Seven New NFS Capabilities

If you’re looking for power from your file system, NFS v4.2 delivers a host of new features that enhance modern scale-out architectures.

Lance L. Smith is CEO of Primary Data.

What many of us in IT love about open standards is that they can be improved upon and leveraged by all. NFS stands out as an archetypical example. It enjoys widespread use, and continues to evolve thanks to the work of a dedicated community of innovators.

The latest release, NFS v4.2, solves a number of v4 performance issues and also introduces many new features developed to meet the demands of data-driven businesses. Let's explore how these breakthroughs can power new capabilities with the file system you may already be using at your enterprise.

Ask most admins about NFS, and high performance isn’t often the first thing that comes to mind. NFS v3 performance was adequate, but not ideal for high-performance applications.  Protecting data integrity required additional steps, as NFS v3 is a stateless file system. Latency was a problem because data access requests typically required 5-6 trips between the client and the NFS server. Since NFS v3 could not cache data on clients, all access requests required these multiple steps every time data had to traverse the network.

NFS v4 worked toward a fix for this problem by evolving to a stateful protocol, which enabled data caching on clients with the delegations feature. But this surface fix ended up creating new problems below since the rest of the NFS v4 architecture wasn’t updated to leverage this change. Trips between the client and the NFS server actually increased from 5-6 to around 10. As a result, in spite of the ability to now cache data, this overly chatty design worsened performance and scalability from NFS v3 to NFS v4.

NFS v4.2 finally sets things right with its Compound Operations feature. Compound Operations reduce the typical number of round trips between client and server to about two, significantly reducing latency. In fact, there is no need to check with the server at all when an application is accessing data cached on the client. In this common operation, data is accessed directly, so performance is limited only by what the storage and the network can deliver.

The Power of pNFS

NFS v4.2 enables clients to access multiple storage devices in parallel with a feature called parallel NFS, or pNFS. This feature makes it possible to finally make multiple storage resources simultaneously available to applications so workloads can be distributed across them. While first introduced in NFS 4.1, the performance improvements created by these features in 4.2 are significant, and make the NFS v4.2 release an excellent choice for high-performance applications.

Seeing the results of these performance improvements requires real-world testing. In fact, it’s hard to see the change if using tests performing simple file operations, such as opening or creating a thousand files, because these kinds of tests don’t exercise capabilities beyond  stateless NFS v3 architectures. Using tests that measure real-world operations, such as how fast a file can be created and then accessed by another application, enterprises can achieve the kinds of results that make the most impact on business.

Flex Files: The Keys to the Automation Castle 

It is difficult to overstate the impact of the Flex Files feature in NFS 4.2. This powerful feature enables live files to be moved without impacting applications. This has never been done before, and it changes everything.

Flex Files can non-disruptively recall layouts (metadata about files) to allow data access and data integrity to be maintained even as files are being copied. This means you no longer need to schedule downtime to perform migrations and upgrades – they can happen even while data is being accessed, without causing any application interruption. But these abilities need to be enabled. Metadata engine software makes it possible to automate the movement and placement of data according to IT objectives.

The non-disruptive movement that Flex Files makes possible also helps to reduce downtime. With human error causing up to 60 to 80 percent of all downtime events, if you are using software that leverages this feature, get ready for much more uptime at your enterprise.

Data is Unaware No More: Free and Accurate Performance Telemetry

With NFS 4.2, all Linux clients continuously report back with performance metrics about the underlying infrastructure. These metrics can be used to optimize service levels while minimizing costs, and can be used to profile your data and savings.  Importantly, NFS clients do not require additional software installation to add this ability. When it’s this easy to see what’s really going on with your infrastructure, as Maya Angelou said, we are only as blind as we choose to be.

Native Support for file Cloning

Performance loves to go native. Turns out, protection does too. Server-side clone-and-copy enables cloning and snapshots of files by any NFS v4.2 storage server. If the NFS storage server is deployed on an NVMe powered flash server, admins can improve service levels by offloading these operations from storage—preserving more storage resources for serving data to applications.

Open Your Windows to Enhanced Security

Let’s be friends – NFS 4.2 ACLs are compatible with Windows ACLs, making is simple to share data securely across Linux and Windows platforms. In addition, the ability to use RPCSEC_GSS for authentication and data access delivers another security boost.

 All Around Support from Enterprise Linux Distributions

Not only is NFS v4.2 supported by all the major distributions, including Red Hat Enterprise Linux, CentOS, Oracle, Ubuntu, and others, but also, Red Hat recently abandoned BTRFS support in favor of NFS and its parallel access capabilities noted above.

If you’re looking for power from your file system, NFS v4.2 delivers a host of new features that enhance modern scale-out architectures. To learn more about how to put them into play at your enterprise, check out the Linux NFS wiki for a repository and documentation. Most Linux vendors’ documentation also serve as great sources for more detailed information. 

Opinions expressed in the article above do not necessarily reflect the opinions of Data Center Knowledge and Informa.

Industry Perspectives is a content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating. View previously published Industry Perspectives in our Knowledge Library.

Hide comments

Comments

  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Publish