronization required by TDM-based wavelength-routed networks.

Due to the recent outstanding advancement of optical networking technology, pervasive Grid computing will be a feasible option in the near future. As Grid infrastructure, optical networks must be able to handle different Grid traffic patterns with various traffic characteristics as well as different QoS requirements. With current optical switching technology, optical circuit switching is suitable for data-intensive Grid applications while optical burst switching is suitable to submit small Grid jobs. However, there would be high bandwidth short-lived traffic in some emerging Grid applications such as multimedia editing. This kind of traffic couldn't be well supported by both OCS and conventional OBS because of considerable path setup delay and bandwidth waste in OCS and inherent loss in OBS. Quasi-Circuit OBS (QCOBS) is proposed in this paper to address this challenge, providing one-way reserved, nearly lossless, instant provisioned wavelength service in OBS networks. Simulation results show that QCOBS achieves lossless transmission at low and moderate loads, and very low loss probability at high loads with proper guard time configuration.

Optical Burst Switching (OBS) is a switching paradigm to efficiently support the ever-growing broadband multimedia traffic over optical networks which has been proposed with the consideration that future optical networks should provide efficient support for bursty traffic. Although research in OBS networks has evolved from theoretical investigations to proof-of-concept demonstrations, several key issues need to be investigated further before OBS prototypes can clearly outperformed Optical Circuit Switched (OCS) networks. In OBS networks, data loss may occur when bursts contend for network resources. There have been several proposed solutions to resolve contention. Most contention resolution techniques only react to the contention, but do not share the contention information. Based on the presented offset time determination technologies, we propose a novel dynamic offset time mechanism which concerns how to adjust offset time at source. In this mechanism, the intermediate node will send a signal to the source to notice the contention. Source will adjust the offset time in different strategies according to the situation of the network. Numerical results demonstrate that the proposed mechanism effectively reduces the contention at the expense of the increase the packet average end-to-end delay.

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.

An automatic synthesis method of equivalent circuits of integrated circuit devices is described in the paper. This method is based on a physical approach to construction of finite-difference approximation to basic equations of semiconductor device physics. It allows to synthesize compact equivalent circuits of different devices automatically as alternative to, for example, sufficiently formal BSIM2 and BSIM3 models used in circuit simulation programs of SPICE type. The method is one of possible variants of general methodology for automatic synthesis of compact equivalent circuits of almost arbitrary devices and circuit-type structures of micro- and nanoelecronics [1]. The method is easily extended in the case of necessity to account thermal effects in integrated circuits. It was shown that its application would be especially perspective for analysis of integrated circuit fragments as a whole and for identification of significant collective physical effects, including parasitic effects in VLSI and ULSI. In the paper the examples illustrating possibilities of the method for automatic synthesis of compact equivalent circuits of some of semiconductor devices and integrated circuit devices are considered. Special attention is given to examples of integrated circuit devices for coarse grids of spatial discretization (less than 10 nodes).

The paper presents some improvements to the transmission line model of photacoustic Helmholtz cell. The improvements were aimed to obtain much better agreement between theoretical and practical values of Q-factors of the cells, and were based on measurements of number of cells and comparison of measurement results with theoretical frequency responses. The measured cells had dimensions typical for photoacoustic setups.

Optical packet switching is commonly considered as a possible technology for future telecommunication networks, due to its compatibility with bursty traffic, eg Internet protocol (IP), and efficient use of wavelength channels. Current transport networks are voice-optimised and connection oriented, however the amount of data traffic is rapidly increasing, resulting in a continuous increase of average traffic through major exchanges exceeding 30% per annum (in Europe). Thus optical packet switching is seen as a future technology that will support diverse traffic profiles and give more efficient bandwidth utilisation through its ability to provide multiplexing at the packet level. In recent years the significance of optical packet switching as an emerging technology has been identified and researched by a number of research groups. Earlier optical packet switching demonstrators presented switching of mainly ATM compatible synchronously transmitted packets at bit rates up to 2.5b/s with the optical header encoded either in series or in parallel to the payload using the sub-carrier modulation technique. More recent projects have demonstrated switching capabilities at 10Gb/s using more advanced approaches with special encoding schemes for header and header detection, together with sophisticated control mechanisms for contention resolution. The capability of switching optical packets at bit rates up to 160Gb/s has recently been demonstrated. This paper discusses the architectures currently proposed for high speed optical packet switching, including the key techniques of header processing and payload switching. The focus is on a high speed demonstrator [OPSnet] capable of operation at rates >100 Gb/s.

