About The I/O Storm Blog
Application response time continues to plague businesses of any size, with more than 80% of the delay being a result of disk I/O bottlenecks. CPUs sit underutilized while waiting for storage I/O to finish. Users also sit idle, in growing frustration, as they wait for data to be delivered. Storage technology has evolved to support the growing need for capacity and performance, but “I/O wait” is still the dominate wait events in many critical application.
This blog discusses what I/O wait is, how it can be measured, ways to overcome it, specifically to improve application performance. A major emphasis will be given to the effect of storage performance on the application behavior and end-user experience and satisfaction. In addition, many of the topics will focus on RDBMS applications, mainly Oracle and SQL Server.
Topics which will be discussed in this blog include, but are not limited to:
- What is acceptable I/O latency?
- What is “I/O wait”?
- When are latency, throughput and IOPS important?
- How to monitor I/O wait
- SQL Server and Oracle I/O bottleneck tuning
- Achieving high performance from Oracle/SQL on fast storage
- SSD market trends
- Simulating application performance improvement of storage refresh
- The affect of mixed workload on the I/O performance
- Achieving data redundancy without compromising performance
Your Entries and Comments are Welcome
This is an open forum and we invite you to contribute your thoughts, ideas, and experiences in any of these, or related areas of storage I/O and application performance. You can submit your comments directly to the blog via “Comment” or email them to Eyal Markovich, our blog moderator, at TheIOStorm@kaminario.com.
