As file sizes grow and retention periods get longer, companies are distributing data across more tiers of storage to satisfy budget and legal constraints. However, spreading data across multiple NAS appliances creates new difficulties when it comes to providing the right level of offsite data protection for these different types of file servers. A file virtualization appliance can help address this situation.
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
File virtualization appliances sit in front of network file servers and create an abstraction layer between file servers and the clients that access the files. The file virtualization layer may catalog files and file systems across the file servers and may optionally present a single logical file mount point that's accessible by all client servers. Once the file virtualization is in place, network file servers continue to host the files and file systems. But the file virtualization appliance provides advanced file management by using the metadata gathered from the network file servers for optimized remote data protection.
Using this metadata, companies can use the file virtualization appliance to introduce file filtering to limit the number of files that are remotely protected. Changes to files may occur on any of the virtualized file servers at any time, but companies may only want changes to files on tier 1 file servers replicated immediately. Conversely, file changes on less-critical tier 2 and tier 3 NAS filers may only be replicated periodically to minimize network bandwidth requirements.
Vendors offer a couple of different ways to control this. Attune Systems' Maestro stores the metadata it collects from the files and file systems of the NAS filers and network file servers it virtualizes in an NTFS file system. Part of the data stored in the file metadata is the tier of storage on which the file is stored and the file modified date and time. Maestro then stores a copy of the metadata in a SQL Server database for faster indexing and search.
This information can be used in two ways. To do remote replication with Maestro, users can set up a job schedule so that snapshots of the volume sets are taken. Maestro then transfers or replicates the snapshot using NDMP to copy to a remote location. Because Maestro is storage-agnostic, the target NAS filer may be of another tier or from another vendor to help control costs.
Backup software can be used in conjunction with Maestro to generate a list of files that have changed since the last backup. By using this list of changed files that delineates what files are on second and third tier file servers, a schedule can be created to copy the files to remote sites based on which tier of file servers that the files are located.
Using EMC Corp.'s Rainfinity file virtualization appliance, policies can be set to redirect file requests so that specified files are stored in multiple physical locations, local or remote, at the same time. Rainfinity leverages Microsoft Windows Distributed File System (DFS), which ships with every Windows client and makes namespace updates transparent to the client. Administrators can set policies so that files on certain NAS filers, file systems or specific file types are automatically replicated to other locations.
To prevent delays in the production application if the file is replicated to a remote location, Rainfinity gives administrators the option to throttle how much data is sent. In this mode, it semi-synchronously copies the files to other sites so companies can use lower bandwidth network pipes.
F5 Networks Inc. Acopia ARX Series file virtualization appliance is another option companies can use to improve remote-data protection. Administrators can configure the ARX appliance to aggregate local file storage with datacenter-based file storage. Once aggregated, administrators can set policies to keep the most actively accessed files local to ensure LAN type access speeds but can use ARX's replication feature to create a copy of all data at the main data center for offsite data protection.
Companies may also opt to introduce a global namespace that works across all NAS filers (CIFS or NFS based) using the file virtualization appliance. If a global namespace is implemented, the file virtualization appliance can redirect file requests to any NAS filer, local or remote, should the primary NAS filer become unavailable. While companies still need to account for and manage network bandwidth to remote locations, a simple change in the network DNS can automatically redirect file requests from the local filer to the filer at a remote location.
File virtualization is a powerful but sometimes overlooked way to provide heightened levels of remote data protection for network file servers. File virtualization appliances can permit companies to leave their existing NAS infrastructure in place while providing a means to achieve a higher level of file recoverability in the event of either a minor and major data loss.
About the author: Jerome M. Wendt is lead analyst and president of DCIG Inc.
Dig Deeper on Remote data protection