In Greek mythology, the chimera is a fire-breathing hybrid animal with the body and head of a lion, an extra goat head, a snake tail, dragon wings — basically, a monstrosity. But a monstrosity that never existed.
Recently I read a competitor’s blog, which was in response to my earlier blog on NetApp FlexGroup volumes. I won’t link to it here, but I will say that it was chock full of mythology, like stories of the chimera. The blog presented FlexGroup volumes as some sort of solution that was hastily cobbled together. I won’t call out competitors by name. It’s poor form. But I will highlight the strengths of NetApp® FlexGroup volumes versus Competitor X’s scale-out solution.
Here are just some of the strengths of FlexGroup volumes:
- Industry-leading NetApp® Snapshot™ technology to protect your large datasets
- Disaster recovery across NetApp ONTAP® systems with SnapMirror® replication technology
- SMBv2.x, SMBv3.x, and NFSv3 support
- Industry-leading multiprotocol NAS functionality in a single OS
- No dependencies on custom hardware
- Inline deduplication (aggregate and FlexVol® level), compression, and compaction
- Ability to scale beyond a single chassis
Now, let’s debunk some myths.
Myth #1: FlexGroup volumes aren’t big, fast, and simple enough for big-data workloads
The aforementioned competitor blog spent a lot of time focusing on the limitations of the FlexVol volume as opposed to the FlexGroup volume. That’s because their narrative was “big, fast, and simple,” and they needed a way to fit that narrative. Let’s dissect this claim point by point.
“Big” isn’t 1.6PB, especially when you need to use an “assumed 3:1 data reduction” to achieve that capacity. For us, BIG, is closer to 20PiB today with a architecture that allows for something a hundred times bigger than that, and that’s before adding in efficiency technology like compression, compaction, and deduplication. “Big, fast, and simple” is the sweet spot for FlexGroup volumes.
FlexGroup is designed as a true scale-out file system that spans across multiple nodes to leverage their CPUs, memory, and network, and it can scale across 24 nodes in a single cluster. We’re scaling capacity *and/or* compute, all under a single namespace. This allows ONTAP to push millions of operations per second at predictable, low latencies, and to aggregate throughput numbers by making the most efficient use of ONTAP resources. NetApp is the only enterprise-grade NAS scale-out platform that can leverage the benefits of All Flash FAS for low-latency, high-performance, and big-data workloads, while also being able to service SAN workloads on the same storage system.
When you start dealing with unstructured NAS file systems at multipetabyte scale, you’re potentially going to see billions of files and folders. The question then is, “Do you want to manage the file system structure from the storage or from the application?” In nearly all cases, the application owners want to drive the file system structure. NetApp has created an elegant solution to a customer problem, a solution that allows customers to achieve the performance and capacity benefits of that complex architecture, but behind a simplified configuration of a FlexGroup volume. In less than a minute, you can present a 20PB NetApp FlexGroup volume that can handle millions of IOPS in a single 24node cluster.
Myth # 2: NetApp is “struggling” with C-Mode
First of all, it’s not C-Mode; it’s ONTAP. We rebranded it as ONTAP in version 9.0 because that’s how confident we are in its power and flexibility. It’s also a tip of the hat to how NetApp can still innovate on top of “legacy” technology like WAFL® and FlexVol volumes. We even have a perfect example – All Flash FAS. All we heard before the release of All Flash FAS was how “flash can only be effective if you build it from the ground up.” Flash forward (pun intended) a couple of years and NetApp is #2 in the all-flash market (with “C-Mode”) and gaining accolades, including being named the IT Brand Pulse’s “2017 Scale-out File Storage” leader. Incidentally, the same competitors that misjudged NetApp’s flash capabilities are now trying to explain our scale-out NAS solution. As our founder, Dave Hitz, once said, WAFL just happened to be built for flash. Guess what else happens to be built for flash? FlexGroup volumes.
Myth #3: Building FlexGroup volumes on FlexVol volumes? Why? They’re old!
As I mentioned with the All Flash FAS anecdote, you don’t have to build something from the ground up or start from scratch to be successful. Technology evolves as the market defines what is needed.