A kind of optical coarse packet switching based on optical label routing is intrduced. In the switching, the swapping granularity is coarse compared to common optical packet switching, and the header of optical packet is labeled by multi-wavelengths. Some realization techniques on core switching node and edge node with terminator are studied. A principle experiment system is established to perform the optical coarse packet transmitting, switching, and receiving, and verify the feasibility of such optical coarse packet switching.

Recent development of ultra high-speed optical code-label processing and its application to packet switching are reported. 10 to 40Gbit/s interface, scalable, optical packet switch prototype based on optical code-label processing is developed. Optical packet switching with 200Gchip/s all-optical label processing, 10 to 40Gbit/s/port variable data rate packet switching, and optical buffering to avoid packet collisions is experimentally demonstrated. A novel packet bit error rate (BER) and loss real-time measurement method is proposed. The first 40Gbit/s packet BER and loss measurement system is shown. 40Gbit/s BER and loss evaluation with various conditions is experimentally demonstrated. In real time, only the payload part of packet and burst stream with fluctuated interval time is evaluated. BER and packet loss of randomly received packet sequence because of routing and buffering can be also evaluated by the proposed system. 10Gbit/s packet BER and loss measurement with optical label switching, buffering, and preamble free optical packet 3R are experimentally demonstrated.

The virtually unlimited bandwidth of optical fibers has caused a great increase in data transmission speed over the past decade and, hence, stimulated high-demand multimedia services such as distance learning, video-conferencing and peer to peer applications. For this reason data traffic is exceeding telephony traffic, and this trend is driving the convergence of telecommunications and computer communications. In this scenario Internet Protocol (IP) is becoming the dominant protocol for any traffic, shifting the attention of the network designers from a circuit switching approach to a packet switching approach. A role of paramount importance in packet switching networks is played by the router that must implement the functionalities to set up and maintain the inter-nodal communications. The main functionalities a router must implement are routing, forwarding, switching, synchronization, contention resolution, and buffering. Nowadays, opto-electronic conversion is still required at each network node to process the incoming signal before routing that to the right output port. However, when the single channel bit rate increases beyond electronic speed limit, Optical Time Division Multiplexing (OTDM) becomes a forced choice, and all-optical processing must be performed to extract the information from the incoming packet. In this paper enabling techniques for ultra-fast all-optical network will be addressed. First a 160 Gbit/s complete transmission system will be considered. As enabling technique, an overview for all-optical logics will be discussed and experimental results will be presented using a particular reconfigurable NOLM based on Self-Phase-Modulation (SPM) or Cross-Phase-Modulation (XPM). Finally, a rough experiment on label extraction, all-optical switching and packet forwarding is shown.

Recent developments in Micro Electro Mechanical Systems (MEMS), promises to solve several technological limitations that have plagued the wireless electronics for decades. This paper reports the applicability of MEMS technology in the implementation of RF switches for the realization of switchable low pass filter. The switch and filter are designed and simulated for 2 GHz, 3 GHZ using HPEEsof simulator.

Leveraging the dominance of Ethernet in LANs and SONET/SDH in MANs and WANs, we propose a service called CHEETAH (Circuit-switched High-speed End-to-End Transport ArcHitecture). The service concept is to provide end hosts with high-speed, end-to-end circuit connectivity on a call-by-call shared basis, where a "circuit" consists of Ethernet segments at the ends that are mapped into Ethernet-over-SONET long-distance circuits. This paper focuses on the file-transfer application for such circuits. For this application, the CHEETAH service is proposed as an add-on to the primary Internet access service already in place for enterprise hosts. This allows an end host that is sending a file to first attempt setting up an end-to-end Ethernet/EoS circuit, and if rejected, fall back to the TCP/IP path. If the circuit setup is successful, the end host will enjoy a much shorter file-transfer delay than on the TCP/IP path. To determine the conditions under which an end host with access to the CHEETAH service should attempt circuit setup, we analyze mean file-transfer delays as a function of call blocking probability in the circuit-switched network, probability of packet loss in the IP network, round-trip times, link rates, and so on.

