Program

Thursday 7, June 2007. 10:30-12:15

Opening address
-Tadanobu Okada, General Chair, NTT
- Bijan Jabbari, General Chair, ISOCORE

Keynote
"Global Optical Networking and The Role Standards Play"
-John McDonough, NEC Corporation of America

John McDonnough is an industry veteran with more than twenty five years of supervisory, managerial, and technical experience in both the North American and international telecommunications industry. He spent a number of years working for both carriers and vendors community. Some of his previous posts include senior technical management positions with Nynex (now Verizon) and Cisco. At the same time, John was very active in different industry standard forums, where among others, he served as the Vice President of OIF Board of Directors, Vice Chair of ATIS OPTXS Optical Hierarchical Interface Working Group, and Vice Chairman of internal NGN Network Management Subcommittee. He joined NEC Corporation of America in 2006 and serves as NEC principal representative in industry forums dealing with optical networking.

Program introduction
- Joseph Berthold, Ciena and Satoru Okamoto, Keio University

Exhibition introduction
- Takeshi Akaike, NTT and Soichiro Araki, NEC

S1: Next Generation Transport from Ethernet to Lambda
Thursday 7, June 2007. 14:00-15:00

S1-1 "Recent Progress in ASON Technologies in the OIF"
Joseph Berthold, Ciena

Biography: Joseph Berthold is Vice President, Network Architecture at CIENA, where he has worked since early 1997. There he contributes to the understanding of future network architecture directions, the definition of CIENA's networking products, and is responsible for coordination of CIENA's work in industry standards. He is also president of the Optical Internetworking Forum. Prior to Ciena he held various research and development positions at Bell Labs and Bellcore from 1977-1997.

S1-2 "Carrier Ethernet: A Packet Optimized Transport Technology"
Dinesh Mohan, Nortel

Biography: Dinesh Mohan has 14+ years of experience in telecom and data networking research, design, development, and product management. He is a subject matter expert on OAM and Carrier Ethernet. He has been an active contributor in IEEE 802.1, ITU-T SG 13 and SG 15, IETF L2VPN & PWE3 WG, MEF, and DSLF. He has also been an editor of several standards and draft recommendations in ITU-T, IEEE 802.1, MEF and IETF L2VPN WG. He is currently a product line manager in Nortel’s Metro Ethernet Networks organization.He had received his B.E. from Delhi College of Engineering and M.Eng. from CarletonUniversity.

Program at a glance

T1: Path computation element
Thursday 7, June 2007. 15:50-16:50, 17:00-18:00

T1-1 "Latest developments in techniques for computing inter-layer and multi-region paths for multi-layer traffic engineering."
Adrian Farrel, Aria Networks

GMPLS is seeing adoption and deployment in various technology-specific networks to provide an automated control plane that ease the provisioning of connections and advanced services such as protection and recovery.

Different technology networks are often operated hierarchically so that a connection in a lower layer network provides a point-to-point link in the upper layer network. For example, links between IP routers are often achieved by multi-hop connections through a TDM network. Traffic engineering is often utilized within each network layer to make best use of available network resources and to route traffic around hot-spots.

As more of the network layers converge on MPLS and GMPLS technologies it becomes possible to consider traffic engineering solutions that span multiple layers. Inter-layer traffic engineering enables optimization of resources across all networks and enables operators to build more robust and economically viable solutions.

This presentation will examine new techniques are being discussed in the IETF's PCE and CCAMP working groups to enable automatic end-to-end computation and provisioning of TE LSPs that utilize network resources at different layers while preserving administrative confidentiality, security, and policy. Those layers may represent different technologies or different administrative domains. The speaker will describe why the options of TE aggregation have been rejected, and show how cooperation between path computation engines and management applications can produce an optimal inter-domain path.
The presentation will conclude with an analysis of existing path computation techniques for multi-layer optimization and will suggest how new, holistic optimization techniques can produce dramatically better results.

Biography: Adrian Farrel is co-chair of the IETF’s Common Control and Measurement Plane (CCAMP) Working Group, which is responsible for the development of the GMPLS family of protocols. He also chairs the Path Computation Element (PCE) Working Group, which is applying remote path computation techniques to MPLS and GMPLS networks, and the Layer One VPN (L1VPN) Working Group, which is developing mechanisms to manage connectivity over optical networks using GMPLS.
Building on his 20 years’ experience designing and developing communications software, Adrian runs a successful consultancy company, Old Dog Consulting, providing advice on implementation, deployment, and standardization of Internet Protocol-based solutions, especially in the arena of MPLS and GMPLS.
Adrian currently serves as Chief Technology Officer of Aria Networks manufacturing intelligent path computation, network optimization, and capacity planning solutions for IP, MPLS and GMPLS next generation networks.
As well as frequently speaking at conferences, giving tutorials on MPLS and GMPLS, and authoring several white papers on GMPLS, Adrian is the author of The Internet and Its Protocols: A Comparative Approach (Morgan-Kaufmann, 2004) which explains many of the IP-based protocols including those that make up MPLS and GMPLS, and GMPLS: Architecture and Applications (Morgan-Kaufmann, 2005).

T1-2 "Path Computation Algorithms for DRAGON Multi-Layer Network"
Bijan Jabbari, George Mason University
Jerry Sobieski, Mid-Atlantic Crossroads
Tom Lehman, University of Southern California/ISI-East
Shujia Gong, GeorgeMasonUniversity
Xi Yang, University of Southern California/ISI-East

This presentation will first review the early path computation implementation in DRAGON (Dynamic Resource Allocation via GMPLS Optical Networks) project. It subsequently describes our approach to implementing enhanced path computation for multi-layer network in DRAGON. The scenario will include both intra-domain as well as inter-domain path determination.
We will discuss the scalability and the computational complexity. Then, we will consider the efficiency of the approach and show that we can attain a high degree of optimality.

The talk will also include the performance aspects in which the performance will be evaluated using simulation, emulation of large networks and real algorithm implantation. The talk will also cover the software implementation aspects.

