Topic
IBM Parallel Sysplex
About: IBM Parallel Sysplex is a research topic. Over the lifetime, 3 publications have been published within this topic receiving 19 citations. The topic is also known as: Sysplex.
Papers
Proceedings Article•
1 Jan 2000
TL;DR: The extension of a middleware infrastructure designed for high performance computing with efficient checkpointing mechanisms is described and the satisfiability checking problem for boolean formulae is considered as an example application from the realm of Symbolic Computation.
Abstract: In this paper we address the issue of dependable distributed high performance computing in the field of Symbolic Computation. We describe the extension of a middleware infrastructure designed for high performance computing with efficient checkpointing mechanisms. As target platform an IBM Parallel Sysplex Cluster is used. We consider the satisfiability checking problem for boolean formulae as an example application from the realm of Symbolic Computation. Time measurements for an implementation of this application on top of the described system environment are given.
9 citations
IBM1
TL;DR: Key technology options are discussed relative to their importance to achieving continuous availability and best practices for effectively using these technology options to improve availability are recommended.
Abstract: IBM Parallel Sysplex® is a clustering technology that was designed to address specific client business objectives for IBM mainframe System z® servers. The nondisruptive addition of scalable processing capacity and improved application availability with respect to unplanned and planned outages were two key design objectives. This paper focuses on the evolving technology options that support the business objective of continuous availability. Key technology options are discussed relative to their importance to achieving continuous availability. Best practices for effectively using these technology options to improve availability, based on extensive client experiences, are recommended.
7 citations
TL;DR: The latest generation of high-fanout and low-latency coupling adapter cards, the associated firmware, and a new protocol are described, which improve connectivity, latency, and throughput in the IBM System z Parallel Sysplex.
Abstract: The coupling adapter hub of an IBM System z® server is a key component for the IBM System z Parallel Sysplex®. The hub is built to exchange messages between systems in a highly efficient manner. This paper describes the latest generation of high-fanout and low-latency coupling adapter cards, the associated firmware, and a new protocol. As in the z10® system, there is a long-range and a short-distance card. The coupling adapter for zEnterprise® 196 (z196) is based on the z10 infrastructure (InfiniBand® link layer), with the internal transport engine for message handling completely redesigned to support the new protocol and improve connectivity, latency, and throughput. In addition to enabling the new adapter's functionality, the Parallel Sysplex support firmware has several significant enhancements in a number of functional areas. Connectivity and utilization are improved through the ability to define more channels and more concurrent connections (message buffer sets) for each channel. Through a combination of hardware and firmware protocols, response time for messages at a short distance is significantly improved. Finally, new methods are presented that support efficient presilicon and postsilicon functional and performance verification.
4 citations
Related Topics (5)
Performance Metrics
| Year | Papers |
|---|---|
| 2012 | 1 |
| 2008 | 1 |
| 2000 | 1 |