Future networks will require very high throughput, carrying dominantly data-centric traffic. The role of Photonic Networks employing all-optical systems will become increasingly important in providing scalable bandwidth, agile reconfigurability, and low-power consumptions in the future. In particular, the self-similar nature of data traffic indicates that packet switching and burst switching will be beneficial in the Next Generation Photonic Networks. While the natural conclusion is to pursue Photonic Packet Switching and Photonic Burst Switching systems, there are significant challenges in realizing such a system due to practical limitations in optical component technologies. Lack of a viable all-optical memory technology will continue to drive us towards exploring rapid reconfigurability in the wavelength domain. We will introduce and discuss the advanced optical component technologies behind the Photonic Packet Routing system designed and demonstrated at UC Davis. The system is capable of packet switching and burst switching, as well as circuit switching with 600 psec switching speed and scalability to 42 petabit/sec aggregated switching capacity. By utilizing a combination of rapidly tunable wavelength conversion and a uniform-loss cyclic frequency (ULCF) arrayed waveguide grating router (AWGR), the system is capable of rapidly switching the packets in wavelength, time, and space domains. The label swapping module inside the Photonic Packet Routing system containing a Mach-Zehnder wavelength converter and a narrow-band fiber Bragg-grating achieves all-optical label swapping with optical 2R (potentially 3R) regeneration while maintaining optical transparency for the data payload. By utilizing the advanced optical component technologies, the Photonic Packet Routing system successfully demonstrated error-free, cascaded, multi-hop photonic packet switching and routing with optical-label swapping. This paper will review the advanced optical component technologies and their role in the Next Generation Photonic Networks.

The Preprocessing GPS - SMS Communication Unit (PCU) is a mobile tracking device used within AVL tracking systems for determining the location of vehicles. It was designed primarily to utilize the SMS service of the GSM network for communicating. The use of SMS messages is part of an effort aimed at providing a cost effective alternative for tracking the location of vehicles. Its primary function is to send information about user location across a GSM network to a Central Base Station (CBS) from which assets are being tracked. Though SMS is the main bearer, the unit is also capable of using Circuit Switch Data Service (CSD) to send and receive data from the Base Station (BS). The PCU was developed as a small hardware unit based on the Microchip microcontroller, with a multiplexer switching two RS 232 serial inputs. One input is dedicated to the GPS receiver and the second one to the wireless modem.

Parallel optical interconnects based on vertical-cavity surface-emitting lasers are being widely deployed today in switching and routing systems. Almost all of the major OEMs today are using parallel optics in their flagship products to solve interconnect problems where they are most relevant: at the bay-to-bay, shelf-to-shelf, and card-to-card levels. The density and capacity of these systems are already being constrained by the capacity of these one-dimensional links. We expect that next generations of systems will need high-density 2D parallelinks to further improve the performance-cost metric. Ultimately, we believe this metric can be optimized by directly integrating interconnect together with data processing and switching circuitry.

We first proposed monolithic optoelectronic integration of Resonant Tunneling Diode (RTD) and Heterojunction Bipolar Phototransistor (HPT). Circuit simulations using simple models of RTD and HPT successfully produced the optoelectronic bistable characteristics of RTD and HPT Negative Differential Resistance (NDR) device. The basic operation mechanism of Photoelectric Monostable-Bistable Transition Logic Elements (PMOBILE's) by utilizing the functionality of two series-connected RTDs in combination with HPT is also demonstrated by simulation.