T1-3 "Advanced Backbone Network Management with PCE (Path Computation Element)"
Eiji Oki, Ichiro Inoue, and Kohei Shiomoto, NTT

Multi-protocol Label Switching (MPLS) and Generalized MPLS technologies provide sophisticated traffic engineering (TE), which controls label switching path with an explicit route to satisfy its constraints such as bandwidth and delay. A PCE is an entity that is capable of computing a network path or route based on a network graph, and applying computational constraints. It is applied to intra-area, inter-area, inter-AS, and inter-layer traffic engineering. A PCE-based architecture for the computation of paths is being discussed in IETF PCE Working Group.

This presentation addresses advanced network management with a PCE. We present several PCE usage scenarios and evaluate their scenarios on how a PCE is useful for network management. We also define database placement for network management.
In a conventional network management, when a path is set up, operators configure the path attributes, which may includes a specified route or may not include one, into a Network Management System (NMS) or a router. When a route is specified, operators consider their path setup policy and constraints in a manual manner. When a route is not specified, they let a node find an appropriate route that satisfies the constraints. However, as various services come out, policy and constraints on the path setup become complex and fast path setup decision will be required, As a result, conventional network management approaches are no longer appropriate. Therefore, a new path computation model is required.

Using PCEs is able to solve conventional problems on network management. A PCE complements a Constraint Shortest Path First (CSPF) engine equipped with a source node. In case that the CSPF engine may not handle complicated path computation, such as inter-domain, interlayer, point-to-multipoint path computation, a PCE helps a node to provide an appropriate route.
PCE also supports an NMS by handling network management policy, complicated constraints, which are manually done by operators in conventional networks. In addition, a PCE is useful for global optimization and fast provisioning when traffic demand changes and network failure occurs. Thanks to PCE, network management will become more advanced in terms of optimality, flexibility and reliability.

Biography: Eiji Oki received B.E. and M.E. degrees in Instrumentation Engineering and a Ph.D. degree in Electrical Engineering from Keio University, Yokohama, Japan. In 1993, he joined Nippon Telegraph and Telephone Corporation's (NTT's) Communication Switching Laboratories, Tokyo Japan. He has been researching multimedia-communication network architectures based on ATM techniques, traffic-control methods, and high-speed switching systems. From 2000 to 2001, he was a Visiting Scholar at Polytechnic University, Brooklyn, New York, where he was involved in designing tera-bit switch/router systems. He is now engaged in researching and developing high-speed optical IP backbone networks as a Senior Research Engineer with NTT Network Service Systems Laboratories. He is active in standardization of GMPLS and Path Computation Element (PCE) in the IETF. Dr. Oki was the recipient of the 1998 Switching System Research Award and the 1999 Excellent Paper Award presented by IEICE, and the 2001 Asia-Pacific Outstanding Young Researcher Award presented by IEEE Communications Society for his contribution to broadband network, ATM, and optical IP technologies. He co-authored two books, ``Broadband Packet Switching Technologies,'' published by John Wiley, New York, in 2001 and ``GMPLS Technologies,'' published by RC Press, Boca Raton, in 2005. He is an IEEE Senior Member.

T1-4 "Interactive operation for highly-scalable multi-domain IP/Photonic networks based on GMPLS and PCE"
Itaru Nishioka and Soichiro Araki, NEC Corporation

We present an interactive network operation model between control plane and management plane in large-scale converged IP/Photonic networks using GMPLS (Generalized multi-protocol label switching) and PCE (Path computation element). Operation using GMPLS in IP/Photonic networks enables to reduce operational expenditure (Opex), because it facilitates more automatic control than that of a current centralized EMS/NMS. However it is necessary to introduce a new network operation and management model due to the following problems. One is a database inconsistency due to a synchronization delay between network information on a centralized management plane and that on a distributed control plane. The other is a scalability problem, because almost functionalities of GMPLS technology are restricted within an IGP area like in MPLS. To solve these problems, we propose an advanced network operation model using PCE that collaborates with NMS/EMS and GMPLS.

PCE can acts as intelligent agents to collect information for a traffic engineering database (TED) from network via OSPF-TE. However, when using OSPF-TE, there is some update delay that induces unsynchronized condition between the latest network topology information and TED on PCE. This unsynchronized condition makes a control plane instable, because PCE calculates non-optimal or wrong routes using TED different from the latest topology information, which may result in GMPLS signaling failure. This instability becomes more significant, when a path provisioning request is issued more frequently. To avoid this instability, the key point is who updates PCE’s TED. In our network operation model, TED in PCE is also updated by PCE itself locally using CSPF calculation results, before updated by OSPF-TE. TED updated by OSPF is used for just checking that local update was not wrong. Using this operation model, PCE provides reliable route using its TED, which is very close to the latest network information, and we can migrate easily from the current centralized NMS/EMS model to the proposed one, since TED is stored like a centralized database in PCE.

PCE also provides scalable LSP calculation functionalities over domains in cooperation with other PCEs. When calculating a route over domains, it is necessary for PCE to determine transit and destination domains in advance or hop-by-hop. If it is the composite PCE on ABR or ASBR, transit or destination domains can be determined based on IP reachability, but it is difficult for external PCEs to use only IP reachability. In the presentation, we discuss a detailed PCE discovery mechanism in the external PCEs.

Acknowledgment: This work was partly supported by National institute of Information and Communications of Technology Japan.

Biography: Itaru Nishioka received the B.E. and M.E. degrees in communications engineering from Osaka University, Osaka, Japan, in 1998 and 2000, respectively. In 2000, he joined NEC Corporation, Kawasaki, Japan, where he has been engaged in the research and development of optical networking systems and the design of its control/management architecture. He is a member of IEEE.

T1-5 "Multi-domain TE Routing: Hierarchical Routing with PCE"
Huiying Xu and Dan Li, Huawei Technologies

With the development of ASON architecture and GMPLS protocols, the management and control of multi-AS networks are very important. Hierarchy routing architecture is defined in ITU-T Rec. G.7715, OSPF based extension for inter-domain (inter-AS) routing protocol is defined in OIF, Path Computation Element (PCE) architecture and related protocols such as PCE Communication Protocol (PCECP) are also defined in IETF. Both OSPF based extension and PCE solutions can handle the main routing issues in multi-AS networks. Huawei did some experiments in multi-AS networks with OSPF based extension solution, and collected the routing performance data.

