Yusuke Fukushima

Fault-Tolerant Routing Algorithm for Network on Chip without Virtual Channels

Constructing 2D mesh topology network on chips (NoCs) without using virtual channels becomes attractive approach to building future massive multi-core computer systems because of its large amount of bandwidths, less design complexity, and less space consumption of routers. Dead lock problem on NoC is critical because it makes data transmission between nodes unreachable, and inevitable failures in hardware make mesh topology irregular. Although several fault-tolerant techniques are available, deadlock-free routing control algorithm for irregular mesh topology is promising approach to utilize large amount of bandwidths of NoC. The main drawback of available routing control algorithms is that many healthy nodes are deactivated to guarantee deadlock-freeness, and a number of deactivated nodes lead to traffic congestion. In this paper, we propose new fault-tolerant routing algorithm on 2D mesh topology NoC constructed without using virtual channels. The proposed algorithm is fully analyzed its dead lock-freeness, and the experimental result shows that the proposed algorithm can achieve both less number of deactivated nodes and higher throughput.

A genetic approach for the reconfiguration of degradable processor arrays

This paper discusses the NP-complete problem of reconfiguring 2D degradable processor arrays under the row and column rerouting constraints. One major drawback of the previous algorithms is that a rerouting scheme to construct logical rows/columns can be applied only in one direction, either row or column direction, to handle the difficulty of this reconfiguration problem. We propose a new approach for the reconfiguration of degradable processor arrays employing a genetic algorithm (GA). The GA is used to evolve rerouting strategies for constructing logical rows/columns. A strategy based rerouting scheme is proposed to reroute the interconnections of fault-free PEs in both row and column directions. The performances of our algorithm are compared with previous studies and it indicates that the proposed algorithm achieves better results in terms of harvest and degradation.

Dynamic mobile sensor network platform for ID-based communication

This paper presents the design of a dynamic mobile sensor network platform, consisting of mobile sensors, mobile sensor gateways, and sink servers. The sensor network platform supports heterogeneous protocols in the network layer and performs ID-based communication to deliver sensor data from the mobile sensors to the sinks, as well as to send control and monitoring commands from a sensor administrator to the mobile sensors. To reliably provide sensor data irrespective of their locations, the mobile sensors and mobile sensor gateways natively support mobility and multihoming and possess network access authentication and data transport security functions. The users can freely install new application programs over an already-deployed sensor network. They can configure the sensor network to operate in light-weight or full-function modes depending on the application requirements or available networking environments. Some components of the proposed sensor network platform are already in the standardization process in ITU and others can be brought for standardization in the near future.

An Improved Reconfiguration Method for Degradable Processor Arrays Using Genetic Algorithm

This paper addresses the NP-complete problem of reconfiguring two-dimensional degradable processor arrays under the row and column rerouting constraint. One promising approach to this problem is to treat the reconfiguration problem as a combinatorial optimization problem of finding the set of rerouting rules for all rows/columns and employ a genetic algorithm (GA) to obtain an optimal solution (Fukushi et al., 2005 ). However, major drawback of this method is poor utilization of processing elements (PEs) in the reconfiguration process. In this paper, the previous method (Fukushi et al., 2005) for efficient reconfiguration was improved. The key idea is to treat the reconfiguration problem as an optimization problem of determining routing directions for all faulty PEs. A new rerouting scheme is also proposed to reroute logical rows/columns efficiently. Experimental study shows that the proposed method produces good results in terms of the percentage of harvest and degradation

New mobility paradigm with ID/locator split in the future Internet

Introduction of ID/locator split into the future Internet architecture can make distributed and dynamic mobility management functions simpler and native. Unlike conventional IP mobility solutions, the new approach can offer smooth mobility across heterogeneous network-layer protocols (e.g. IPv4, IPv6) and other benefits. This paper first gives an overview of new features that ID/locator split-based architecture can bring about besides overcoming the limitations of conventional IP mobility protocols and then discusses the architecture, implementation, and performance of HIMALIS as an example of ID/locator split-based architecture. The performance results validate that HIMALIS has end-to-end throughputs and latencies similar to those of the conventional TCP/IP network. Moreover, the results verify that HIMALIS is capable to provide seamless handover across heterogeneous network-layer protocols.

Routing Algorithms for Packet/Circuit Switching in Optical Multi-log2N Networks

The multi-log2N network architecture is attractive for constructing optical switches, and the related routing algorithms are critical for the operation and efficiency of such switches. Although several routing algorithms have been proposed for multi-log2N networks, a full performance comparison among them has not been published up to now. Thus, we rectify this omission by providing such a comparison in terms of blocking probability, time complexity, hardware cost and load balancing capability. Notice that the load balance is important for reducing the peak power requirement of a switch, so we also propose in this paper a new routing algorithm for optical multi-log2N networks to achieve a better load balance.

Implementation of communication path recovery mechanism in a multihomed ID/locator-split network

To promptly identify a path failure and transfer an ongoing data stream to another lively path, we have proposed a path failure detection and recovery mechanism in an ID/locator-split network architecture, called HIMALIS. Unlike traditional solutions, the proposed mechanism provides better resiliency to path failure independently of the network and transport layers protocols. This paper presents an implementation overview of the mechanism. The implemented functions are verified and evaluated in a 21-node-scale network-emulation environment. The results show that this mechanism provides fast failure recovery under both site- and host-multihoming configurations.

ID-based communication framework for new generation networks

The current Internet follows a location-based communication paradigm, thus is being insufficient to satisfy lately emerged communication services requirements, namely mobility, security, and scalability. Therefore, the new-generation network or future Internet architectures research has been pursuing the identifier (ID)-based communication paradigm where the locations of communicating devices that are requesting for services or that are providing the requested services do not have any significant role in the network protocol. The communication objects such as devices, contents, users, and services are assigned with IDs, which are used for identification, authentication, discovery, and delivery of services. In this paper, we first present a generalized ID-based communication framework and then show how the recently proposed ID/locator split-based networking and information-centric networking architectures fit into the framework. We then summarize the features and challenges brought about by the new proposal as well as some possible solutions to address the challenges.

Self-routing design of nonblocking WDM switches based on arrayed waveguide grating

Arrayed waveguide grating (AWG) is a promising technology for constructing high-speed large-capacity WDM switches, because it can switch fast, is scalable to large size and consumes little power. To take the full advantages of high-speed AWG, the routing control of a massive AWG-based switch should be as simple as possible. In this paper, we focus on the self-routing design of AWG-based switches with O(1) constant routing complexity and propose a novel construction of self-routing AWG switches that can guarantee the attractive nonblocking property for both the wavelength-to-wavelength and wavelength-to-fiber request models. It is expected that the proposed construction will be useful for the design and all-optical implementation of future ultra high-speed optical packet/burst switches.

Scalable Directory Service for IoT Applications

Unprecedentedly, a massive number of devices are getting connected to the network in the new era of the Internet of Things. For the discovery, authentication,, secure access and management of these Internet of Things devices, an Internet of Things directory service is needed. Its responsibility would be to store and provide the attributes of Internet of Things devices, such as identifier, location, owners name, generated data types, and security parameters. In this article, we present the architectural design of an Internet of Things directory service that is capable of storing a huge number of records and providing functions for their fast lookup (latency of a few milliseconds) and dynamic update (latency of a few seconds), while fully complying with owner-centric security and privacy policies. To demonstrate its feasibility and overall quality, we also discuss the experimental results obtained from a proof-of-concept implementation of the record database and related communication interfaces.