Charging technology for solid laser power storage capacity is the key part technology of laser power. Because the capacity voltage always changes in the course of charge, and the equivalent load also changes in a large range, the load characters of charge circuit are strictly required. This paper uses active clamp technique to realize using constant-current to charge storage capacity in the stage of switching tube working at zero voltage, reduce switch loss, get higher charging efficiency.

Generalized multiprotocol label switching (GMPLS) has emerged as a very promising technology for the next generation carrier networks. GMPLS successfully combines the best features of IP and ATM in terms of QoS, privacy, flexibility and scalability. Besides, GMPLS introduces enhancements to the existing IP routing and signaling protocols by supporting not only networks that perform packet switching (IP), but also networks that perform switching in the time (TDM), wavelength (DWDM), and space domain (circuit switching). This paper discusses the design and implementation of a simulation-based GMPLS Optical Router (GOR). We build our GOR model on OPNETtm platform with all the necessary GMPLS functionalities of an optical backbone router, for the GOR to behave as a virtual machine. The description on the essential design features and the key implementation elements of the internal mechanism of GOR form the highlight of this paper. Our virtual GMPLS Optical Router has the capability of giving a more integrated and realistic simulation on wavelength routing, wavelength assignment, wavelength switching, dynamic label switching path (LSP) setup, distributed routing table calculation, and blocking mechanism of GMPLS light paths, all within a scenario. Our simulation results include the blocking rate, which is highly comparable to a real GMPLS optical network. In addition, our proposed GOR is able to provide a simulation platform for further development and future enhancement of GMPLS technologies.© (2002) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Recently, researchers have investigated the use of photonic circuit switching, or optical time division multiplexing, as a mean to sub-divide the bandwidth of wavelength channels in Dense Wavelength Division Multiplexing networks. In this paper, we present our study on the use of optical buffer in a photonic circuit switching network and how routing and time-slot assignment can be used to optimize the effectiveness of optical buffer. We will also discuss how photonic circuit switching may be integrated into the emerging IP-centric control and management scheme of GMPLS.

Optical burst switching (OBS) is a promising solution for building terabit optical routers and realizing IP over WDM. In this paper, we describe the basic concept of OBS and present a simulation model. The objective of our simulation is to estimate the throughput and the burst loss probability in such a network of OBS nodes. The traffic characteristics and reservation schemes are essential on burst loss probability. We give some comparison of these schemes by using our simulation model, as we can see, while low self-similarity traffic distribution and JET (just enough time) reservation schemes may have a potential positive effect on the overall blocking probability of the OBS network.

As many kinds of media converge to the Internet, the backbone will face new demands not only in terms of bandwidth, but also in terms of services. Therefore it is certain that Internet will evolve into IP-over-WDM based all-optical network. Obviously, the optical router will be a crucial network device. In the paper, we propose a practical architecture of optical router. Based on all-optical labeled burst switching, this architecture supports all-optical edge-to-edge data routing and transport services with transparent data rate and format, and simplifying protocol and operations. From the viewpoint of practicality, some key issues during the design have been studied and analyzed, and corresponding feasible solutions to them are put forward respectively. Finally an experimental system and results are shown.

Optical Internet carries IP traffic over a WDM optical network by taking advantage of huge bandwidth of optical network. The optical burst switching (OBS) was proposed to enhance the low utilization problem of a wavelength routing (or optical circuit switching) and at the same time, relax the technical difficulty in implementing optical packet switching. The optical burst switched networks consists of the edge routers and the core routers. In particular, the edge routers perform the adaptation function, which interfaces between the electronic layer (IP, ATM, SONET) and the optical layer (OBS networks). One of the key functions of the edge router is the burst assembly, in which the data from the electronic layer assembled into a super packet (called a burst). In this paper, we compare three burst assembly schemes: constant assemble time (CAT), variable assemble time (VAT), and adaptive assemble time (AAT). The performance of burst assemble schemes are compared based on the extensive simulations.