FlexVol volumes are old. They were released in Data ONTAP® 7.0, nearly 20 years ago. They’re almost old enough to drive! But here’s the thing: They’re also tremendously effective and stable across thousands of installations. So the question shouldn’t be “why build on top of FlexVol volumes,” but instead, “why wouldn’t you?” FlexVol volumes – and FlexGroup volumes – have been in development longer than some startup flash storage companies have been in business, much less developing their own all-flash scale-out NAS arrays “from the ground up.”
FlexGroup volumes were designed from the start for big data and high-tech workloads, with the goal of providing:
- Configurable fault domain
- Performance and capacity scaling based on workloads
- End-to-end data protection
- Ability to use different classes of media based on workload requirements
Myth #4: FlexGroup volumes are simply another layer of virtualization
The FlexVol volume was designed to virtualize the container namespace in ONTAP to allow more flexible ways to carve up physical disks. It just didn’t make sense any longer to present volumes at a physical level, because capacity and storage needs were changing. This evolution was similar to the need for servers and workstations to move away from physical hardware footprints to a single large server that virtualizes operating systems.
NetApp FlexGroup volumes, however, are not virtualized. They’re just another component in WAFL. They use FlexVol volumes as the underlying containers and are built on top of the WAFL file layout and data management engine that has 25 years of innovation and is enterprise hardened.
Myth #5: FlexGroup docs are too long!
One point was made that is fair: our “Implementation Guide” is 100 pages long. I’m guessing that they were referring to TR-4571 and/or TR-4557. That’s on the TME (me). I like to make sure that people know what they’re dealing with. But we have shorter docs for people who just want to set it up, such as Scalability and Performance Using FlexGroup Volumes Power Guide (32 pages, with complete steps for creating, managing, and protecting FlexGroup volumes). And the new FlexGroup Volumes Top Best Practices is just 4 pages. We have actual documentation sets out there for everyone to see that aren’t behind a paywall or a lead-generation form.
Myth #6: FlexGroup volumes are only good for equally distributed workloads, such as EDA
FlexGroup volumes were designed with multiple, diverse workloads in mind, including (but not limited to):
- Electronic design automation (EDA)
- Media and entertainment
- Software development
- Seismic; oil and gas
- Financial transactions
- High-tech engineering
- Manufacturing and automotive
FlexGroup volumes (as our lead developer put it) thrive on chaos. They crave as many different workloads as your clients can throw at them. The higher the rate of new files and directories, the better the FlexGroup response.
As features and functionality are added to FlexGroup volumes, expect this list of ideal workloads to grow. Don’t believe it? Email us at firstname.lastname@example.org to find out more, and to learn how to get a proof of concept started. Or try it out yourself! The ONTAP flexibility and multitenant architecture allow you to stand up storage virtual machines independent of production to conduct tests of new features.
Myth #7: Our competitors know our technology better than we do
Any time you read a competitor’s assessment of NetApp, be sure to consider the source. We don’t pretend to be experts on our competitors’ solutions. If you’re looking for a detailed, accurate, and factual representation of NetApp technology, ask NetApp. We’ll be glad to have a conversation about FlexGroup volumes, ONTAP, or any part of our portfolio. We can give you the information so you can make decisions based on fact and accuracy. For more information about FlexGroup volumes, email us at email@example.com or tweet me at @NFSDudeAbides.
Find out more about NetApp FlexGroup volumes
NetApp FlexGroup volumes are available in ONTAP 9.1 and later. For technical reports, best practices, and information on what is and is not supported, check out:
- TR-4571: FlexGroup Best Practices and Implementation Guide
- TR-4557: NetApp FlexGroup Volume Technical Overview
- Watch NetApp Insight 2016: NetApp FlexGroup Volumes
- Watch Deploying and Managing FlexGroups using OnCommand Management Software
If you’re interested in a deeper technical look on FlexGroups, check out episode 46 of the Tech ONTAP Podcast.
Last, if you spot any chimera, please send a photo my way.