As Ajay described in his blog, we spent over a year talking to more than 100 IT professionals to better understand their challenges, and their storage and application environments, while we developed our first products. We also surveyed a nationwide group of 150 professionals across a wide range of company sizes and industry verticals to get a broad perspective on their challenges in managing storage, backup and disaster recovery. We announced the results of that survey yesterday. The complete survey data can be found here. The survey corroborated much of what we learned in individual discussions: the cost and complexity of storage, backup and DR are pervasive challenges.
High RPM = $$$
For example, 66% of respondents told us that a lack of available budget for new systems or upgrades is limiting their organizations’ storage systems. Application data continues to grow at a rapid clip, and most business-critical applications require high performance primary storage that is usually provided by arrays with expensive, high RPM drives. Flash holds a lot of promise to improve performance, but existing options come at a cost that remains hard to justify for most mainstream enterprises. Note that I’m not referring to those of you who are short-stroking racks of 15K RPM disk shelves to keep your seismic analysis or animation render farm humming – high-end flash solutions can certainly both improve performance and reduce costs for you today. Emerging tiered architectures that can automatically move data between flash and HDDs are certainly a step in the right direction, but as we’ll explore in more depth in coming blogs, we think they are insufficient to significantly reduce primary storage costs for mainstream applications.
Disk Backup Systems are Expensive
Another big driver for increasing equipment cost as application data grows comes from the need to continually upgrade or add new disk-based backup systems. Here the problem is not the expense of the drives themselves. Because backup workloads are sequential, high-capacity, low-RPM drives work well, and dedupe, by eliminating redundancies across backups, makes storing 60 to 90 days of backups relatively affordable. But even with these optimizations, the capital cost of disk-based backup systems usually exceeds tape libraries. Despite the up-front cost, we’ve seen widespread adoption of disk-based backup thanks to the significant operational advantages over tape.
These insights were instrumental in the development of our CASL™ architecture. CASL allows our customers to converge primary and backup storage, eliminating expensive, high RPM drives in primary storage and the entire standalone disk-based back system. As a result, we’ve been able to deliver savings of 60-75% in upfront capital costs, along with big ongoing operational benefits.