1) Simulation test with 200 nodes in 4 levels hierarchy routing network

200 simulated nodes are grouped into 4 levels hierarchy routing network (Figure 1 and Figure 2). Functionalities such as node addresses auto-aggregation, routing information feed up/feed down between two adjacent levels, inter-domain TE links auto-discovery are implemented, and the corresponding performance data like addresses aggregation time, topology information flooding time, multi-AS service setup/restoration time have been collected.

Figure 1. Hierarchy Topology

Figure 2. Domain Topology

2) Trial with 30 real nodes (Huawei OSN3500) in 2 levels hierarchy routing network

30 real equipments (Huawei OSN3500) grouped into 2 levels hierarchy routing network (Figure 3), which can support multiple protection types service. The typical routing performance data has been collected.

Figure 3. Domain Topology

The result of the experiments shows more work has to be done on the current OSPF based extension solution in order to support the commercial use:

1) Only the data format of inter-domain routing information is defined, but doesn't define the remote path query interface;
2) The imprecise abstract routing topology may result in signaling roll back;
3) The Transport Network Assigned (TNA) Address aggregation by each node may result in numerous network addresses;
4) The multi-AS restoration procedure;
5) The separation of call and routing in multi-AS;
6) Diverse routing in multi-AS
7) Policy process in multi-AS;

The introduction of PCE architecture can provide the multi-AS path computation, node address translation, policy process, and avoid the imprecise abstract routing topology.

OSPF based extension solution provides the way of routing information flooding in multi-AS network, and the PCE architecture provides the path computation (such as remote path query, diverse paths, protection or restoration path computation), policy process and address translation as the complemental functionalities. So the OSPF based extension solution with the PCE architecture may provide a good way of routing solution in multi-AS network.

Biography: Huiying Xu, a system engineer, supervisor, in Advanced Technology Department of Wireline Network Business Unit of Huawei Technologies Co., Ltd., a leader in providing next generation telecommunications networks. He joined Huawei Technologies Co., Ltd. in 1999. He received his B.E in computer application from Tianjing University. His research interests include GMPLS control plane、PCE and network planning tool.

T1-6 "Global Concurrent Optimization (GCO) Application in PCE-based Network"
Young Lee, Huawei, J-L Le Roux, France Telecom, Daniel King, Aria Networks, Eiji Oki, NTT

In this paper, we present an application of the PCE concept to Global Concurrent Optimization (GCO). Global Concurrent Optimization is concurrent path computation application where a set of TE paths are computed concurrently with a global objective function in order to efficiently utilize network resources and to avoid blocking problems. A GCO path computation is able to simultaneously consider the entire topology of the network and the complete set of existing LSPs, and their respective constraints, and look to compute or re-optimize the entire network to satisfy all constraints for all LSPs.

The need for a global concurrent path computation may arise when network operators need to set up a large number of TE LSPs in their network planning process. When a new TE network needs to be provisioned from a green-field perspective, a set of TE LSPs need to be created based on traffic demand, network topology, service constraints, and network resources. Under this scenario, concurrent computation ability is highly desirable, or required, to utilize network resources in an optimal manner and avoid blocking risks. Sequential path computation could potentially result in sub-optimal use of network resources or even blocking issues.

The need for a global concurrent path computation may also arise in existing networks. When an existing TE LSP network experiences sub-optimal use of its resources, the need for re-optimization or reconfiguration may arise. The scope of re-optimization and reconfiguration may vary depending on particular situations. The scope of re-optimization may be limited to bandwidth modification to an existing TE LSP. However, it could well be that a large number of TE LSPs may need to be re-optimized concurrently. In an extreme case, the TE LSPs may need to be globally re-optimized. Note that sequential re-optimization of such TE LSPs is unlikely to produce substantial improvements in overall network optimization except in very sparsely utilized networks.

A global concurrent path computation is typically an off-line computation; however, there may be the possibility that network operators might require on-line computation for a global concurrent path computation in the event of catastrophic network failures and/or in overload and any abnormal traffic load condition, where a set of TE LSPs need to be optimally rerouted in real-time.

In this paper, we present the PCE-based network architecture that supports a global concurrent optimization application, the requirements in support of a global concurrent optimization application, the protocol extensions to the PCEP (PCE communication Protocol) to satisfy the identified requirements.

Biography: Young is currently Principal Technologist at Advance Technology Division, Huawei Technologies USA Research Center, Plano, Texas. He is leading optical transport control plane technology research and development.His research interest includes distributed path computation architecture, multi-layer traffic engineering methodology, and network optimization modeling and new concept development in optical control plane signaling and routing.
Prior to joining to Huawei Technologies, Young was a co-founder and a Principal Architect at Ceterus Networks (2001-2005) where he developed topology discovery protocol and control plane architecture for optical transport core product. Prior to joining to Ceterus Networks, Young was Principal Technical Staff Member at AT&T/Bell Labs in Middletown/Holmdel, New Jersey. At AT&T Labs (1996-2000), he was responsible for core IP/MPLS network architecture evolution and AT&T End-to-end architecture planning. He also involved voice/data convergence architecture planning and evolution. At Bell Labs (1987-1995), Young was responsible for developing dynamic routing schemes and traffic network management control and measurement development. He is currently active in IETF PCE and CCAMP WGs and is a co-author of RFC 3214. He holds several patents in the area of dynamic routing and switching technology and several patents pending in optical networking.
Young Lee received B.A. degree in Applied Mathematics from the University of California at Berkeley in 1986, M.S. degree in Operations Research from StanfordUniversity, Stanford, CA, in 1987, andPh.D. degree in Decision Sciences and Engineering Systems from Rensselaer Polytechnic Institute, Troy, NY, in 1996. He is a member of IEEE and Alpha Phi Mu honor society.

