1992 Fiscal Year Final Research Report Summary
Fault Tolerance of PC-Server for Computer Lab
| Project/Area Number |
03680251
|
|---|
| Research Category |
Grant-in-Aid for General Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
科学教育(含教育工学)
|
|---|
| Research Institution | Musashi Institute of Technology |
|---|
Principal Investigator |
MATSUYAMA Minoru Musashi Inst. of Tech., Fac.of Engineering, Associate Professor, 工学部, 助教授 (60102703)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
YOKOI Toshiaki Musashi Inst. of Tech., Fac.of Engineering, Associate Professor, 工学部, 助教授 (80182682)
|
|---|
| Project Period (FY) |
1991 – 1992
|
| Keywords | PC-server / Fault tolerance / Fault avoidance / Multi-address communication / Computer lab / Report-reception system |
|---|
| Research Abstract |
The purpose of the present work is to investigate a feasible mothdology of the fault-tolerance of PC-server used in a computer laboratory for educational purpose, and is to embody such technique into a real system. To realize the fault-tolerance, various arrangements of equipments and techniques can be selected according to object and level of fault-tolerance. In this work, two PC's were employed as servers because PC's were low in price. The fault-tolerance can be achieved by multi-address communication from a PC-terminal to the two servers and recording the same information in each. At first, developed were programs by which a file transmission take S place from a terminal to the servers. It was confirmed in a test that the server function continued even after one of the servers was ceased purposely. The function of the fault-tolerance of the programs were included into the report submission/reception system developed and operated previously. The improved report system was tested, and is under operation at the present days. To absorb the difference in operational speed between two servers, and to record the same information in each, synchronizing signals must be sent and received between two servers. This caused the deterioration in response, although it seemed that the students did not sense it. In general, a hard-disk unit with many mechanical-moving parts is often out of order compared to other units. In this work, therefore, the report system was further improved to decrease the number of access time S to hard disks of servers and hence to increase the fault-avoidance by exploiting the RAM-disk as a disk-cache. In the present work, the fault-tolerance was achieved by software for PC's on market without any hardware modification. Thus, the system is easily transferred to new PC's which are usually replaced every few years.
|
