Cloud gateway vendor Panzura recently posted a blog comparing their snapshot methodology vs. the technical approach used by Nasuni enterprise file services. While they got some of it right, there were some blatant falsehoods. One that stood out was their claim Nasuni is limited to 256 snapshots per volume. Nasuni is known for our global file system innovation, which supports unlimited version history. Many of our case studies feature Nasuni customers who have been storing versions of files as often as every 15 minutes for years, resulting in hundreds of thousands of versions of a single volume. It shouldn’t have been hard to fact-check that erroneous claim.
Let’s examine each of Panzura’s claims in detail and separate the truth from the vendor “snap talk.”
Panzura: “Copy on Write incurs higher performance overhead during the snapshot process than Copy on Redirect.”
Reality: True. However, this is only relevant during the actual time a snapshot is being taken – typically a few seconds every 15 minutes, every hour, or every day, depending on your versioning frequency. The overall impact on operations ends up being negligible. As with most technologies, there is also a tradeoff. Copy on Redirect’s use of pointers ends up with a fragmented data structure that does not scale over time as versions add up. This may be why Panzura claims a theoretical limit of 264 snapshots, but only tests to 10,000.
Panzura: “Because Nasuni’s snapshot technology is based on the Linux Volume Manager (LVM), they are in fact limited to 256 snapshots per volume.”
Reality: False. Nasuni doesn’t have any reliance on LVM’s snapshot limit. Nasuni’s key innovation is our global file system, Nasuni UniFS®. As the world’s first file system designed to live in the cloud (i.e. the entire inode structure resides on cloud object storage instead of hardware), UniFS has no limit to the number of snapshots that can be stored. UniFS scales as well as Azure, AWS, IBM COS, and other underlying cloud object stores. We understand vendors that depend on legacy, device-constrained file systems like Panzura might naturally assume others have similar limitations. But that’s why we built UniFS as a cloud-native file system. If you require a 15-minute Recovery Point Objective (RPO), you’ll exceed 10,000 versions in 3 ½ months. Walsh Construction has been using Nasuni for years with a 15-minute RPO. They now have hundreds of thousands of snapshots of a single volume stored with Nasuni.
Panzura: “Nasuni says they have tested up to 27,000 snapshots. That’s a lot of volumes.”
Reality: False. Panzura is likely referring to our public scalability demonstration of a single Nasuni volume storing 1 billion files. Their erroneous claim that Nasuni is limited to 256 snapshots per volume led them to incorrectly assume our 27,000 snapshots were spread across multiple volumes. As explained above, Nasuni has no limit to the number of snapshots that can be stored per volume. The 27,000 snapshots in the demonstration were indeed on a single volume. And we only stopped there because that’s when the writing of the 1 billion files completed. We finished that demonstration nearly three years ago, but the actual testing continued well past that point.
Panzura: “The difference is that one solution is designed for the modern data center and does not impact performance, while the other is based on legacy technology that is known to cause performance problems.”
Reality. True. The Nasuni platform, built on the world’s first cloud-native file system, is the only enterprise file services solution designed for modern data centers that need multi-site scalability, local NAS-like performance, advanced RPOs and RTOs, and global file access. With our unique cloud-centric architecture, customers can scale to any number of sites without performance degradation, since file system metadata must only be updated once in the cloud. This is in stark contrast to device-centric architectures that suffer performance degradation as more sites are added, due to the need to update all metadata caches every time a write is made in one site.
The bottom line: enterprises with multiple locations and rapid unstructured data growth are consistently selecting Nasuni cloud-scale enterprise file services over Panzura cloud gateways for our:
- patented global file system that scales in (and with) cloud object storage to provide limitless primary NAS and archive capacity
- global namespace that provides multi-site file access without version conflict
- unlimited version history that enables fast recoveries from file deletions, corruptions, and malware/Ransomware events
- stateless edge appliance architecture that enable any location with power and cloud connectivity to become a DR site in as little as 15 minutes.
Enough of the vendor “snap talk.” Have you looked at Nasuni and Panzura in depth? Weigh in with your thoughts below in the comments.