Storage System Bottlenecks And Their Solutions

In today's demanding and competitive business world, it is critical for Information Technology (IT) departments to obtain optimum end-to-end performance from their companies' computing solutions (with end-to-end defined as involving the servers, the storage, and the applications in between). To achieve optimal performance, it is essential to configure each computing resource in the end-to-end solution to perform optimally to meet application workload needs and specific business performance requirements. To accomplish this, however, they must first understand one of the most critical pieces of the end-to-end computing solution puzzle: the storage system.

There are many storage vendors today, each offering a wide range of storage system options. This includes Data Center storage systems, Mid-Range storage, Workgroup Storage, NAS and Backup, (tape and software.) Though there are many choices available, this paper focuses on disk-based storage systems.

As noted earlier, storage system performance is a crucial piece of the IT solution, for it typically has a significant impact on the overall performance. Consequently, the specific purpose of this paper is to educate the reader concerning typical storage system performance criteria (e.g., input and output (I/O) rate, response time, and data rate) determined by the storage system workload. Additionally, it will discuss how to resolve storage system bottlenecks. A bottleneck is any limiting factor or component in the storage system which negatively impacts its performance; i.e., a storage system will only perform as fast as its slowest component.

Diminishing storage system bottlenecks should help improve storage system performance, which can assist in meeting, and hopefully exceeding, overall business performance requirements.