buchachon - Fotolia
It makes sense to store files in the cloud for several reasons. First, files are not as sensitive to latency as application data, where the added latency of the trip to the cloud could crash the app if the data comes back slowly. Storing files in the cloud generally involves users browsing and sharing files, where some amount of latency is less critical. Second, files are increasingly shared with users around the world, so it becomes less important to centralize their location at an enterprise data center. If most users will have to access files from far away anyway, they might as well get them from the cloud, which might offer ways to accelerate the traffic and position data at key points of presence more effectively than on-premises resources can. Finally, file data is growing rapidly for most companies, but most of it isn’t accessed frequently. This means that it might not make sense to keep all of it at your fingertips -- enterprises can instead save money while ensuring that the data is protected within the infrastructure of a cloud provider.
However, no major cloud providers offer a true file storage service today. The most common standalone cloud storage service is object, which can be used to store files (as a single object or otherwise) and can be accessed similarly, but the syntax and processes used for accessing files doesn’t work for objects. For a user that wants to store files in the cloud, there are several approaches to take:
- Run a file server within a cloud server and use block storage as the back end. You can run most anything within a cloud server, including a file server. This allows you to build out a high performance connection between the server and associated storage, and use the Windows or Linux file serving that you’re already familiar with. Unfortunately, this can also be expensive. It also doesn’t take advantage of the unlimited scalability and inherent resiliency of object storage services.
- Write your own file front end to the cloud object repository. It’s not rocket science, but there is a level of effort required to build out the browsing, sharing and updating capabilities to use object storage as a file store.
- Go to a third party that has already built out a file front end for the object back end. Vendors like Dropbox, Box, Bitcasa among many others give you the ability to leverage object-backed cloud services, with a more familiar, intuitive file oriented front-end.
- Use a cloud gateway for file to object translation. Cloud gateways do a variety of things, but one of them is to translate data into objects for storage in the cloud. If you have a library of files that is saved within a gateway, the system will use policies to determine what data to keep on-premises, and what data to store in the cloud. Some gateways are intended to keep 100% on-premises, and use the cloud as an automated backup and data protection solution, while others are intended to use the cloud as the primary repository for all but the most actively used files. Some gateways have file capabilities built in, where others use a block connection, and require you to run a file server within the system. One potential downside of this approach is that some gateways prevent you from accessing cloud resident files directly, meaning that without the gateway, files are obscured or inaccessible.
Given the generally low sensitivity to latency of files, the need for sharing files worldwide and the high growth of file data, it seems likely that storing and sharing file data will be a huge growth engine for cloud providers, but today, there’s no way to do this elegantly and without third-party tools. Keep watching this space for new options, and move forward carefully if your organization is ready for cloud file storage now.
Useful third party tools for AWS management
Choose the right enterprise cloud storage and file-sharing service for you
Dig Deeper on Cloud object storage
Have a question for an expert?
Please add a title for your question
Get answers from a TechTarget expert on whatever's puzzling you.