This post will discuss what the general availability of Azure File Sync will mean to customers that are continuing to use on-premises file servers.
For over 5 years, you’ve been hit over the head repeatedly with the mantra of cloud is cheap, highly available, and easily accessible, and that you should store your files there. But many people do not store files there, or only store some of them there. The role of the file server has continued for many businesses, but with continued pain points.
Azure isn’t just about “move everything to the cloud”. Azure is built to do that “digital transformation”, but it also offers many hybrid services to improve existing on-premises investments in IT. One of these is Azure File Sync which is designed to improve how file servers work, and to reduce capital and operational costs & complexities.
I don’t think I’ve ever been in an office where there has been enough disk capacity in the file server. No matter what is done, there is never enough disk. You can keep buying disk until the server is full, then add on a NAS or LUN capacity from a SAN. In many organizations, the only server left in the building is a file server, so keeping backup and disaster recovery (DR) solutions around is an excessive but necessary cost. Good news: Azure File Sync, which is generally available now, aims to solve these problems and more.
Azure File Sync leverages the low costs of Azure Files storage in The Cloud. An agent is installed onto your file server (Windows Server 2012 R2 and Windows Server 2016 only) and registered with a sync service in Azure. A cloud endpoint is deployed into a general purpose storage account. And then you decide which folders you want to synchronize to that storage account. This differs from other cloud synchronization solutions, such as Microsoft’s own StorSimple service, which requires you to move files to an appliance because you synchronize with Azure File Sync from the original storage location.
Once you synchronize the selected folders to the cloud, the copy in Azure should be considered the master copy, and what is on the file server should be considered a hot local replica.
Note that the engine being used for data transmission isn’t new – it’s based on technology from SQL Server.
If you are struggling with never having enough disk capacity and cannot remove old files (for whatever reason) then you will like this feature. You can enable cloud tiering of cold files – files that are rarely or never used. Once a policy is enabled, the least used files are removed from the file server and replaced with metadata in the file system.
From the users’ point of view, the files are still there, just with a different icon, and an “O” attribute if they go looking for it. The same permissions still apply too. If they try to open the file, it is downloaded by the sync agent from Azure to the file server and presented to the user as normal. With this feature, the ever expansion of storage in the file server will cease and you will expand into the low-cost storage of Azure by moving old files there, without you having to say what is old and what is not.
You will no longer backup your files on the local file server; instead, you will back up the master copy of the files in the Azure storage account. This reduces costs even more:
If you lose the file server, you’ll be back online in no time. Drop in a new file server, install the sync agent and register it with the old storage account. The metadata of the shares/files/folders and the permissions will be downloaded and presented to the users as if they were using the old server. Files will be downloaded on demand from Azure and eventually the file server will be populated with the hot files.
A company with multiple locations can use Azure File Sync instead of DFS-R to synchronize shares. The benefit is that Azure is the master copy that offers centralized backup and DR.
There is a catch. Today, Azure File Sync does not have the ability to distribute file locks. The solution is the “OneDrive solution”. If two people open a file and save changes, then both copies are saved, and it’s up to the users to sort it out. That’s not perfect but:
I have been waiting on AFS for quite some time – I asked Microsoft for something like AFS years before the announced it! I know that many companies are in a hole with their file servers, and AFS does almost everything that these organizations need. I can see AFS taking off and being widely used, and like with Azure Backup, being one of the entry points into the cloud for cloud-first-timers.