Q&A with Paul Feresten of NetApp

Paul Feresten (Profile)
Tuesday, October 9th 2012

VSM: A lot has been happening in the world of Flash on both the server and storage side. What is NetApp’s overall strategy with regard to the use of Flash technology?

PF: It has been quite an exciting time in the world of enterprise Flash technology. NetApp views Flash as a game-changing technology and is pursuing an extensive innovation agenda to address evolving customer needs. With our Data ONTAP platform, NetApp today is providing customers who have mission critical and performance sensitive applications with the combination of seamless scale-out, non-disruptive operations, and Flash technology required to accelerate business.

NetApp currently offers both PCI-e based Flash technology (storage controller attached) and SSD technology (storage array attached).  Flash technology is primarily used to maximize the effectiveness of our intelligent caching thus enabling the NetApp Virtual Storage Tier (VST). NetApp’s VST allows customers to speed access to critical data while lowering the cost of physical storage infrastructure. In addition to the storage layer, deployment of Flash and intelligent caching at the server layer provides additional performance enhancement at the application level while maintaining consistency with back end storage.

VSM: Where do you think that the use of Flash makes the most sense? Server, network, storage array, or all of the above?

PF: Flash technology yields benefit at both the server and back-end levels. For lowest latency at the application level, placement of the “hottest” data sets at the server will speed performance and improve server efficiency. For data shared across multiple servers and data centers, it makes sense to apply Flash and caching at the storage back end. A key requirement will be integration of the end-to-end deployment, as data consistency must be maintained.

VSM: Which applications and workloads benefit most from the use of Flash technology?

PF: We believe that to determine the optimal deployment of Flash technology, it’s best to map it to workload needs. There are three types of customer applications that are ideal for Flash technology: shared virtual infrastructures, large capacity and high IOP transactional workloads.

Shared virtual infrastructures represent the largest segment of the storage market today and covers more than 50% of the total storage opportunity. Most of the IT workloads in virtualized environments run on hypervisors such as VMware, XenServer, or Microsoft Hyper-V™. These customers require a system that provides the ideal balance of performance, capacity, and cost in a multi-tenant shared infrastructure. In order to achieve this balance, a hybrid system of Flash and HDDs is ideal to automatically manage the appropriate location of data based on workload need to the storage media.

Customers in the large capacity segment can best leverage Flash technology within the array for read caching, which is a cost-effective way to provide a performance boost when infrequently used data is suddenly in great demand.

Flash technology also provides incredible value for high IOP transactional workloads due to its low latency and high IOPS. At the same time, customers in this segment require high availability, resiliency, extreme quality, and reliability.