Initialising ...
Initialising ...
Initialising ...
Initialising ...
Initialising ...
Initialising ...
Initialising ...
Miyamura, Hiroko; Hu, H.-Y.*; Yoshida, Masahiro*; Ozahata, Satoshi*; Nakao, Akihiro*; Takahashi, Shigeo*
Shingaku Giho, 112(463), p.577 - 582, 2013/03
This report presents a method for visualizing scale-free networks including Overlay networks as an typical example. Visualizing large-scale and complicated networks such as social networks has recently been very popular while conventional network visualization techniques cannot allow us to understand the topological structure of the scale-free networks. This is because the vertex degrees vary at an exponential rate in the scale-free network and thus special attention should be given when visualizing the network connectivity and analyzing the network traffic there. In this report, we employ the hierarchical representation of the scale-free network by referring to their vertex degrees and autonomous system relationships, so that we can clearly visualize the topological structure of the network in 3D space and retrieve the traffic paths over the network.
Miyamura, Hiroko; Ozahata, Satoshi*; Nakao, Akihiro*; Kawashima, Konosuke*
Shingaku Giho, 110(190), p.103 - 108, 2010/09
In this paper, we propose a visualization technique for handling large-scale experiment datasets in distributed file system. Traditional network visualization techniques have problem that not all data can be displayed at the same time because the file-sharing network data is too large-scale. Therefore, we propose a concept of adaptive network graph display, in which the graph style can be changed according to the details of user's observation. We construct a visualization system based on this concept. When observing data globally, the proposed system selectively displays information based on the clustering result, and when observing data locally, the system displays detailed information. This technique is a basic technology for computational science to achieve large-scale datasets handling which is a problem in atomic energy related fields.
Miyamura, Hiroko; Ozahata, Satoshi*; Nakao, Akihiro*; Kawashima, Konosuke*; Suzuki, Yoshio
Shingaku Giho, 109(448), p.357 - 362, 2010/03
We propose a visualization system for file-sharing network data. The file-sharing networks have some problems, for example: the contents distribution of the copyright infringement files, an increase in the amount of traffic, and so on. To handle these problems, techniques for observing the network state and controlling the file sharing are proposed. In addition, the proposed techniques realize effective control by giving retrieval control to the key nodes, which are important roles in the file-sharing network.
Miyamura, Hiroko; Ozahata, Satoshi*; Nakao, Akihiro*; Kawashima, Konosuke*; Suzuki, Yoshio
Shingaku Giho, 109(188), p.85 - 90, 2009/09
We propose a network visualization method for a large-scale file-sharing network dataset. The network dataset has nodes and links that represent the connection between two nodes. When the number of links is large, it is difficult to recognize the structure of the network data, because many links are crossed. In this context, for recognizing the structure and exploring the key nodes, we propose a link-less network visualization method using matrix based representation. In addition, we demonstrate the effectiveness of the proposed method by applying it to the file-sharing network log data.
Miyamura, Hiroko; Yoshida, Masahiro*; Ozahata, Satoshi*; Takahashi, Shigeo*; Nakao, Akihiro*; Kawashima, Konosuke*
no journal, ,
We propose a multilevel graph layout technique for visualizing large scale network datasets. Large-scale datasets are said to be difficult to analyze and visualize. However, with our technique, the datasets can be displayed in multistep. For instance the outline of network dataset is shown first, and when focused on a certain region, the detailed information is shown as well. By using this technique, users can observe a large scale dataset from the outline to the detail seamlessly, and therefore this is valuable for the structural analysis of network datasets.