An efficient fault-tolerant multicast technique for hypercubes /
Abstract: "This paper proposes a fault-tolerant multicast technique, called the acknowledged DIPAT (AIPAT), for the hypercube multicomputers. AIPAT is extended from the DImension PAtition Technique (DIPAT) [1] for efficient detection of message loss due to transient faults. A DIPAT-graph is...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Book |
| Language: | English |
| Published: |
College Station, Tex. :
Texas A & M University, Computer Science Dept.,
[1993]
|
| Series: | Technical report (Texas A & M University. Computer Science Department) ;
93-026. |
| Subjects: |
| Summary: | Abstract: "This paper proposes a fault-tolerant multicast technique, called the acknowledged DIPAT (AIPAT), for the hypercube multicomputers. AIPAT is extended from the DImension PAtition Technique (DIPAT) [1] for efficient detection of message loss due to transient faults. A DIPAT-graph is first constructed from a multicast set, and any node in the multicast set can reach every destination node along a spanning tree. Since a lage number of rings exist in a typical DIPAT-graph, the leaf nodes of the spanning tree are likely connected with each other. Therefore, the acknowledgment (ACK-) messages can be passed between the leaf nodes along the links between them and eventually only one ACK-message is needed to be sent back to the sender of the multicast message. In this way, the average number of ACK-messages is significantly reduced, and any message loss can be detected effectively. The multicast algorithm can also be integrated with high level activities, like concurrency control, synchronization, cache coherence management, etc., in a fair, symmetric manner." |
|---|---|
| Item Description: | "April 29, 1993." |
| Physical Description: | 20 leaves : illustrations ; 28 cm. |
| Bibliography: | Includes bibliographical references. |