NeoPath offers clear route to flexible file server management

File Director's virtualized network shares make storage changes a walk in the park

Many organizations find that file servers tend to pop up all over the place and are often hard to manage. Worse, consolidating the servers may disrupt end-users, because consolidation usually requires changing network mappings and user permissions.

NeoPath’s File Director offers a solution to all those woes: virtualizing network shares and presenting them in a manner that users are familiar with. Also, the file system hardware can be changed in the background as desired, with the users remaining blissfully unaware of the alterations.

This virtualization capability isn’t unique (see our review of the OnStor Bobcat 2200 NAS Gateway), but NeoPath makes it simple to use and provides great storage management tools as well.

Share and share alike

File Director handles both CIFS (used by Windows) and NFS (used by Unix and Linux) volumes, allowing the administrator to create new shares or virtual volumes that match the names of existing network file servers, removing the need to redo drive mappings but requiring that the original file servers’ IP addresses be changed or new names be created.

Either way, once File Director creates the virtual shares, there is a complete disconnect between the hardware and what the end-user sees in the file browser. Files can be migrated to second-tier storage, moved from one file server to another, consolidated on a single system, or distributed for fault tolerance — all transparently.

Users can access the files, and even change them, while the administrator is moving files from one back-end location to another. Any writes to the files are preserved and made to both copies of the file if necessary.

In addition to virtualization, File Director offers management tools. Files can be identified by size, last time accessed, file type, frequency of use, and more; they can also be automatically or manually moved from one back-end volume to another based on these criteria.

For instance, you might move all JPEG files from the high-speed file server to a second, lower-performance file server, and users would never know the files had moved because they would still appear to be in their original locations. Even if File Director is removed, hidden directories pointing to the new locations are maintained, so the files will not be lost.

Naming rights
I reviewed a single File Director, but most enterprises will opt for the clustered pair configuration for high availability. The price is $29,995 for one and $49,995 for a pair. According to NeoPath, one pair should be enough to support as many as 20 back-end file servers or a maximum of 100TB of user data.

File Director’s setup is straightforward, with no surprises in its initial configuration. The only tough decision is whether to duplicate existing volumes for transparency to end-users. If you create share names identical to your existing shares, you won’t have to make changes to the workstation drive mappings, but you will have to change your DNS mappings for the host names of the existing back-end servers, by changing their host names, their IP addresses, or both.

Alternatively, you may create new share names and make all the old shares available through the new names. This has some advantages when it comes to consolidation. For example, you can create a new share called “home” and virtual directories under /home that correspond to the old shares: /home/fs1/users, /home/fs2/users, and so forth. Users would simply look in /home/users for their home directory and wouldn’t need to know which file server they were using.

This process is especially useful if you need to create a new, logical structure for file services. The nice part is that you can consolidate multiple file servers into a single directory structure without changing the file servers.

It’s also possible to consolidate directories on different types of file servers, presenting both CIFS and NFS volumes as a single virtual volume. File attributes, permissions, and other properties are properly maintained depending on which type of user accesses the volume.

Virtual meets physical
File Director’s storage management utilities are easy to use, and it’s simple to set up queries to figure out which files a new policy would affect before you actually implement the policy. You may then move files from one back-end storage device to another based on nearly any imaginable file attribute.

Policies can be created manually and run once, or you can schedule them to run at any desired interval. You can nest search criteria, too: For example, you could select all MP3 files of more than 100KB that are not in a specific directory to be migrated to second-tier storage.

Because File Director offers a unified view of all storage in the enterprise, it allows full utilization of existing storage without running out of physical space. Virtual volumes can span multiple physical volumes, so you can add space to a virtual volume without having to add physical drives to any server -- either take space from an existing share or add drives anywhere on the network that’s convenient, and link their capacity to the existing virtual volume.

The NeoPath File Director offers a simple way to establish order from chaos in any organization with a fragmented directory structure. Because File Director transparently virtualizes storage to end-users, you can migrate some data to less-expensive storage without having to train users, or you can change the logical structure of your network data without having to move data.

With File Director, administrators get great flexibility for reconfiguring data services with no impact to users, even while moving files. In a sense, it’s like having a very sophisticated, monolithic SAN-based storage system without having to buy any new hardware.

InfoWorld Scorecard
Setup (20.0%)
Performance (25.0%)
Value (10.0%)
Ease of use (20.0%)
Manageability (25.0%)
Overall Score (100%)
NeoPath File Director 8.0 9.0 8.0 9.0 9.0 8.7
Recommended
Join the discussion
Be the first to comment on this article. Our Commenting Policies