Last month, the United States Patent and Trademark Office issued U.S. Patent No. 8,566,362, titled Method and system for versioned file system using structured data representations. As the title implies, this invention purports to provide the ability to create a versioned file system from the object-based storage infrastructure of public cloud storage providers.
When deployed at scale, object storage systems are cost-effective, stable and highly available; therefore, they are well suited to the needs of the large Web companies that operate them. Unfortunately, until now, object storage was not well adapted to handle changes quickly or in a consistent manner, which makes the technology unsuitable for data center infrastructure in its raw form. Thus, it was necessary to bridge the gap by leveraging local snapshots to create a direct mapping between a high-performance file system and a series of immutable versions of the file system that are then committed to the object store.
According to Andres Rodriguez, CEO of Nasuni, which is the owner of the ‘362 patent, the goal when the company set out to develop a versioned cloud based object-based storage system was to “consolidate all of the storage functions into a simple service offering… The era of glorified hardware boxes is coming to an end. It is too expensive to keep up with the sheer growth in data and the need to have that data protected and available everywhere. Enlightened IT organizations are thinning their data centers and leading the charge towards a new era where simple appliances offer direct access to the vast resources that are available in the cloud.”
The way the invention works is by creating a snapshot backup. For example, first the interface creates and exports to the data store a first structured data representation corresponding to a first version of the local file system. Upon a change within the file system (e.g., file or directory creation, deletion, or modification), the interface creates and exports a second structured data representation corresponding to a second version of the file system.
Thus, the second structured data representation differs from the first structured data representation in one or more (but not necessarily all) respects when compared with the first structured data representation. Thus, the interface generates and exports structured data representations to the data store when changes within the file system have occurred. This new approach to cloud back-up combines the performance and consistency of a traditional file system with the scale and stability of the object storage systems used by all of the leading public cloud providers.
Unlike other cloud gateway technologies, this technology has no limits on the number of files or snapshots that can be supported in a single logical volume. Regardless of the number of snapshots, there is no performance degradation, and restoring data from any time in the past can take place without the need for a separate backup or archiving application.
The broadest claim in the patent, claim 1, covers:
A non-transitory computer-readable medium having stored thereon instructions that, when executed by a processor, perform a method associated with a local file system, the method comprising: at a first time, creating and exporting to a cloud-based storage a first structured data representation corresponding to a first version of the local file system, wherein the first structured data representation corresponds to a complete copy of the first version of the local file system at a point-in-time and comprises, for at least one file of the local file system, a tree having a root element that includes all the data from one or more inode or inode-equivalent data structures of the local file system describing the file, and at least one object that contains at least a portion of the file, and one or more directory elements associated with the root element; and upon a change within the local file system, creating and exporting to the cloud-based storage a second structured data representation corresponding to a complete copy of the second version of the local file system at a point-in-time, the second structured data representation differing from the first structured data representation up to and including the root element of the second structured data representation; wherein at least one of the first and second structured data representations is self-contained in that it includes or points to all data structures and data needed to reconstruct the local file system at a point-in-time; wherein the cloud-based storage is a write-once data store.
Tags: cloud computing, patent, patents
You share in the PLI Practice Center community, so we just ask that you keep things civil. Leave out the personal attacks. Do not use profanity, ethnic or racial slurs, or take shots at anyone's sexual orientation or religion. If you can't be nice, we reserve the right to remove your material and ban users who violate our Terms of Service.