Optical burst switching (OBS) is a promising solution for realizing IP over WDM since it combines the advantages of optical circuit switching and optical packet switching (OPS) while overcoming their shortages. In this paper, the mechanisms to implement burst assembly and scheduling in an OBS edge router are discussed. The performance of the output traffic from the edge router, e.g. traffic shape and delay, is analyzed with simulation results. The results show that the traffic shape and the delay are affected by maximum delay specified by different classes of services, the maximum length of the burst, the burstiness of the traffic, and the mechanisms of burst assembly and scheduling. By properly designing, the traffic shape can be improvement while the delay keeping in the bounds.

The paper describes our activities on optical packet switching by means of our out-of-band optical label switching technique. It also discusses our recent progress in some key components for optical packet/burst switching, including an optical packet synchronizer and a burst-mode optical receiver.

The realization technology of a new kind of multi-wavelengths label optical packet switching technology is discussed. In this switching, optical header is labeled by several optical pulses at different wavelengths in the same optical communication channel band as optical payload. The optical transmitter with header generation and packet formation, receiver with data restoration and switching node with route processing are introduced. The switching principle is verified by a simplified experiment.

Optical Burst Switching combines the advantages of circuit switching and package switching, and it can improve the efficiency of system resource with high speed transmission and Qos. Now the rationale of OBS has been discussed a lot, and many schemes have been given. While there are still some key problems need to be resolved, such as in the model of IP over WDM, how to assemble the IP packages into burst that will be transmitted in the optical networks; and how to set the offset time between payload and its control package. In this paper, we provide a part randomized offset-time method to calculate the offset time upon the basis of: the network condition of the optical path that the payload will pass, the shaping scheme which is introduced to reduce the drop probability. At the same time, although there are many factors should be considered when setting the offset time, too sophisticated algorithm will be discarded. Because our ultimate end is to get equilibrium between the control complexity and bandwidth efficiency even some redundancy will be introduced in this method. We also consider a balance between end-to-end delay and blocking probability introduced by the new offset-time determination algorithm. At last, we will analyze the improvement in the network performance if the offset time has been set properly, especially when associated with delayed reservation technology.

In this paper, we propose a novel optical packet switching mechanism incorporating void filling that makes some packets switch in optical domain, others in electronic domain. The optical packet switching is solved in time, space, wavelength and electronic domain. We also present numerical simulations under Poisson traffic and variable IP packet length condition. Simulation results have shown improvement in packet loss probability by using our switching mechanism.

For DWDM networks, optical burst switching has been proposed as a combination of the best of optical circuit switching and optical packet switching while avoiding their shortcomings. One of the most challenging aspects of implementing burst switching DWDM networks is the design of data channel scheduling algorithms for core routers. In this paper, we present a channel scheduling algorithm, discuss its hardware implementation issues, and propose a hardware design.© (2002) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

We consider optical networks with nodes interconnected by links comprising multiple channels. The blocking performance of such networks depends on the channel-switching capabilities of the nodes. In this paper, we focus on developing analytical models for evaluating the blocking performance in circuit-switched optical networks. Several architectures for channel-switching are presented. Multi-rate circuits may be established if different circuits are allocated different number of channels. Depending on how the network can assign the channels to circuits that request more than one channel, it is classified as a Data Splitting Network (DSN) or a non-DSN. We consider multi-rate circuits which require either a single channel or a given number of channels k (>1). Analytical models for computing blocking probabilities are developed for various channel-switching architectures at the nodes. The validity of the models is shown by comparing the analytical results with simulations. Numerical results in a single-fiber TDM wavelength-routing network suggest that limited channel-switching may be sufficient even for circuits that require more than one channel or slot. A comparison of DSN and non-DSN shows that data splitting can accommodate more multi-slot circuits at the expense of blocking more single-slot circuits.

