In this short video, Andres Rodriguez, Nasuni Founder and CTO takes to the whiteboard to answer the question: What is a Global File System?
Today we want to answer the question “What is a global file system?” What can we use it for?
So let’s start with the makeup of it. At the core of a global file system you have cloud storage, also some people call it object storage, it essentially is a system for replicating any asset that you put into it to many servers and to many geographic locations across data centers. That’s the foundational hub for a global file system. Within that, you are going to be running your global file system – and in the case of Nasuni that file system is UniFS®. A file system must first of all be able to store files and then every change to those files must be stored as a stable immutable version of that file.
In a global file system you typically have an infinite time-stream of those files as they continue to accumulate in the system and so if you rename a file, if you change a file, if you create a new directory structure, it’s all being represented in this version stream. One of the great advantages of that back-end is that it’s distributed it’s fully version and you can attach now multiple access points to that file system so that you can synchronize workloads around the world. Now what you have here is appliances – edge appliances – that are going to be inside of your data centers and they’re going to be communicating to your application layer with CIFS or NFS or any other traditional file access protocols, HTTP. And your application layers are running here now the powerful concept here is that as the applications change file data locally, what is happening is the system is resynchronizing those changes up to the cloud storage system. It then propagates the changes in a hub and spoke architecture so that every one of these other sites that could be you know this could be across the ocean, it could be on the other side of the world, it’s going to be replicating those changes around and around and as you are able to access those changes say here via CIFS on a second application you’ll be able to make some other changes to the files and be able to synchronize those changes again here and then be able to see them not only in the first site but in every other side that are sharing that global file system. We adjust the size of these appliances so that they correspond to the IOPS, so the performance levels that are needed by these applications, but really the capacity of the file system and the stability and the data protection of that file system is all here in the back end of the system.
At Nasuni we work with [Amazon] AWS, we work with Microsoft Azure, IBM COS Cloud Object Storage as well as the EMC ECS system, so you have plenty of choice in terms of what you can build this back end for cloud technology and then you overlay the Nasuni file system on top of it and then these appliances give you access to the file system. What that means to a large global organization is that the files which have always been the kind of work product of many, many businesses in media, manufacturing, engineering, architecture firms – the files are now available at very high levels of performance on full read/write across any number of locations and unlike traditional WAN acceleration technology, you’re not really trying to accelerate just the network part of the protocol, but you’re truly replicating the file system across all of those locations. This becomes a very powerful way of orchestrating file heavy workloads around the world for teams that are trying to collaborate on very large design projects it may be that they have Autodesk Revit or creative teams that are working with Adobe media applications. This really eliminates the problem of, “How do I get a lot of that big heavy file data to all of my locations around the world so that we can make the business far more efficient?”