Presentations

In wide area Layer 2 (L2) networks, there is a problem that broadcast packets for address resolution fill up link capacities. As a result, address resolution using Address Resolution Protocol (ARP), which uses broadcasts, cannot be used in wide area L2 networks. A conventional research has proposed an architecture using Distributed Hash Table (DHT) and ARP cache for reducing broadcasts. However, since this architecture uses simple Least Recently Used (LRU) method for caching, some packets may be discarded because of continuous cache misses which require long address resolution response time. Therefore, we propose a cache retention time change method considering address resolution response time. In addition, we also propose cache insertion position determination methods based on several ideas. Finally, we evaluate the effectiveness of the proposed method.

In order to respond to the rapidly increasing traffic and the diversification of communication services in recent years, the Photonic Layer 2 (L2) network attracts attention. The aim of the photonic L2 network is providing various services on a common photonic L2 with high-speed transmission by light paths. A wide area L2 network assuming a carrier network capable of accommodating one million users is considered as one of the services provided by the photonic L2 network. In wide area L2 networks, however, there is a problems that broadcast packets for Address Resolution Protocol (ARP) fill up link capacities due to the enormous number of users and packet losses due to buffer overflow by accumulation of long address resolution time. In this paper, we propose hierarchical address resolution architecture with Distributed Hash Table (DHT) in wide area L2 network to reduce broadcast, and caching method which based on the retention time to reduce address resolution time in the proposed architecture.

This paper introduces the remote control of an industrial robot performing surface blending by using SDN-based optical transport networks. The SDN orchestrator enables the industrial robot located on the production floor and the robot control system located in the cloud/edge network to jointly and successfully carry out an unstructured task while minimizing both data transfer latency and volume across the transport network. Finally, the paper shows experimental validation of the proposed system.

次世代光メトロ・アクセス統合型ネットワーク「エラスティックλアグリゲーションネットワーク（EλAN）」が提案されている. EλANにおいて大規模災害時の接続性を確保するために, アクティブ型のOptical Distribution Network (ODN)を活用し, 通信途絶に陥ったOptical Network Unit (ONU)が1台以上のRemote ONU (R-ONU)を経由して生存Optical Line Terminal (OLT)まで通信を確立する光アドホック通信が提案されている. 本稿では, R-ONUが隣接ONU群内のR-ONUを発見する際のルーチング手法に関して検討を行う.

複数のハードウェア(CPU, FPGA, LSI等)の組み合わせにより構成される再構成可能通信処理プロセッサを用いたプログラマブル光エッジ(PPE)が提案されている[1]. 本稿ではPPEの仮想化を提案, QoSおよび省電力性の観点からリソース割当アルゴリズムの要件を明確化する.

In recent years large-scale parallel distributed processing by clustering such as Hadoop is executed frequently in the data center. Depending on the job(processing content), the network load is large and an increase in power consumption is regarded as a problem. The authors are working on the reserch and have found a discriminate jobs with such a heavy network load. With using it, we develop the thereby reducing the power consumption Data Center Network. In this paper, we describe the method of determining the network load and the operation method of the proposed architecture, and then evaluated the proposed architecture in terms of power comsumption. The result of the experiment shows that the proposed architecture achieves drastically higher energy consumption reduction than it was.

In the elastic lambda aggregation network (EλAN), access paths are flexibly reconfigured by the active optical distribution network (ODN) when a network failure occurs. It is expected that an optical line terminal (OLT) of another central office (CO) recovers the connectivity of optical network units (ONU). However, there is an upper limit on the number of ONUs that a single OLT can accommodate. Therefore, some of ONUs cannot be recovered in a large-scale disaster where many COs become out of service. In order to overcome this limit, the time division multiple access (TDMA)-based logical OLT sharing method among ONU groups has been proposed. The proposed method maintains connectivity of the excessive numbers of ONUs from a single OLT. In this paper, we report the experiment results of the restoration and the TDMA-based OLT sharing method conducted on the EλAN prototype system.

In the elastic lambda aggregation network (EλAN), access paths are flexibly reconfigured by the active optical distribution network (ODN) when a network failure occurs. It is expected that an optical line terminal (OLT) of another central office (CO) recovers the connectivity of optical network units (ONU). The optical ad hoc networking has been proposed as a method to overcome the transmission distance limitation between the OLT and the ONU in the case of a large-scale disaster in which a large number of COs are out of service. In this paper, we propose the wavelength reservation method to realize the optical ad hoc networking. By computer simulation, the optical ad hoc networking applying the proposed wavelength selection method achieves ONU group rescue rate higher than 20% the restoration method when the failure rate of the network device is 40%.

In the Software-Defined Transport Network (SDTN), there are some problems in a conventional approach which is controlled by a single top orchestrator in case of multi-carrier environment. Therefore, we propose a scale-out type SDTN approach based on assumption that each carrier has its own top orchestrator in the future, and also define an Orchestrator-to-Orchestrator Interface taking into account the confidentiality of information between carriers. We report that the effectiveness of the proposed approach is confirmed compared to the conventional SDTN approach by an evaluation, and the feasibility is also verified by implementation of the proof of concept system.

The concept of Programmable Photonic Edge (PPE) has been proposed which uses reconfigurable communication processers composed of multiple hardware such as CPU, FPGA, and LSI. This paper proposes virtualization of PPE for the sake of high utilization rate of these hardware as resource pool. We formulate the optimization problem of hardware resource allocation for edge computing in terms of QoS and power consumption, and clarify the requirements of the algorithm development with the performance analysis of two simple algorithms as basis examination.

Enhancing quality of service that perceived by customers is an important measure for web service providers. Major web service providers create their infrastructure by locating several computing resources inside the Internet to assist its quality enhancement. For services using Internet of Things (IoT), there are strong interests in controlling IoT devices through network. We propose network assisting functions and supporting architecture for such services. In our proposal, we spin off parts of IoT device or server functions and locate them on network located computing resources. In this report, we compare and evaluate our approach with other method using real devices and virtual network test bed.

Hadoop has swiftly been accepted in both industry and academia in order to offer enormous storage for any kind of data, massive processing power and the capability to manage virtually jobs. This paper designs a Hadoop system, which considers both slave server's processing capacity and network delay for wide area networks in order to reduce the job processing time. The demonstrated testbed results esignate that the task allocation scheme in our designed Hadoop system is effective.

Recently, the energy consumption in data centers is increasing. Introducing optical circuit to the inter-rack communication is considered an effective solution. The authors are working on the research and development of the High-speed Optical Layer 1 Switch system for Time slot switching based optical data center networks (HOLST). In this paper, several ideas on how to connect PLZT optical switches and MEMS optical switches, the composing modules of the HOLST, are proposed and evaluated in order to design the optimal architecture. The result of the numerical simulation shows that the proposed architecture achieves significantly higher energy consumption reduction than the conventional architecture.

再構成可能通信処理プロセッサをリソースプールとして取り扱い, 網内のリソースプールを光網で接続してサービス用仮想通信処理プロセッサ(VRCP)を構築することを提案している. リソースプールから処理モジュールを選択して割当てる手法として, 省電力優先及びQoS優先の準最適化アルゴリズムを提案した. 本稿では, 動的にVRCP再構成を行うことでより一層の最適性を実現することを目的として, リソース割当に利用可能なゆらぎ制御手法を分散制御可能とした分散ゆらぎ制御でVRCPリソース割当を実現する手法を提案する.

We have proposed a new data-center network architecture “HOLST” with dynamic and adaptive network configuration for reducing power consumption. This demonstration shows the method of detecting and assigning flows from traffic and application trigger.