Improving Storage Subsystem Reliability - A Case Study

The importance of a business data cannot be understated. Custom-system builders protect vital data by building redundant technology into both servers and desktop workstations. The most common type of redundant technology is RAID. The term RAID is short for redundant array of independent (or inexpensive) disks. RAID is established when two or more hard disk drives are combined in a way to either increase performance, add data protection (fault tolerance), or both. However, many key components such as power supply units (PSU), Motherboards/controllers or Backplanes (B/P) in servers or workstations do not always come with a redundant design, causing a single point of failure. Single points of failure prohibit a system from functioning when a single component fails causing the complete loss of the system to the customer. These failures are disruptive due to the need for replacement or repair of failed components. High reliability is a requirement in every storage subsystem, more so with entry-level systems with nonredundant hardware or RAID redundant HDD configurations. These products pose more challenges in design and manufacturing to meet customer expectations for reliability, performance and low cost. This paper will share and discuss an example of how Promise Technology, Inc. implemented more effective reliability approaches in ESS and G8D (FRACAS) processes to address 1) a non-redundant power supply field reliability issue, 2) system box with B/P, and controller issue in a promise storage subsystem. Based on our test results we used the information to address future new product design requirements to provide enhanced reliability. Shipments of our new HBA and subsystem product field return data has shown very encouraging results over the last eighteen months.