NetApp Doubles Down On Cloud, Data Fabric With Ontap 9.5

NetApp Wednesday used the second day of its NetApp Insight 2018 conference to expand its Data Fabric platform from the core to the edge and the cloud with the introduction of a new and more powerful Ontap storage operating system.

The Sunnyvale, Calif.-based vendor also formally introduced new software that uses in-server persistent memory to provide extremely low latency to applications, and unveiled a unique take on performance guarantees that focuses on low latency.

It is refreshing to see how much of a singular focus NetApp has had on the cloud and its Data Fabric architecture, said Mark Vaughn, senior director of integrated services at Presidio, a Greenbelt, Md.-based solution provider and longtime NetApp channel partner.

[Related: NetApp Insight: NetApp Significantly Expands Ability To Treat On-Premises, Cloud As One]

"The cloud is an area where NetApp is in the lead," Vaughn told CRN. "The company was quiet about it at first. But their ability to create hybrid solutions is second to none. Data Fabric is real."

NetApp has really moved the conversation from on-premises infrastructure to a focus on how to integrate the cloud into every aspect of customers’ business, said Bob Hankins, vice president of technical sales at Logicalis, a New York-based solution provider and NetApp channel partner.

"We used to talk about replacing five-year-old arrays," Hankins told CRN. "Now we're talking about changing the infrastructure. It's a whole different level of conversation, especially around Data Fabric. NetApp is really opening customers’ eyes around the cloud."

All of NetApp's new hardware and software introductions Wednesday are tied together by its Data Fabric architecture, said Bharat Badrinath, vice president of storage systems and software marketing at NetApp.

Data Fabric is NetApp's platform for allowing data to be managed and migrated seamlessly across on-premises, public clouds, private clouds and hybrid clouds.

NetApp is pushing forward on its Data Fabric platform with a new version 9.5 of its Ontap storage operating system, Badrinath said.

Ontap 9.5 features increased performance and efficiency thanks to a reworking of its code base, Badrinath told CRN. But those increases also stem from several other enhanced products and services, including for the first time support of NetApp's FlexCache technology for speeding up access of data from remote sites across clusters, he said.

Another big enhancement to Ontap 9.5 is the ability to manage multiple NetApp Fabric Pools, which are a part of automated data tiering, as a single Fabric Pool stretching from on-premises to the cloud, Badrinath said.

Ontap 9.5 also now includes NetApp Data Availability Services, which allows backups of data from primary to secondary storage, and secondary storage to the cloud, with two clicks, Badrinath said.

"This turns an existing backup into one that is managed by the cloud, with a copy in AWS S3," he said. "From S3, it can then be moved to NetApp Cloud Volumes."

NetApp, which has made a big investment in the high-performance NVMe storage protocol, is now the first to support Red Hat Linux on NVMe thanks to Ontap 9.5, Badrinath said.

Ontap 9.5 also allows for metro clusters to span to over 700 kilometers versus the previous limit of 300 kilometers, and provides zero data loss when replicating data with NetApp SnapMirror, he said.

FlexCache was available in the past, but NetApp is bringing it back with Ontap 9.5 in a more usable fashion, said Glenn Dekhayser, field chief technology officer at Red8, a Costa Mesa, Calif.-based solution provider and longtime NetApp channel partner.

FlexCache in the past was more about content distribution, but not for read and write use across an environment, Dekhayser told CRN.

"Now, a lot of engineering and legal and other customers are very dispersed geographically," he said. "FlexCache is going to have a huge uptake in those industries with distributed data requirements where they used WAN optimization in the past."

With Ontap 9.5, NetApp also officially unveiled MAX Data, or NetApp Memory Accelerated Data. MAX Data, which NetApp first started discussing in July, is based on technology NetApp got with its 2017 acquisition of PlexiStor. It is aimed at helping channel partners provide low-latency acceleration services for persistent memory while still providing enterprise-class data management.

MAX Data is server-side software that lets partners use persistent memory in servers such as Intel's Optane 3D XPoint memory technology that offers performance approaching that of DRAM but retains memory if power is lost.

With MAX Data and Optane, customers can experience performance improvements of five times to 100 times, Badrinath said.

"We are improving the latency by orders of magnitude without needing to rewrite the applications," he said.

MAX Data currently works in conjunction with NetApp storage as part of the Data Fabric, but in the future will also be able to accelerate storage from competing vendors, he said.

Back in the old days, it was possible to carve a RAM drive out of server memory, but it wasn't useful because it didn't have persistent memory, so when the server power was off, the data was lost, Dekhayser said.

"NetApp, using the PlexiStor technology, has been able to create a LUN in RAM and making it persistent by backing it up with AFF [All Flash FAS] and NVMe over Fabric," he said. "That's pretty hot stuff because, by providing a LUN that's accessed in RAM, it's the end-stage of storage. The only way to get a LUN with faster performance is by getting the data into the CPU Level 1 and Level 2 cache, and that's never going to be a thing because there's not enough memory in there to do it."

Think of the possibilities, Dekhayser said.

"You'll be able to take legacy database applications like Oracle or MongoDB, and put them in this layer," he said. "You start to see the possibility of putting new data types in this tier, which will create new opportunities for application developers who before had to deal either with slow I/O or lack of persistence. You don't have to make a choice anymore. You can have your fast I/O and your persistence. And not only that, but you also have all the Data Fabric management that NetApp brings to the table."

MAX Data will be a hit with workloads where performance is an absolute necessity including genomics or energy where new technologies like artificial intelligence are pushing the need to ingest data ever faster, Presidio's Vaughn said.

"Latency is important, especially in data analytics," he said. "We're looking at high-speed ingesting of data. And as we do more work with AI, analytics and IoT, the number of devices pushing data back to the core and the amount of data they are pushing is growing rapidly. It's important to keep up with that in real time."

NetApp is also approaching low latency with what Badrinath called a unique performance guarantee. For NetApp's flagship A800 all-flash array, NetApp will guarantee that latency will be 500 microseconds or less 99 percent of the time as measured in 100-hour increments. The latency guarantee for other all-flash arrays will be 1 millisecond.

"We're providing peace of mind. … If we don't meet [the guarantee] for 100 hours, we will immediately remediate that," he said.

The performance guarantee moves the conversation away from needing to buy controllers over and over forever, Dekhayser said.

"While some vendors focus on the six- to eight-year life span of storage, NetApp with the flash performance guarantee is earning their keep 100 hours at a time," he said.