Yasuhiro Mizukoshi

Torus-Topology Data Center Network Based on Optical Packet/Agile Circuit Switching with Intelligent Flow Management

We review our work on an intra-data center (DC) network based on co-deployment of optical packet switching (OPS) and optical circuit switching (OCS), conducted within the framework of a five-year-long national R&D program in Japan (~March 2016). For the starter, preceding works relevant to optical switching technologies in intra-DC networks are briefly reviewed. Next, we present the architecture of our torus-topology OPS and agile OCS intra-DC network, together with a new flow management concept, where instantaneous optical path on-demand, so-called Express Path is established. Then, our hybrid optoelectronic packet router (HOPR), which handles 100 Gbps (25 Gbps × 4-wavelength) optical packets and its enabling device and sub-system technologies are presented. The HOPR aims at a high energy-efficiency of 0.09 [W/Gbps] and low-latency of 100 ns regime. Next, we provide the contention resolution strategies in the OPS and agile OCS network and present the performance analysis with the simulation results. It is followed by the discussions on the power consumption of intra-DC networks. We compare the power consumption and the throughput of a conventional fat-tree topology with the N-dimensional torus topology. Finally, for further power saving, we propose a new scheme, which shuts off HOPR buffers according to the server operation status.

Flexible and automated operational control in SDN transport-base virtual router

We propose and demonstrate a virtual router based on SDN transport which has automatic topology discovery function. We show that the virtual router contributes to flexible and automated operational control.

Study on performance of networked dead reckoning for real vehicles

Vehicle-to-vehicle (V2V) communication technology based on Wi-Fi technology has recently been used to prevent vehicles from colliding with each other. However, in V2V communications, radio waves cannot be diffracted well enough at intersections at which vehicles cannot detect each other. Therefore, such technology might not prevent traffic accidents from occurring. In this study, we used mobile networks to exchange location information, such as longitude and latitude, every 100 msec and enable vehicles to use the information to predict the locations of surrounding vehicles of near future and avoid traffic accidents in any situation. However, the LTE network will become overloaded with radio resources, and data fees might become too expensive. Therefore, we have to reduce the amount of data packets to 2% and maintain location accuracy in the location-management servers at the same time. In this study, by using networked dead reckoning and our vehicle motion model, we developed a data reduction/compression and communication method to reduce the uploading frequency of location information from clients to servers in the LTE network and maintain the accuracy of location prediction at location-management severs.

Short-term rainfall attenuation prediction for wireless communication

This paper describes a method for predicting short-term rain attenuation for ground wireless communications. We propose machine learning methods using the k-nearest neighbour (KNN) and artificial neural network (ANN). We used a time-series of link attenuation calculated from a set of high-precision rainfall radar maps as training data. These two methods were verified with data from a different rainfall event. The results show that the two methods provide rainfall attenuation for wireless communications with sufficient accuracy.