Program at a glance

S2: Progress of GMPLS and its applications
Thursday 8, June 2007. 10:00-11:30

S2-1"GMPLS Real-Time Deployment: Challenges and Current Status"
Rajiv Papneja, ISOCORE

The presentation will examine the technical challenges in deploying GMPLS in a large scale environment. The presentation further presents the current state of the GMPLS based technology and the advancement made during the last year and analyze the issues associated with the introduction of GMPLS in the next generation architecture of service providers and carriers. The motivation for this work is primarily driven by network convergence and the promise for operational and capital expenses reduction. Architectural simplification is seen as well as a driver. The areas examined include GMPLS provisioning, MPLS/GMPLS integration, GMPLS protection. Finally we will conclude on standardization status and how it can help to speed up deployment.

S2-2"Challenges of e-Science Applications with On-Demand Optical Grid Networks: e-VLBI over GMPLS based Lightpath Networks"
Sugang XU, NICT

The performance/cost ratio of both computers and networkshas been are improving rapidly during the past few years.To perform larger scale computing with lower costs or enable collaborations widely among geographically distributed scientific research communities, Research and development oin both computer and communication technologies for large-scale distributed grid systems has been are emerging to enable larger-scale computing at lower cost or wide-spread collaborations between geographically distributed scientific-research communities. e-VLBI(Very Long Baseline Interferometry) [1] as one effective observation approach has been employed on the most frontiers of the research in Astronomy, Geodesy, and Spacecraft Navigation. During observation 256Mbps - 1Gbps (60TB per day) of data is generated at each observation site. Three essential requirements of e-VLBI (long baseline for high angular resolution, large bandwidth for high sensitivity, and real-time data transfer for fast/rapid turnaround) make WDM lightpath network the ideal candidate for high-quality data transmission.

Generalized Multi-Protocol Label Switching (GMPLS) suite of protocols allows carriers to automate the provisioning and management of the network, and promises to lower the cost of operation by several orders of magnitude compared with manual operation. GMPLS based lightpath network is playing a key role that enables quick turnaround from observation to data processing and results display. This presentation briefly reviews our previous efforts on optical grid networks [2] first, and then addresses our filed trials on e-VLBI over GMPLS based lightpath networks. It was a joint work of e-VLBI and GMPLS, lightpath, and optical grid infrastructure.

Figure 1 depicts the network architecture in the e-VLBI demo. Before the data transmission, when a demand of lightpath network comes, at a proper network node, this network request was decomposed into multiple bidirectional-lightpath creation requests starting at different source nodes. Then these bidirectional lightpaths were established automatically with a path message extension of RSVP-TE [3]. Figure 2 presents this path message extension. Figure 3 shows a snapshot of a RSVP-TE console, which was showing a bidirectional and an unidirectional lightpaths. After that, datum from Haystack and Kashima observation sites were firstly transferred to Tokyo via transpacific L3 and domestic L2 links, respectively, then, decomposed and delivered therein over the on-demand multiple lightpaths network to three correlation calculation computers, which were located at different places for parallel correlation computation.

To verify the possibility of new services that support the emerging large-scale distributed computing applications such as e-VLBI we implemented GMPLS based lightpath control system on a DWDM optical network. As the final part of this presentation, we will describe the implementation and experiment in detail.

Reference:
[1] e-VLBI at NICT (http://www2.nict.go.jp/w/w114/stsi/e/research.html)
[2] S. Xu and H. Harai, “Optical Ring Services in GMPLS Based Mesh Networks: An Implementation of Optical GRID,” OFC/NFOEC 2006 (Optical Fiber Communication Conference & Exposition and the National Fiber Optic Engineers Conference 2006), no.JThB74, Mar. 2006.
[3] IETF-RFC [3473].

Fig. 1. Network architecture of the e-VLBI demo.

Fig. 2. Path Message and Bidirectional Lightpath Flag Object.

Fig. 3. A snapshot of RSVP-TE console showing a “Bdlp” field (bidirectional and unidirectional lightpath provisioning)

Biography: Sugang Xu received his B.E. and M.E. degrees in computer engineering from Beijing Polytechnic University, Beijing, China, in 1994 and 1997, respectively, and Ph.D. degree in information and communication engineering at the University of Tokyo, Tokyo, Japan, in 2002. He joined the Global Information and Telecommunication Institute, Waseda University in 2002, as a research associate there. Since 2005, he joined National Institute of Information and Communications Technology (NICT), Tokyo, Japan, as an expert researcher. His research interests include algorithms, network architectures, photonic network control, optical grid network systems, parallel and distributed processing. He received the IEEE APB-APCC 2000 Best Paper Award in 2000, the IEICE Communications Society Conference English Section Best Paper Award in 2004. He is a member of the IEICE and IEEE.

S2-3"Highly Sensitive Radio Interferometer Observation with High Speed Data Transmission through Optical Fiber Communication Network"
Noriyuki Kawaguchi, Yusuke Kono and Tomoaki Oyama, National Astronomical Observatory

Radio interferometer observation connecting radio telescopes remotely separated in the long distance, i.e. Very Long Baseline Interferometer (VLBI) produces images of a compact radio source with the highest resolution ever achieved in the astronomical instruments. The sensitivity is, however, severely limited by a reason that astronomers have to use a data recorder to store their data, a record of a radio signal arriving to the radio telescope onto magnetic tapes at individual observatories for later correlation processing. The recording rate is usually 256 Mpbs. The most advanced Japanese VLBI array named the VERA, was established in 2002 and equipped with the up-to-date recorder, limited in the recording rate at 1024 Mbps. That is a reason why VLBI is only involved in the observation of bright galaxy cores or maser emission around stars.

National Astronomical Observatory, Japan (NAOJ) has started the development of optical fiber connected VLBI array since 1996 in cooperation with NTT laboratories. The first connection of the Usuda 64-m telescope and the Kashima 34-m telescope at the data transmission rate of 256 Mbps through an STM-16 ATM line of 2.5 Gbps was successfully performed in 1998 on the detection of the flare from the RC CVn binary star, HR1099. In 2001 the first advanced academic communication network, the Super SINET, realized the connection of another radio telescope, Tsukuba 32-m telescope to a correlation station in NAOJ, Mitaka, Tokyo through two communication lines of the data rate 2048 Mbps each. The network was expanded to the Gifu 11-m telescope in 2003 and to Yamaguchi 32-m telescope in 2006. Now four large radio telescopes operating in an 8-GHz band, Usuda 64-m, Tsukuba 32-m, Kashima 34-m and Yamaguchi 32-m are connected at the minimum transmission rate of 2048 Mbps with a correlator located in Mitaka. The network is now regularly operated for various kinds of engineering tests and scientific observations. In near future we are expecting to have a higher frequency 22-GHz network to be formed with Gifu 11-m, Nobeyama 45-m, Kashima 34-m and Tsukuba 32-m aiming at detecting high-energy electron emission from a galaxy in the early-stage formation of the structure.

