I’ve been in IT for 10 years, and now work primarily in the MS stack of technologies with MCSA/MCSEs.
On a day-to-day basis, I consult for customers and partners alike, looking at how Azure can help solve business challenges.
Every week brings a new challenge, and by ‘getting my hands dirty’ with Azure services I can see the solutions in action – so, I’m going to be blogging regularly on these challenges, and how Azure can help solve them.
So, what’s first?
I have a challenge around getting a lot of files into Azure, what’s the best way to do that?
Azure Storage is a scalable, durable, highly available cloud storage solution in every Azure region.
Azure storage can store Blobs, Tables, Queues and files. Each type has a particular use case, and I hope to write about using each type as time goes on. In this blog, though, I want to look at a real-world challenge faced by a customer and a potential route to solving it.
Consider the case of Contoso. They have a file server on premise containing 5TB of sensitive documents. They need a solution to get these files off their on-premises server and onto an IaaS virtual machine running in Azure, but not to impact on-going operations on those files.
Enter Azure Files and Azure File Sync (currently in preview.)
Azure Files is a type of Azure storage that is “wrapped” around an SMB share. It supports most of the SMB 3 suite, supporting encryption in transit for example. It is an ideal solution for somewhere to drop files either to be archived off, or to support legacy applications that aren’t configured to use cloud storage. You can mount it as an SMB share on a server just like you would a normal file share.
Azure File Sync is an agent based solution that can synchronise files between groups of servers and an Azure Files share. Install the agent, and register it to your Azure subscription and you can configure a file path to sync to the Azure Files share. The brilliance is you can add additional registered servers to the sync group to bring those files back down again.
I have just set up a proof of concept for Contoso, installing the sync service to their on-premises server, and watched as the files synced from the on-premises server, into the file share and then back down onto the new VM running in Azure. Over the course of a few days the 5TB of data will gracefully sync up and with no impact to users still working on them. Going further, Azure File Sync supports cloud tiering – I can actually provision significantly less storage to the Azure based VM, with the majority of least accessed files stored in Azure Files storage. Only a subset of these files is replicated to the virtual machine for access by users. Anytime they request a local copy, it is transparently and seamlessly copied down from the storage account to the server.