The cloud is enabling IT to rethink its use of legacy tools and processes that, for years, have been costly and cumbersome. Storing and protecting files is a prime example. Does anyone in IT really love administering backup? Or trying to recover files from their backups? Thanks to cloud storage and a new technology called continuous file versioning, backup as we’ve known it can finally be eliminated. Continuous file versioning takes advantage of cloud object storage to protect files as they change. While the first version of a volume sent to the cloud generally covers an entire disk, subsequent versions only capture the changes, ensuring more efficient use of capacity and bandwidth.
Read on to understand how continuous file versioning works, how it is different from traditional snapshot technology, how it scales by leveraging the cloud, and why enterprises of all sizes are now using it to replace backup as their new continuous data protection solution.
How Does Continuous File Versioning Work?
Continuous file versioning starts out like traditional snapshot technology by taking an image of an entire volume. Files are chunked, deduplicated, compressed and encrypted, and the chunks are stored in the cloud object store. But this is where the technology diverges from traditional snapshots.
After that initial image, any change to a chunk is noted, time-stamped, and saved as part of each subsequent version. Performing this process of capturing changes with a snapshot is fast, but not instantaneous. Think about trying to take a large family photograph. Someone is always moving, refusing to freeze for the picture. In the same way, users are often reading, writing and making changes to disks while snapshots are in progress. To address this issue, technologies have been developed that effectively hold the image still during the snapshot. These solutions track the changes, but they can impact overall performance during the snapshot.
Continuous file versioning builds on this by breaking down the snapshot process into smaller sections of a volume using a technique called “data sharding.” In this manner, smaller groups of changes (the shards) can be captured less disruptively and processed in parallel. This enables the process of capturing versions to be continuous rather than a large single point in time snapshot of an entire volume. As a result, continuous file versioning can be done less disruptively, and at unlimited scale. Continuous file versioning also enables the ability to bring changes in from many locations sharing the same data to a central, authoritative ‘gold copy’ in the cloud.
Three Challenges with Using Traditional Snapshots for Data Protection
Let’s clarify how traditional snapshots differ from continuous file versioning. Although snapshots do allow IT to quickly roll back data to a specific point in time, there are three major problems with using traditional snapshots for data protection.
- The technology typically writes the snapshots to the same disk it copies – In one sense, this is great because the snapshots are local and easily accessible, which translates into quick rollbacks. But these snapshots could never be used for data protection because they would never survive the failure of that disk.
- Traditional snapshots use up precious capacity on the local disk, increasing costs – Primary file storage is already growing at unprecedented rates. Companies are having enough trouble finding the capacity to manage their unstructured data; they do not need to fill up expensive local disk space with snapshots.
- Snapshot solutions typically limit the number of snapshots you can retain – Each snapshot must be tracked with metadata, including timestamps and tags, so file systems put a cap on the number of snapshots that can be retained. This caps the amount of history you can keep. A viable data protection system should have no cap on history. All data should be protected and recoverable in the most efficient way possible.
There are snapshot solutions that circumvent some of these issues. Snapshot vaulting and snapshot mirroring move the snapshots to other disk-based systems, reducing the burden on primary storage.
Snapshot vaulting is better suited for standard backup, but even then there are limitations on the retention, as well as added expense to keep them on disk, which is why enterprises still have a separate backup solution for long-term storage.
“Continuous file versioning applies the best properties of snapshots – efficient capture and storage of changes and fast rollbacks.”
Snapshot mirroring is more of a disaster recovery solution. All file data is replicated – or mirrored – to another site. In the event of a local disaster, you can retrieve your history from that remote array. Furthermore, since you are dealing with an image on a disk, the restores are faster.
Does any of this sound familiar? It should because vaulting and mirroring are fundamentally not that different from running disk-to-disk backup or offsite replication. The process is more efficient, but the data is still moving from one physical medium to another. You still need replication software, WAN acceleration for extra bandwidth, physical disks, and a location to host everything, plus the budget to pay the added licensing fees. Enterprises have begun offering vaulting and mirroring to the cloud, but these solutions burden IT with another system and another vendor – the cloud storage provider – to manage.
Snapshot Efficiency at Cloud Scale with Continuous File Versioning
Continuous file versioning applies the best properties of snapshots – efficient capture and storage of changes and fast rollbacks – and merges them with the cloud to create a new kind of data protection. Instead of being forced to rely on separate tools and processes for snapshots and long-term backups, companies deploy a single, comprehensive solution that stores all files and all file version history in the cloud.
How is this possible? That’s where Nasuni comes in. We’ve invented a cloud-native file system called UniFS® that leverages cloud object stores to capture unlimited, perpetual file history using continuous file versioning. I’ve already explained the “data sharding” snapshot advancement that makes this process highly efficient, but you can find more technical detail in Nasuni patent filings.
Nasuni can be configured to capture versions in real-time, with all versions being securely stored in almost any public or private cloud object store. Nasuni can even span multiple cloud object stores. Users can self-restore files, and IT can reestablish access to business files within minutes of a disaster.
Finally, Nasuni wraps everything into a single-pane-of-glass enterprise file management solution built on top of our UniFS foundation, so you don’t have to deal with managing different tools and processes for primary NAS, backup, archive, DR, and file sync and share.
Nasuni’s file services innovation, combined with the continued maturity and growth of the cloud, is why so many enterprises are now turning to continuous file versioning as their enterprise backup alternative.
Would you like to finally eliminate your traditional backup and recovery process? What would stop you from doing this? Comment and let us know.