Systems Performance

Brendan Gregg (San Francisco, CA area) is Lead Performance Engineer at Joyent, where he analyzes performance and scalability for small to large cloud computing environments, at any level of the software stack. He is primary author of DTrace and co-author of Solaris Performance and Tools (both from Prentice Hall); as well as numerous articles about systems performance. Previously performance lead and kernel engineer at Sun Microsystems, he has also consulted and trained IT professionals on performance optimization. He developed DTraceToolkit and ZFS L2ARC, and many of his DTrace scripts are shipped by default in Mac OS X and Oracle Solaris 11. His recent work has included performance visualizations.

出版者:Prentice Hall

作者:Brendan Gregg

页数:792

出版时间:2013-10-26

价格:USD 64.99

装帧:Paperback

isbn号码:9780133390094

图书标签:Linuxperformance性能计算机system性能优化操作系统Performance

The accelerating deployment of large-scale web, cloud, Big Data, and virtualized computing systems has introduced serious new challenges in performance optimization. Until now, however, little reliable, practical information has been available to IT professionals who are responsible for running these systems efficiently and cost-effectively.

Systems Performance: Enterprise and the Cloud is the solution. Internationally renowned performance optimization expert Brendan Gregg brings together state-of-the-art techniques and tools for analysis and tuning of large-scale web/cloud computing environments.

Gregg focuses on Linux/Unix/Solaris performance issues, while offering proven methodologies and discussing key issues that apply to all enterprise operating systems. Coverage includes:

Modern performance analysis and capacity planning, including key issues such as latency and dynamic tracing

New performance and reliability challenges associated with cloud computing

Methodology, concepts, terminology, tools, and metrics

Key tradeoffs, including problems of load vs. architecture

Tuning operating systems, CPUs, memory, file systems, disks, networks, and busses

Tuning virtualized systems

Programming language issues related to performance — including application profiling for C, C++, Java, and node.js

Benchmarking strategies and pitfalls, including custom microbenchmarking

读后感

《性能之巅：洞悉系统、企业与云计算》基于Linux 和Solaris系统阐述了适用于所有系统的性能理论和方法，Brendan Gregg 将业界普遍承认的性能方法、工具和指标收集于本书之中。阅读本书，你能洞悉系统运作的方式，学习到分析和提高系统与应用程序性能的方法，这些性能方法同样适...

相对来说，读起来累是比较次要的。 读到第 83 页读不下去了，82/83 页有一段是这样的： ...... 而是只添加一个系统调用：perf_event_open()，它会用 ioctl() 返回一个文件描述符。...... 在这种情况下，开发人员可以很容易地对参数 PERF_EVENT_IOC_ENABLE 做添加和修改。 找了...  

用户评价

太宽了就显得范了。

700多页我竟然读完了，感觉讲得太罗嗦了。。。

大师之作，不仅介绍了性能分析和排障的各种工具，更是总结了性能分析和排障的方法论，帮助我们不仅看到性能问题，更为我们指了一条追根溯源的路。

非常实用，介绍到很多工具，目前最好的性能调优工具书。

参考书