Data Storage Digest

Do-It-Yourself Windows File Recovery Software: A Comparison

results »

Examining the New Storage Spaces Direct Feature of Windows Server 2016

Although Windows Server 2016 has not been released at the time of this writing, preview and beta versions have been made available to select professionals in the IT industry. With ample time to test out the software up to this point, experts are beginning to grasp the concepts behind some of the new functionality. This includes the new feature, Storage Spaces Direct, which serves as a complement to the previous software-defined storage stack of Windows Server.


Storage Spaces Direct can be likened to the Storage Spaces feature seen in Windows Server 2012 R2, which enables the creation of highly available (HA) networks of attached storage. While an incredibly efficient and useful means of combining data from multiple physical devices, the outdated Storage Spaces feature had many shortcomings. Firstly, in order to enable the physical connection of disks across all storage nodes, every drive must be stored in an external JBOD chassis. Secondly, only SAS devices are compatible with Storage Spaces, which is sometimes referred to as Storage Spaces with Shared JBOD.

Another system setup involves the use of various internal disks, which are actually located internally within each separate storage node. While this is more efficient that Storage Spaces with Shared JBOD, it still wasn’t quite what Microsoft’s development team was aiming for.

Thanks to Storage Spaces Direct, which is being dubbed the evolution of Storage Spaces, each storage node is connected to its own JBOD. As a result, the requirement of the shared SAS architecture is completely eliminated. Instead, SATA disks can be used.

The Storage Spaces Direct Stack

Microsoft’s Storage Spaces Direct stack includes seven different layers, all of which combine to form Microsoft’s software-defined storage system. Starting at the bottom of the stack we find the primary hardware, which includes a minimum of four nodes of local storage. Each node is capable of utilizing internal disks or external disks via SAS connection.

At the next level of the stack we find the software storage bus, which establishes consistency and visibility amongst all the storage nodes. Moving upward is the storage pool, which spans all local storage sources. Storage Spaces, or virtual disks, are contained within the next layer of the stack. These offer resiliency and redundancy by storing copied data across the various storage nodes.

Next is the Resilient File System, or ReFS. The ReFS controls the architecture that contains Hyper-V virtual machine files. Above this layer we find the Clustered Shared Volumes, or CSVFS layers. These serve to collate all of the individual volumes into a centralized entity.

Finally, the top layer of the storage stack is covered by the scale-out file server. Facilitating remote accessibility to the storage system, the scale-out file server relies on the SMB3 protocol to dictate user access.

Current State of Testing

The easiest and most straightforward way to evaluate Storage Spaces Direct within the Windows Server Technical Preview is to use four Hyper-V virtual machines, generation 2, as well as two disks per virtual machine. PowerShell should be used to manage the deployment and day-to-day functionality of Storage Spaces Direct.


No comments yet. Sign in to add the first!