In this paper the current state of the 8-GHz network is presented with a future plan expecting much higher transmission rate on the communication network for higher sensitivity. The topics relating to the higher data acquisition system and a request of high-bandwidth demands by radio astronomers are also presented.

Biography of Kawaguchi
1949 Born in Gifu
1975 Graduated from the master course of the University of Electro Communications
1977 Starting developments of VLBI instruments with the Radio Research Laboratory, Ministry of Posts and Telecommunications
1982-1986 Conducting the US-Japan Joint geodetic VLBI experiments and successfully detecting the plate tectonic motion of Hawaii Island.
1989Assistant Professor in National Astronomical Observatory Starting developments of the Space VLBI instruments.
1996 Starting the project OLIVE, Optical LInked Vlbi Experiment
1997 Launch of the Space Orbiting Telescope, HALCA. Ph. D. from The Graduate University for Advanced Studies
1999 Professor of National Astronomical Observatory. Starting developments of astrometric VLBI, VERA
2003 Starting VERA and eVLBI observations
Now A chief of eVLBI Project of NAO

Program at a glance

T2: Network Management
Thursday 8, June 2007. 14:00-15:20

T2-1 "What GMPLS Can Do? Analysis on Operation and Service Aspects"
Tomonori Takeda, Takashi Miyamura, Ichiro Inoue and Kohei Shiomoto, NTT

This presentation addresses benefits that GMPLS brings in terms of operation and service aspects, mainly in the context of transport networks.

Firstly, GMPLS has a potential to enable “plug and play” in the transport network. Traditionally, operators have to configure many parameters for nodes and links in advance so that the network is ready to be used. Management systems can help to assign parameters and/or to check configuration consistency, but it is hard to eliminate all of manual configuration and miss-configuration.

Possible solutions for plug and play will be presented, including the use of LMP. Currently parameters that LMP supports for automatic configuration are limited even for GMPLS specific ones, and very little for traditional ones. However, this can be improved by extending LMP and combining with automatic parameter assignment at the network equipment level as well as at the management system level.

Secondly, GMPLS has a potential to enable “simplified network operation”, where a set of tools are available for operators to operate and manage the network. Traditionally, operations such as network planning, service provisioning and fault management are manual intensive, especially under multi-vendor networks.

Possible solutions for simplified network operation will be presented, including PCE (Path Computation Element) based architectures. GMPLS allows automation distributed at the network equipment level (e.g., service provisioning, failure recovery), but how to apply carrier own policies on such automation requires some analysis. For example, sophisticated network design may be required considering constraints over multiple network layers to fully utilize the recovery capability while maintaining high efficiency. A PCE based architecture can be applied by enhancing to apply carrier own policies, so that operations such as network planning, service provisioning and fault management are simplified and manageable.

Thirdly, GMPLS enables not only operation enhancement, such as “plug and play” and “simplified network operation”, but also offers an opportunity to provide new transport services. For example, by extending the notion of such operation enhancement, connections can be provided to customers flexibly (i.e., fast service delivery with no or little manual operations).

Note that in order to deploy GMPLS technologies to the existing transport network, migration aspects need to be considered. Migration issues from non GMPLS based networks and possible solutions will be presented, including how to migrate management system owned connections into GMPLS-controlled connections. The impact on GMPLS protocols is referred.

Biography: Tomonori Takeda received the B.E. and M.E. degrees in Electronics, Information and Communication Engineering from Waseda University, Tokyo. Currently, he is with NTT Network Service Systems Laboratories, where his work is focused on the next-generation network architecture, IP optical network architecture and related protocols. He has been involved in standardization activities, including IETF and ITU-T, and co-chairs the Layer 1 Virtual Private Network (L1VPN) working group in the IETF.

T2-2 "Operational and Maintenance Challenges for GMPLS Deployment"
Zafar Ali, Cisco Systems, Inc.
Tomohiro Otani, KDDI R&D Laboratories, Inc.

GMPLS protocols for control of optical networks are now mature enough to claim deployment. However, one of the major obstacles for GMPLS deployment faced by the service provider is “how to operate and maintain a GMPLS based IP + Optical network”. This problem is more acute when the network is based on all optical nodes, and the fact that data and control channels are separate. This presentation details these challenges faced by GMPLS, along with solutions to address them. The presentation also outlines some implementation gaps and discusses next steps to overcome obstacles for large scale deployment of GMPLS based IP+Optical networks.

In a network consisting of optical cross connects (OXC) and DWDM gear, there are requirements to perform commissioning of the links, e.g., monitoring and power-checking, before GMPLS LSP can carry user’s traffic. This requirement presents some issues that hinder GMPLS deployment. E.g., alarm reporting needs also be disabled during the commissioning phase. If the commissioning is done for a specific route that GMPLS LSP takes and when the LSP is rerouted due to a failure or reoptimized, the LSP needs to be commissioned before it can carry user traffic. These aspects need to be considered in implementing protection and restoration, or in realizing make-before-break procedure for reoptimization.

In addition to LSP commissioning requirement mentioned above, there are additional operation requirements on GMPLS technology. E.g., on one hand, GMPLS needs to communicate alarms along the path of the LSP within the optical network; on the other hand it needs to be able to provide alarm suppression, aggregation and correlation to avoid undesirable actions at multiple layers of the network. Similarly, due to traditionally heavy use of management plan in control of optical network, GMPLS needs to be able to provide more management functions that its packet counterpart. It is also an operational requirement that IP applications running on top of GMPLS LSP should not need to boot strap in the event of a failure in the optical network or during reoptimization. Trace route and LSP ping are also important OAM requirements.

