Home > Data Protection > PowerProtect DD Series Appliances > Dell PowerProtect Data Domain SISL Scaling Architecture > Overview
Despite tape being the dominant storage medium for data protection for decades because of its low cost, it is steadily losing ground to disk-based deduplication storage systems. The CPU-centric design of Dell PowerProtect Data Domain systems takes the pressure off disk I/O as a bottleneck. Over the last 25 years, CPUs have improved in speed by a factor of millions, meanwhile disks have improved by about 10x. It appears this performance gap will continue to grow well into the future. It is reasonable to assume that each doubling of cores could improve the Data Domain system speeds by about 50 percent.
In Stream-Informed Segment Layout (SISL), Data Domain has developed a proven architecture to deliver high-throughput deduplication storage systems with economical storage hardware. Over time, this will allow the continued scaling of CPUs to add direct benefit to system scalability.
The Deduplication approach can deliver a magnitude of data reduction compared to traditional compression. In theory, that means a deduplication system needs fewer disks. Therefore, the configured costs of a disk storage system are comparable to tape automation.
Customers are often surprised that most emerging deduplication products use a lot more disks than expected. Without careful thought about how to implement it, deduplication can become a disk-intensive activity. The conventional way to increase disk systems’ performance is to use more, faster, and more expensive disks. This can spread the load across the relatively low per-spindle access and transfer speeds. Unfortunately, using this approach in a dedupe array can quickly make it more expensive than the tape library against which it will be compared. When using low-cost, high-capacity SATA drives, much of the capacity is wasted, because each disk comes with a lot of space. By adding disks just for better I/O performance, the customer could pay for a lot of unnecessary capacity.
Dell PowerProtect Data Domain solved this problem early on with the Stream-Informed Segment Layout (SISL) scaling architecture within the Data Domain Operating System (DD OS). It optimizes deduplication throughput scalability and minimizes disk footprint by minimizing disk accesses. Doing so allows the system throughput to be CPU-centric. Speed increases directly as CPUs improve in performance.