Go to Top

Newly Implemented Technologies in Windows Server 2016: Better for Data Recovery or Not? (Part 2)


In the first part of this article, we explained the benefits and challenges concerning the new file system, ReFS v3, intended for storage use in Windows Server 2016.  We now want to focus on another technology which has a deep impact on data recovery in case of loss – Storage Spaces and Storage Spaces Direct.

Storage Spaces Direct: Microsoft´s answer to VMWare vSAN making data recovery even more complex

Storage Spaces Direct is the evolution of its Storage Spaces technology which was introduced in Windows Server 2012.  In short, Storage Spaces means that all data is stored in a virtual storage pool based on physical hard disks or SSDs. The advantage of this concept is that one can easily add new disks to the pool without any huge adjustments or problems.

Storage Spaces Direct, which is only available in the Windows Server 2016 Datacenter edition, is brought to a higher level – it is now possible to combine local volumes of different servers in a cluster to one single pool (instead of just all physical drives of one server into a data pool).  In contrast to the previous version, which was only capable of using SAS-JBODs, it’s possible to use different hard or flash disks – SATA, SAS or SSD.  All of these storage types, except for disks that are attached using “multipath,” can now be used to create one single storage pool.  Additionally, Microsoft integrates the Software Storage Bus, which – to keep it simple – replaces the old Fibre Channel Cables, integrating networking between several servers and establishing a software-defined storage fabric where all servers can see all connected drives.


The redundancy concept of Storage Spaces is similar to the concept of RAID, but the technology used is different and software-based.  During setup one can choose between a simple, mirror or parity storage arrangement.  Simple stands for no redundancy.  Mirror simply mirrors the disks when two disks are available (One disk can fail).  If three disks are available, two can fail.  Finally, parity is based on RAID 5 and needs three disks in one storage pool to prevent data loss.

With this new technology implemented in Windows Server 2016, it’s possible to implement both a converged storage solution (storage and compute in separate clusters) and a hyper-converged solution (one cluster both for compute and storage).  This is the main difference between VMware vSAN and Windows Server 2016.  Microsoft also allows its users to create its own storage tier because Storage Spaces Direct is not combined with the Hypervisor.

Whether one deploys a converged or hyper-converged solution with Windows Server 2016, Microsoft claims that with its new enhanced version of Storage Spaces the data stored inside the pool is even more secure than ever before.  Microsoft states, “Storage Spaces protects data from partial and complete disk failures by maintaining copies on multiple disks.  ReFS interfaces with Storage Spaces to automatically repair the corruption.”

What now?

Although Microsoft claims that Windows Server 2016 is the basis for the future of server technology and implemented a lot of interesting and useful (but complex) features, this can also pose a threat for successful data recovery.  With the combination of two advanced technologies like ReFS and Storage Spaces Direct, which in reality is Microsoft´s (hyper-) converged storage and software-based solution based on HyperV virtualization and virtual machines, it makes data recovery a more complex and challenging task.

And even though professional data recovery specialists at Kroll Ontrack already have successfully recovered data from multilayered storage systems or servers like VMware vSAN or HP EVA 6000, and have developed tools to rebuild both the system as well as the data structures, it’s  a more  challenging, time-consuming and costly project.

So when you’re using Windows Server 2016 try to keep your system’s design as simple as possible, hold back from using too many features at once (e.g. deduplication, virtualization and data compression all in one system) and document everything on the system for future reference in case of an emergency.  You will be thankful to yourself!

Leave a Reply

Your email address will not be published. Required fields are marked *