GMPLS protocol extensions provide mechanisms that can be used to address the above-mentioned requirements. Nonetheless, there are implementation gaps. This presentation outlines how GMPLS is able to address the OAM requirements, what are implementation gaps and key interop issues. It summarizes the major obstacles for large scale deployment of GMPLS based IP+Optical networks.

Biography:Zafar Ali is a Technical Leader at Cisco Systems, Inc. where he leads software protocol development for Cisco Systems high speed, carrier-class routing business unit. Most recently Zafar has been focused on designing and developing GMPLS protocols and high availability solutions. Prior to joining Cisco, Zafar worked at Nortel Networks and Hughes Network Systems.
Zafar Ali is quite active as a member of the MPLS, CCAMP, BFD, IDR, RTG and other working group within IETF. He has authored a number of IETF drafts. In addition, Zafar Ali has also authored several Journal papers, conference publications, and book chapters. He is also author of several patents. Zafar has a broad range of interests include optical networking, MPLS Traffic Engineering and IP routing.
Zafar Ali received his Ph. D. and MS in Electrical and Computer Engineering from Purdue University, West Lafayette, Indiana. He obtained his Bachelor of Engineering in Electrical Engineering from NED University, Karachi, Pakistan where he was awarded the University Gold Medal.

T2-3 "IP/MPLS plug & play in GMPLS based IP Optical infrastructure."
Hidetsugu Sugiyama, Juniper Networks Inc.

One of the biggest concerns in today’s network is how to deploy GMPLS technology to realize the benefits of it. Although one benefit of GMPLS technology is the on demand optical path setup by ingress node, it is not enough to enable for the on demand dynamic IP data service unless IP/MPLS layer can support plug & play. Since SPs are trying to provide application aware network on demand service using the distributed applications with XML-RPC, SOAP-RPC, etc, the on demand based dynamic IP data network is becoming key infrastructure. In order to build more flexible infrastructure, simplifying configuration and management of IP/MPLS devices and networks will play a key role in the adoption of IP/MPLS/GMPLS in wider contexts, such as metro networks and enterprises. This paper describes current data plane issue in GMPLS IP Optical infrastructure and also details several techniques that can be applied to simplify IP/MPLS provisioning.

The outline of this presentation is as follows.

- IPv4/v6 data plane issue in GMPLS IP Optical infrastructure
- IP/MPLS plug & play for data plane
- IP/MPLS plug & play for Multi-layer Service Network in GMPLS IP Optical infrastructure
- IP/MPLS plug & play issue
- Conclusions

The goal is to bring these together in a coherent approach to reduce OpEx and advance the applicability of (G)MPLS.

Biography: Hidetsugu Sugiyama is an R&D Support Director, APAC Technical Operations at Juniper Networks and is currently focused on next generation IP Optical network technology related MPLS/GMPLS products. He has over 18th years experience in Telecommunication industry as a Network consultant engineer in data communications.

T2-4 "Management, Planning, and Provisioning of Point-to-Multipoint Transport Connections"
Daniel King, Aria Networks

Point-to-multipoint (P2MP) MPLS-TE is now an accepted technology that is approaching the completion of standardization within the IETF, and is being deployment in for live traffic by Service Providers.

Less consideration has been given to the deployment of GMPLS-based P2MP in transport networks, but P2MP technologies such as drop and continue have existed in transport networks for a long time, and new applications of content distribution make such physical connectivity techniques a reality.

The new protocol extensions for P2MP MPLS-TE are equally applicable to GMPLS. This means that optical transport networks can be enabled for P2MP services using a dynamic control plane.

This presentation will outline the signalling and routing procedures for GMPLS P2MP LSPs and will examine the requirements for traffic engineering a P2MP GMPLS network. It will compare the benefits of least cost destination and Steiner tree solutions.

Linear programming techniques traditionally used to compute paths for point-to-point LSPs do not extend well to the complexities of P2MP traffic engineering. Additional requirements to support P2P and P2MP LSPs within the same network, and to offer end-to-end protection of P2MP traffic make the planning and operational management of networks hard. The speaker will explain why what the challenges are and why it is difficult to coordinate these services using traditional network planning techniques. The presentation will conclude by explaining how holistic traffic engineering can achieve significant improvements in network optimization and the delivery of P2MP services.

Biography: Daniel King is a co-founder of Aria Networks, who provide path computation solutions for network modelling and traffic engineering. Daniel has broad experience of IP, MPLS, GMPLS, and optical networking having worked extensively with equipment vendors and Service Providers. Formerly, he worked at Movaz Networks where GMPLS control of P2MP services was pioneered. Prior to Movaz Networks he worked at Redback Networks, Cisco Systems, and Bell Labs. He co-authored RFC4687 that defines the OAM requirements for P2MP MPLS, and is co-author of several Internet-Drafts related to path computation and optimization. His research interests include the development and application of evolutionary computing and machine learning techniques to complex path computation and network optimization problems.

T3: GMPLS E-NNI
Thursday 8, June 2007. 15:30-16:30

T3-1 "Inter-carrier photonic networking developing project in Japan"
Satoru Okamoto, Keio University, Eiji Oki, NTT, Wataru Imajuku, NTT, Tomohiro Otani, KDDI and Naoaki Yamanaka, Keio University

Keio University, NTT Laboratories, and KDDI R&D Laboratories promote developing inter-carrier photonic networking technologies. This consortium is a subset of the Kei-han-na Info-Communication Open Laboratory Interoperability Working Group. The technical target of this consortium is to promote the rapid development of IP+Optical networking technologies under multi-carrier circumstances. IP+Optical networking is key to develop a new IP backbone network.

In this paper, we will present a framework of the consortium and results.The contents will be described e.g.:

• Multi-carrier IP+Optical network architecture
• Hierarchical inter-carrier interface architecture
• Interworking of PXC based Lambda Switch Capable; LSC network domains

Biography: Satoru Okamoto received the B.S., M.S. and Ph.D. degrees in electronics engineering from Hokkaido University, Hokkaido, Japan in 1986, 1988 and 1994 respectively. In 1988, he joined Nippon Telegraph and Telephone Corporation (NTT), Japan. Here, he has been engaged in research on ATM cross-connect system architectures, photonic switching systems, optical path network architectures, and developed GMPLS controlled HIKARI router systems.
He leads several GMPLS related interoperability trials in Japan, such as the Photonic Internet Lab (PIL), OIF world wide interoperability demo, and Kei-han-na Interoperability Working Group.
He moved to an Associate Professor of Keio University at 2006. He is now promoting several research projects in the photonic network area.

