What you will learn in this tip: File virtualization software can be a good solution for growing storage environments and can ease
the burden of managing multiple file servers. This tip will look at using file virtualization as a solution for Windows file server sprawl.
Anyone who has ever managed file servers or network-attached storage (NAS) typically loves the simplicity and ease of this technology. This is why so many small- to medium-sized businesses (SMBs) utilize it for their shared storage. Many of you reading this tip will recognize it as your "D" drive or shared drive.
But as more file servers are added to the network, this eventually causes a management nightmare for unstructured data and applications that utilize file server storage. This is because the data has to be load balanced and often migrated between systems, and even so it still creates uneven storage utilization. Each system has to be managed individually and each storage system also increases power, cooling and infrastructure costs. And maintenance of older device is more frequent and more costly than newer ones.
To solve this problem, you could buy a scalable NAS system and consolidate the file servers or buy a pay-as-you-go NAS system. However, a better alternative may be to take a hard look at file virtualization software, which includes the following options for dealing with Windows file server sprawl:
- Free Microsoft's Distributed File System (DFS) software
- Sanbolic's Clustered File System (Windows software)
- AutoVirt File Virtualization (Windows software)
Microsoft DFS for Windows
Microsoft DFS has the very positive benefit of being free. It only works on Windows file servers; however, it enables multiple different Windows file servers' shared folders to be grouped into one or more logical namespaces where it is seen as a single shared folder with multiple subfolders. DFS automatically connects the user or application to the shared folders in the same available Active Directory Domain services site. DFS alleviates unnecessary LAN or WAN routing while its optional replication capability enables folders to be automatically synchronized between file servers locally or geographically WAN distributed.
But there are serious shortcomings with DFS, such as requiring extensive Windows knowledge and expertise. Additionally, there is loose synchronization between files on different servers, especially geographically dispersed. This means users or applications at a remote location may be accessing an older file before it can be updated. Micosoft DFS doesn't work with NFS-based or Linux file servers; has limited scalability; doesn't provide file-level granularity; provides poor storage utilization from large amounts of duplicate files; typically requires significant additional hardware infrastructure costs; and it's not really architected to scale to very large numbers of filers.
Microsoft DFS works with Windows file servers, and the backend storage of those file servers can be internal direct-attached storage (DAS), external DAS, external storage area network (SAN), or external NAS.
Sanbolic Clustered File System for Windows
Sanbolic's Clustered File System is a lot more functional and intuitive than Microsoft's DFS with none of the inherent limitations. It was designed from the ground up to fix Windows file server sprawl at the OS level. The software provides byte-range locking. That's important because byte-range locking enables high-performance concurrent shared read/write access to data on SAN-attached storage from multiple physical or virtual servers. This in turn enables high availability, horizontal file server scaling, app scaling, higher performance via clustered processing, with intuitive storage provisioning and management. The file system also incorporates quality of service assignment to allow prioritization of specific workloads in storage bandwidth-constrained environments. Sanbolic's Melio FS additionally has optional configuration automation management reducing much of the headaches that typically come from configuring clusters of multiple Windows file servers.
This software has none of the drawbacks of DFS and NTFS, but it's not free. Fortunately, it's not very expensive either. But just like DFS, it doesn't work with other NAS systems. SANbolic takes the place of the Windows file system. In other words, instead of NTFS, (which is the Windows native file system) it replaces it and provides a distributed or virtual file system across multiple systems. The backend storage can be a variety of storage types including NAS, but it's not clustered NAS (specifically CIFS or NFS).
AutoVirt File Virtualization for Windows
AutoVirt is a pure file virtualization software option similar to Sanbolic or Microsoft. AutoVirt is a split-path Windows-based file virtualization software appliance. It creates a shared global namespace that virtualizes Microsoft's UNC (Windows path name to a file) and disconnects it from its actual physical location, enabling one or more logical naming structures for all files across the network, independent of the physical file servers. In other words, it allows multiple physical or virtual Windows file servers to appear as a single file server image or mount point, and it's extensively scalable because of its split-path architecture (the control path does not get in the way of the data path). AutoVirt also enables automation of many management and data protection tasks aiming at reducing admin workloads. Ultimately, AutoVirt solves Windows file system sprawl.
Of course, nothing's for free and, in addition to AutoVirt's licensing and maintenance, there will be additional physical or virtual server costs and infrastructure costs. AutoVirt also doesn't work with NFS or NFS NAS systems. AutoVirt is a GNS appliance that only works with CIFS, not NFS.
There are the file virtualization software choices for SMB Windows file system sprawl. All of them are viable; however, keep in mind one size never fits all.
About the author: Marc Staimer is the founder, senior analyst and CDS of Dragon Slayer Consulting in Beaverton, OR. The consulting practice of 12-plus years has focused in the areas of strategic planning, product development and market development. With more than 30 years of marketing, sales and business experience in infrastructure, storage, server, software and virtualization, he's considered one of the industry's leading experts. Marc can be reached at firstname.lastname@example.org.
This was first published in February 2011