| Project/Area Number |
09680429
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
社会システム工学
|
|---|
| Research Institution | TOKYO METOROPOLITAN UNIVERSITY |
|---|
Principal Investigator |
IWASAKI Kazuhiko Tokyo Metropolitan Univ. Prof., 大学院・工学研究科, 教授 (40232649)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
DESAKI Yoshihisa Tokyo Metropolitan Univ. Assistant Prof., 大学院・工学研究科, 助手 (50291571)
MIURA Yukiya Tokyo Metropolitan Univ. Associate Prof., 大学院・工学研究科, 助教授 (00254152)
|
|---|
| Project Period (FY) |
1997 – 1999
|
| Project Status |
Completed (Fiscal Year 1999)
|
| Budget Amount *help |
¥3,600,000 (Direct Cost: ¥3,600,000)
Fiscal Year 1999: ¥500,000 (Direct Cost: ¥500,000)
Fiscal Year 1998: ¥500,000 (Direct Cost: ¥500,000)
Fiscal Year 1997: ¥2,600,000 (Direct Cost: ¥2,600,000)
|
|---|
| Keywords | Metropolitan Area Network / Local Area Network / Dependable Computing / Packet Loss / Convolutional Code / Multicast / TCP / IP Packet / FPGA / インターネット / ネットワーク / フォールトトレランス / 誤り検出 / 障害解析 |
|---|
| Research Abstract |
One purpose for this research is to analyze how failures occur in metropolitan and Local are networks. We also developed fault-tolerance designs for network systems that can improve the dependability. Results obtained from 1997 to 1999 are briefly described in the following.
We investigated the error detecting capability of the Internet checksum used in TCP and IP. We modeled the set of packets which are regarded error-free in a receiver as a nonlinear code and computed a subset of the distance distribution of the code. We also evaluated the lower and upper bounds on the exact undetectable error probability.
We developed an estimation software tool which uses UDP datagrams as probing packets. Using the tool we measure how Internet packets are lost. The measurements are done between Tokyo Metropolitan University and other four universities. As a result, we modeled the burst packet losses occurred among five universities.
A new method to recover packet losses using convolutional codes is proposed. An independent erasure channel is assumed for a packet loss model. The sufficient conditions are theoretically derived for packet losses that can be recovered. We compared the proposed method with a simple parity method having the same redundancy. It is shown that our method has the better performance. An application to multicasting is also studied.
With the advent of the Internet, the processing of TCP/IP is becoming to be performed on an application specific devices. Using VHDL we designed a TCP/IP processing unit. The unit can execute a subset of TCP/IP and POP3 functions. It is implemented on an FPGA device that can integrate about 20,000 gates.
|