T3-2 "Development of GMPLS inter-carriers E-NNI prototype node based on Linux"
Hideki Otsuki and T.Morioka, NICT

1. Abstract
The Generalized Multi-Protocol Label Switching (GMPLS) [1] is a set of network control protocols and an important controlling technology for next generation’s photonic core networks. It is strongly required to support the end-to-end LSP under multi-carrier environment. However, GMPLS has some issues for interoperability between carrier domains to solve. The Interoperability working group (IWG) of Kei-han-na info-communication open laboratory studied such issues and we developed a prototype node of GMPLS E-NNI based on the proposals of IWG discussions. In this paper, we introduce the prototype node as the E-NNI border node. We also verified that LSPs is established between carrier domains using this prototype.

2. Issues of E-NNI
There are two big issues on inter-working between carrier domains. One is the difference of architecture. There are GMPLS overlay and peer network model by IETF and ASON by ITU-T. Inter-working between these different networks, translation of single-session to/from multi-session of RSVP-TE signaling is required. Another one is routing protocol for E-NNI. Currently, routing protocol for inter domain is not standardized. Routing protocols based on IGP like as OSPF-TE may have the problem of scalability and policy.

3. Prototype node
We developed a prototype as the E-NNI node for GMPLS networks. This prototype runs on Linux and Zebra [2] is used as routing engine. The functions of network model translation and E-NNI routing are implemented. For network model translation, objects of UNI information are translated at E-NNI node. This function make possible to connect single-session and multi-session. And E-NNI nodes hold LSP association to connecting multi-session LSPs. For E-NNI routing, BGP is extended to carry TE information. BGP advertises TE information in the Path attribute of UPDATE messages. Corresponding to separation of E-NNI and I-NNI routing, route calculation of E-NNI and I-NNI is also separated each other, Policy control between E-NNI and I-NNI is also possible. Furthermore, this prototype is now able to be control physical equipments such as OXC, L2 switche via telnet interface.

4. Trials with prototype
We evaluate this prototype in multi-carriers environment. In this field test, a nation wide testbed , Japan Gigabit Network II (JGN II) is used as both of C-plane and D-plane. We made 3 carrier domains and prepared both of IETF GMPLS network model and ASON model. Each carrier has this prototype as the border (E-NNI) node and these networks are connected via E-NNI node. However, physical equipments were not used under the E-NNI prototype. We verified establishing end-to-end LSC LSPs between ASON-GMPLS, ASON-ASON and GMPLS-GMPLS network models.

5. Conclusion
IWG studied issues of GMPLS E-NNI for interoperability and developed a prototype node. This prototype runs on Linux with Zebra. We evaluate the prototype node in multi-carrier environment filed test using nation wide area testbed JGN II. LSPs between ASON-GMPLS, ASON-ASON and GMPLS-GMPLS architecture domains are established. IWG was finished first term in last year and we are just started second term for further issues.

6. References
[1] E. Mannie (Editor), “Generalized Multi-Protocol Label Switching (GMPLS) Architecture,” IETF RFC3945, Oct. 2004.
[2] http://www.zebra.org/

Biography: Hideki Otsuki was born in Kanagawa, Japan on March 11, 1967. He received B.E and M.E degrees from Musashi Institute of Technology, Tokyo, Japan, in 1989 and 1991, respectively. Then he joined the Communications Research Laboratory, Ministry of Posts and Telecommunication, (Now, National Institute of Information and Communications Technology: NICT). He received the Doctor Degree of Electrical Engineering from Tokyo Institute of Technology, Tokyo, Japan, in 2002. His major field of study is network architecture. He has been involved in Intelligent Networks since 1993. He developed the Intelligent Network Service Simulator for B-ISDN, and a testbed with ATM switches. He has supported ATM satellite communication experiments. He continues to study Internet transport and active-net technology. His current research interest is GMPLS protocols.

T3-3 "GMPLS inter-domain network control and management"
Shuichi Okamoto, KDDI R&D Labs., T. Otani, NICT

Due to the explosion of various bandwidth demanding applications, 10 Gigabit Ethernet (10GbE) service is getting significant even over geographically distributed locations and it is a challenge to globally and dynamically control and manage such a high bandwidth connection. With the consideration of such a situation, GMPLS inter-domain network control and management is a key to develop, since global network connection is usually established over multiple administrative domains on demand.

This presentation describes the architecture and requirements for GMPLS inter-domain
network control and management to establish the global 10GbE connectivity over multiple domains.
The basic architecture considering domain-to-domain operation independence will be presented, and some requirements to extend GMPLS protocols will also be touched upon to achieve such
operational environment.

The early field trial of dynamic 10GbE path provisioning was conducted so far in order to
support scientific experiments over multiple domains between the US and Japan, and the results and operational experience will be shared in the presentation. We will also show our developed GMPLS LSP management tool to supervise the dynamic path creation and deletion over multiple domains.

Biography: Shuichi Okamoto received the B.E. and M.E. degrees from Osaka University, Japan, in 2000 and 2003, respectively. He joined KDDI R&D Laboratories, Inc. in 2003. He was a researcher of the National Institute of Information and Communication Technology (NICT) Tsukuba Research Center from 2004 to 2006. Currently he is a reseacher of KDDI R&D Laboratories Inc., and investigating the GMPLS technology from the operational point of view.

T4: GMPLS technology
Thursday 8, June 2007. 16:40-17:40

T4-1 "A ROUTING ALGORITHM FOR TRANSLUCENT OPTICAL NETWORKS"
Giovanni Fiaschi and Diego Caviglia , Ericsson

Recent progress in optical and photonics technology has made possible the realization of all-optical networks and interconnection systems (OXC).

While offering clear cost advantages with respect to full electronic regeneration (opaque) networks, all-optical (transparent) networks present well known limitations in lightpath length, due to physical impairments, and in wavelength assignment, due to limited wavelength availability and matching restrictions.

