News Stay informed about the latest enterprise technology news and product updates.

NetApp launches Flash Pools for SSD, HDD RAID groups, FAS2220 NAS

With its FAS2220, NetApp adds Flash Pools that combine solid-state drives and hard drives to accelerate read/write performances.

NetApp today launched the FAS2220, a new entry-level storage system, and Flash Pools, a caching method for RAID groups which combine solid-state drives (SSDs) and hard drives.

The new 2U FAS2220 scales to 180 TB, and will replace the FAS2020 array as NetApp’s entry-level system. It supports NAS and iSCSI. The FAS2220 will compete with EMC VNXe unified storage and Dell EqualLogic and Hewlett-Packard P4000 iSCSI SANs. Pricing starts at $8,000 for a 6 TB single-controller unit. The system supports SAS, SATA and single-level cell (SLC) solid-state drives. The FAS2220 replaces the FAS2020 array as NetApp’s entry-level system.

The FAS2220 supports Data OnTap 8.1.1, which enables support of Flash Pools. Flash Pools mix SSD and hard drives to accelerate reads and writes. They let customers combine SSDs and hard drives into an aggregate of RAID groups. Hot data in the aggregate gets copied onto flash for faster performance. Flash Pools can include 300 GB per volume of SSDs.

The company's Flash Pools technology is different than NetApp’s Flash Cache product, which has shipped with its enterprise arrays since 2009. Flash Cache is a PCIe card inside the storage array that handles read-intensive workloads. Flash Cache is better suited for multiple storage pools, while Flash Pools technology is intended for a single storage pool. Flash Cache is not available on the FAS2200 series.

Although NetApp calls Flash Cache and Flash Pools virtual storage tiers, they are really caches rather than tiers. NetApp approaches its flash strategy different than other storage vendors. Most of its competitors use automatic tiering software to move data from flash to disk tiers, but NetApp caches hot data on its arrays.

Jim Sangster, senior director of solutions marketing for NetApp, said the vendor’s flash strategy does not require customers to set policies and make choices on data movement.

“The end user doesn’t have to actively manage,” Sangster said. “Data placement is automated based on hot data.”

Flash Pools will eventually become available on the FAS2240, which was launched in November, and NetApp’s midrange FAS3200 and enterprise FAS6200 platforms. Flash Pools will give NetApp FAS6200 and FAS3200 customers the option to use Flash Cache for read I/O acceleration, while Flash Pools can be used for read and write acceleration. Flash Cache is not available on the FAS2200 series.

Enterprise Strategy Group senior analyst Terri McClure said the major differences between the two products are that Flash Pools supports read and write capabilities - while Flash Cache is read-only - and Flash Pools can target specific workloads.

“With a Flash Pool, I can give a workload its own aggregate and not have that aggregate disrupted by other workloads that server might be dealing with,” she said. “I can pin workloads to pools with a higher performance level.”

As for NetApp’s caching strategy versus automated tiering, she said, “there are broad ways that people implement automated and policy-based tiering. Some have 200 policies that you can set for one volume. NetApp’s philosophy is you don’t need to put that kind of work into it. Let the system figure it out and it will put hot data into flash. You don’t have to worry about setting policies and changing them over time.”

NetApp has future flash plans. In recent earnings calls, CEO Tom Georgens spoke about deploying flash throughout the hardware stack and that NetApp’s flash products will be “multi-faceted.” He also said that the most effective way to use flash is as cache, so it’s unlikely that the vendor’s primary strategy will change.

Dig Deeper on Solid-state storage

Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

Is caching writes to flash a feasible (and cost effective) option? How would it compare to creating a NVRAM write cache that becomes a part of the storage stack?
And how does the sizing work?
This is NetApp adding Flash to a 20 year old architecture and burying it's true value under a RAID management stack. It is added in and not designed in. A real SSD caching solution in a hybrid storage system will not put RAID group restrictions on SSD and force you to manage RAID groups to determe where to place SSD. it will enable the pooling of SSD resources across all volumes to provide performance as a Utility. See my blog here for how next generation storage handles SSD caching efficiently