This paper outlines two ways in which liquid crystal devices can be used as optical switches in telecommunications networks. This has become a very important area of research as electronic switches are becoming bottleneck in both circuit switched and packet switched networks. These two applications have been highlighted as they require radically different switching techniques, yet liquid crystals can be used in both as the reconfigurable element. The first application is in circuit switching, where optical loss, crosstalk and polarization dependence are all critical parameters, whereas reconfigurration time is not. For this application a reconfigurable hologram written in phase with liquid crystal modulators is used to diffract optical beams in an optically transparent switch fabric. This technique does not suit packet switching as a much faster reconfigurration rate is required and loss and crosstalk are not critical. Hence a shutter based crossbar is used as a packet switch using a liquid crystal which can be reconfigured in 1 msec. A hybrid electronic/optical approach is also taken as this allows a much more scalable packet switch to be built.© (2001) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Standardization of 3rd Generation (3G) mobile communication systems has produced the first specification releases and the commercial deployment of the 3G systems has started. Whereas 1G and 2G focused on efficiently providing voice services, in 3G a lot of attention has been devoted to solutions that support both Circuit Switched (CS) and Packet Switched (PS) communication. That has called for very flexible air interface and network solutions. 3G will continue to evolve and there are already on-going standardization activities that will, for example, boost the peak data rates up to 5-10 Mbps and improve spectral efficiency by 2-4 times. In the future, 3G evolution will be going towards 10/100 Mbps peak data rates in wide/local are coverage, respectively. This will take place partly because of technical improvements of 3G radio interface solutions, but also due to network evolution which will allow the integration other radio access methods like radio LANs into the 3G system. In longer term the 3G network evolution will be going towards ALL-IP networks. As 3G evolution seems to be going towards 10 Mbps/100 Mbps peak data rates and ALL-IP networks any beyond 3G air interface or network solution should be clearly better in order to justify its technical and commercial feasibility. Given the long evolution time of 3G and integration of other radio access schemes with 3G radio we may not even see a new, complete beyond 3G system being developed. Maybe we will just witness the emergence of a new, more advanced radio access solution which will then be connected to the evolving 3G network. As 3G evolution will continue for several years to come the research targets for any beyond 3G solutions must be set very high. When it comes to air interface, we should aim at 100 Mbps peak data rates for wide area access with high mobility, and at 1 Gbps for local area access with low mobility. Regarding possible commercial launches of any beyond 3G systems or solutions they could then take place around year 2010 or even later.

This paper discusses optical-label switching and an experimental demonstration of high-speed all-optical packet switching with all-optical label-swapping capabilities. The subcarrier optical-label content was optically extracted and compared against the forwarding table, which induced packet forwarding as well as label-swapping decisions. The all- optical label processing technique was immune from dispersion induced fading and polarization induced degradation effects. The packet switching was accomplished by a combination of rapidly tunable wavelength conversion and a uniform-loss cyclic frequency (ULCF) 8 X 8 arrayed waveguide grating router (AWGR). The packet routing system achieved 600 psec switching time with a forwarding decision time of 250 nsec. All optical label swapping incorporated optical subcarrier filtering using fiber Bragg gratings. Accumulated packet-by- packet Bit-Error-Rate measurements confirm the successful and low-penalty packet routing with label swapping technique, while maintaining optical transparency for the data payload.© (2001) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Recent advances in optical devices greatly enhance the feasibility of optical code-division multiplexing/wavelength-division multiplexing (CDM/WDM) network. In this paper, the performance in blocking probability in a two-dimension code/wavelength routing network is investigated. Facilitated by the optical code and wavelength conversion, the capability in reconfigurability, restoration and protection of the optical layer is enhanced. The network blocking performances under various conversion capabilities are investigated and the blocking probabilities are given in closed-form expressions. Under certain conditions, the analysis of the two-dimension routing network can be extended to N-dimension routing network by iterative decomposition procedure. Numerical results are obtained to show the performance improvement from the optical code conversion in terms of code conversion gain. With wavelength conversion provided in the network, the addition of the code conversion achieves a conversion gain as high as 10 4 even when the code conversion is provided with a low placement density of 0.2.