For this reason, it is preferred to provide a certain conversion and/or regeneration capability in the network nodes to realize “translucent” networks. To keep the advantage of equipment cost reduction, this capability is added in a limited amount.

Recent studies propose translucent networks implemented as either sparsely placed opaque nodes or with translucent OXC nodes. The method here presented assumes the latter architecture, where each node implements a certain amount of electronic regeneration. This choice is motivated by a higher flexibility and scalability of the solution. Moreover, simulations show that flexibility in wavelength choice at the network access point is far more efficient than wavelength conversion within the network.

As the conversion / regeneration capacity of a translucent node is limited, it is obviously important to utilize it to the best, assigning it in accordance with criteria that allow good operation of the network and at the same time minimize the chances of blocking of the network because of exhaustion of said capability in the nodes of the network.

In itself, the problem of automatic routing of circuits in transport networks is well known and belongs to the fields of network management and network control planes (ASTN and GMPLS).

In a translucent node network, the additional requirement is to complement the routing algorithms with functions dealing with optimized electrical conversion. The target is to minimize the probability of blocking of the network due to exhaustion of the limited conversion and/or regeneration capabilities.

The purpose of the method here presented is to enhance known routing algorithms to adapt them to take into consideration in an effective manner the existence of partial conversion / regeneration capabilities in the network nodes.

The presented method (patent pending) is a practical solution to the problem, in that it offers no demonstration of optimality, but the produced solutions are correct and reasonably good. In addition, the algorithm has the advantage of simplicity and the enhancements to known routing algorithms use only node local decisions, thus reducing computation and communication overhead over the control network.

Biography: Giovanni Fiaschi works at Ericsson in the Product Line of Optical Core, his main expertise being connection oriented packet transport and control plane for TDM and optical switching. He authored several patents and followed joint research activity with universities and technological partners. Previously he carried out specification activity for access and packet products, where he gained experience on ATM and IP network protocols and attended IETF meetings. Earlier, he was involved in system design of Management Systems of the SDH network, participating to joint study groups with customers and partners and to international standardization bodies (ETSI and ITU-T). Giovanni Fiaschi received a degree in Computer Science in 1987 from the University of Pisa.

T4-2 "GMPLS-based Reliable Photonic Cross Connect"
Eiichi Horiuchi, Yoshimasa Baba, Sota Yoshida, Hiroyuki Sato, Kiyoshi Onohara, Yuji Akiyama, and Takashi Mizuochi, Mitsubishi Electric Corp.

By the rapid spread of optical access networks and movie sites on the Internet, the Internet’s traffic keeps increasing. Therefore, the next generation optical transport network is required to support the function of dynamic optical path establishment in response to changing traffic demands, and to support fast recovery mechanisms when a transport failure occurs.

We have developed a prototype of the Photonic Cross Connect (PXC), which enables setup and
change of optical wavelength paths by using all-optical switches. It supports GMPLS (Generalized Multi-Protocol Label Switching) protocols for control and management of the photonic network. It can establish reliable optical paths cooperating with core routers and WDM equipments. Based upon reliable architecture such as redundant control cards and optical switching fabrics, it realizes carrier-grade high availability.

This presentation shows our approach for high availability of the data plane and the control plane, which are features of our PXC prototype. Figure 1 shows the architecture of the PXC prototype, consisting of 1+1 redundant control cards, optical interface cards, and 1+1 redundant optical switches which have 128 x 128 ports. Input signal is split into a working optical switch and a backup optical switch by using a 3dB coupler on the input side. One of the signals is selected as output signal by using the 2 x 1 switch on the output side.

From a point of view of reliable control plane, 1+1 redundant control cards are connected via redundant HA (High Availability) links, for taking over the control data such as GMPLS signaling data. The control lines between control cards and optical switches are also redundant. When the working control card can’t control the optical switches because of the control line failure, the backup control card takes over the control from the working control card.

At the end of this presentation, we will describe evaluation tests of the PXC such as performance evaluation of GMPLS based protection and restoration.

Fig.1 Architecture of the PXC prototype

Biography: Eiichi Horiuchi received the B.E. degree in electronics engineering from Osaka University, Toyonaka, Japan in 1990. In 1990, he joined Communications System R&D Center, Mitsubishi Electric Corporation, Kamakura, Japan. Since 2001 he has been engaged in research and development of IP-optical networking. He is a member of Information and Communication Engineers of Japan (IEICE).

T4-3 "Distributed optical layer monitoring"
Mark Lourie, Aegis Lightwave Inc.

Core optical networks are evolving to a new transport architecture where native Ethernet is being transmitted directly over DWDM, without SONET and its associated performance monitoring features. Lower cost is one of the strong drivers of this streamlined architecture. Another benefit is the open architecture where the on-ramps and off-ramps of the optical layer are colorless and can accept a multitude of client equipment regardless vendor, native protocol and modulation format. This increased transparency in the optical network drives the need for distributed optical layer monitoring to achieve a richer control plane.This paper will describe how DWDM channel monitoring provides the foundation of the measurement plane which in turn serves the control plane with all the relevant information to achieve fault isolation, fault correlation and wavelength inventory. The control plane can then provide feedback to network elements for real-time optimization of fiber-optic transmission. The control plane can also pass along highly correlated information to the management plane for the purposes of optical layer alarming and end-to-end wavelength tracking. To enable a ubiquitous measurement plane, new low-cost monitoring technologies have been deployed at the ingress and egress of many optical network elements with minimal impact to overall network costs.

Biography: Mark Lourie is responsible for product and corporate marketing at Aegis Lightwave. Since 2001, he has defined and taken to market low-cost tunable devices and systems for reconfigurable DWDM fiber-optic network applications. Mark joined Aegis from Lucent Technologies, where he led market development activities for optical networking products in the service provider market. Prior to Lucent, he held product marketing roles at Corning and before that at EPITAXX (now JDSU). Mark Lourie holds an MSEE from the University of Southern California.

Program at a glance

Friday 8, June 2007. 17:40-17:55
Naoaki Yamanaka, Organization Committee Chair, Keio Univ.

Program at a glance