For many years, storage and storage interfaces could have been viewed as the Rodney Dangerfield of the IT infrastrure. It “didn’t get no respect.” Long considered only as a cost — measured by dollar per-megabyte ($/MB) and, more recently, as dollar per-gigabyte ($/GB) — storage has only recently (coinciding with the growth of e-commerce) been seen as cool and trendy.
For those of us who have spent a career working with storage and I/O technologies, it’s refreshing to see storage and storage interfaces open up. What has been considered a boring, back-of-the-house necessity is no longer seen only as a cost item that needs to fit the budget. Storage has finally found its role and importance as an IT enabler.
Why the sudden popularity of storage? It’s partly due to emerging technology that has enabled content-rich media applications that require massive amounts of storage capacity and I/O bandwidth. As such, storage has grown to occupy a larger portion of many IT budgets in recent years. Another factor is that people have become aware of the need for storage at home and at work. This is similar to what occurred in the networking world in the late ’80s and early ’90s.
Some activities that have resulted in an increase in storage and I/O growth include:
- Workload, server, storage, and data center consolidation
- e-Commerce and content-rich applications that employ images, video, and voice data
- Increased interest in business continuity and data protection, resulting in more copies of data
- Technology and storage upgrades
- Continued decreases in storage costs.
Implementation of enterprise storage for zSeries and other large systems continues, but not at the sensational rates seen prior to Year/2000 (Y2K). There are indications that an increase in storage spending and implementation is starting to occur as the year progresses and the $/GB cost of storage continues to drop. Some roadblocks do exist, however, to slow down and hinder storage expansion and use. Current needs include:
- Reducing the cost to manage storage through system automation and human efficiency
- Increasing the amount of storage that can be managed per person
- Enhancing technology to relieve physical and logical address constraints
- Improving performance over longer distances to protect data and business resiliency
- Improving storage resource utilization and sharing across platforms
- Further shrinking the time required for data backups and recovery.
Accessing and Using Storage
Storage can be accessed using dedicated Direct Attached Storage (DAS) and shared Storage Area Network (SAN) attachment.
The traditional method for accessing storage has been DAS, and it’s still common on stand-alone systems, small servers, and PCs. DAS involves attaching a server to a storage device using an I/O bus (channel or storage interface).
The storage network model utilizes a storage interface such as Fiber Channel to attach two or more systems to shared storage resources (disk and tape). Those who have implemented Enterprise Systems Connection (ESCON) on an S/390 or zSeries have implemented a storage network. The storage networking police (if they, in fact, existed) might argue that ESCON isn’t a SAN interface and that a SAN only exists with Fiber Channel. While the IP folks might disagree with the Fiber Channel assertion, it’s safe to say an ESCON environment is a form of storage networking, although it’s proprietary.
Storage networking enables multiple servers, regardless of operating system type, to share storage-to-support server consolidation, improve backup, and enhance disaster recovery with increased distance compared to DAS over an open interface.