In IP over WDM networks, optical paths (OP's) form a virtual topology seen from the upper IP layer, and greatly relieve the burden of electronic processing thanks to optical bypasses. However, in contrast to the circuit-switching, connection-oriented and coarser OP's, the IP is packet-switching, connectionless and finely granular in nature. Special care should be taken in the control plane to bridge the gap. In our research, we incorporate the idea of optical flow. Optical flows are optically switched in intermediate nodes and converted into electronic form at termination to treat small nested flows. Two modes of optical flow switching (OFS) are examined. One is optical burst switching (OBS), which sends control packet ahead of data to announce a switching and make future reservations. The other is Optical Grooming Switching (OGS), which put packet flows into finer container with discrete yet flexible bit rate. We find that both of these two modes use statistical multiplexing to bridge the gap between IP packets and WDM OP's, and such function is fulfilled respectively by scheduling in OBS and by framing in OGS. Taking snapshots of statistical multiplexing, we formulate the bandwidth distribution (BWD) of the whole OFS networks and the bandwidth usage (BWU) over each node pair. The BWD can be expressed by the product of a traffic-independent grooming matrix with the traffic demand matrix under a given virtual topology. Changes to the grooming matrix are reflected by adjustments in framing or scheduling. Three distinct states of BWU are also identified to determine the initiation of such changes. The BWU states are hence proposed to add to the extension of existing routing protocols.

This paper discusses the new architecture and protocol of Optical-Label switching in light of its interworking with MPLS and MPLambdaS networks. The dynamic and agile characteristics of Optical-Label switching indicate its support of diverse QoS, CoS and ToS services while interoperating packet and circuit switching. Optical-Label switching routers employ time, space and wavelength domain switching for contention resolution, and the interface with MPLS, MPLambdaS, or any client networks with proper signaling such as extension of GMPLS.

In this paper, we introduce deflection routing for irregular meshed networks in an optical packet based environment. By using the MPLS (Multiprotocol Label Switching) control plane to manage the deflected paths, we offer a relatively simple way of controlling the otherwise complicated deflected paths. This is beneficial due to the way it simplifies the optical packet switch so it only needs to deal with simple forwarding and thereby relaxing the need for buffering. The latter is becoming increasingly more important in optical networks since there is no practical optical memory insight for the near future. By introducing deflection routing along with statistical wavelength multiplexing, the need for buffers is greatly reduced while the network resources are better utilized. MPLS is emerging as a main protocol for future networks, whether being based on packet switched or circuit switched technology, hence rendering this approach to be compatible with most modern and future systems.

A stable high-resolution electromagnetic deflection control circuit for an electron-beam lithography (EBL) system has been developed. This deflection control circuit has enabled an EBL system to deal with a wide deflection area of 2.5-mm square having fine address units for a pattern placement of 1.25 nm. The deflection-control circuit consists of a new digital to analog converter (DAC) circuit, whose resolution is 21 bits, and a low-drift current-amplifier circuit. To achieve such high-stability and high-resolution, we had to develop a low noise-current cell structure for the new DAC circuit, because the output-signal noise of the DAC circuit is a major source of interference at the desired resolution. A local temperature control technique has been incorporated into the circuit to reduce fluctuations of the deflection control signal caused by ambient thermal variations. The low noise-current cell structure, which consists of multiple current buffers and low-pass filters, is placed between a constant current source circuit and a differential-switch circuit for each bit of the DAC circuit. The simulation results of the DAC circuit showed that the output-signal noise of the DAC circuit could be reduced to less than 0.4 nm rms, which is small enough to achieve the desired resolution. As the results of the experimentally evaluation of the deflection control circuit show, the total noise of the deflection-control signal obtained was less than 0.6 nm rms and the signal stability obtained was better than 0.3 nm rms. An evaluation of the performance of the new EBL system to which the new deflection control circuit was applied, showed that the critical-dimension accuracy obtained was better than 5 nm (3sigma) and the positioning accuracy obtained was better than 10 nm (3sigma) for the area controlled by electromagnetic deflector.

This paper addresses the suitability of WDM coarse packet switching solutions for IP traffic. Our findings show that the combination of traffic grooming at the higher layers and coarse packet switching at the optical layer provides at least the same performance as more sophisticated and difficult to realize photonic packet switching solutions. We propose a network architecture named files-over-lightpaths that not only simplifies the network optical and electronic design by making use of coarse packet switching, but also serves to the purpose of decreasing the TCP transaction latency in comparison to a flat or split Internet organization with fine grain photonic packet switching.