Analyst View: How to lower file-based storage costs


Noemi Greyzdorf, IDC
For a number of years, organizations have been experiencing exponential data growth and have been buying more and more storage to support it. The slowdown in the economy is not expected to impact the creation of content or the requirement to retain it. As a result, the need to store data will continue to grow while budgets will be cut, which creates an incongruity: If there is more data, then money must be spent on acquiring more storage.

With fewer dollars available for more storage, organizations will be seeking ways to achieve the required result with fewer dollars. This creates an opportunity for storage administrators to design for efficiency and for vendors to bring innovative solutions to market.

Files are the fastest growing data type. Unstructured, file-based data is consuming more space and is being retained longer. This is a challenge that cascades down from the pri- mary storage tier to the data protection and archiving tiers.

The three main categories of file-based data are file and print, application-created data, and content depots. Some suggestions in this article will apply across all categories while others will be specific to a particular category of file-based data.

Here are a few suggestions on how to keep your file-based storage costs under control:

  • Nothing can be done to control the costs of file-based storage without knowing who is creating the majority of the data, and to what end. I have often been surprised when organizations reveal that employees can store any data on their network shares. With advances in imaging technology, higher resolution files consume more storage capacity; so, if employees are storing pictures, music, or videos, they are consuming a significant amount of storage with content that is of no value to the organization. A typical .jpg file can be as large as 2.75MB with a resolution common to today’s digital cameras. That can add up pretty quickly.

Keeping an eye on what files are being stored on the network can help you save significant capacity on your file servers. To achieve this, you can leverage storage resource management (SRM) software, or you can use open-source software for basic SRM reporting.

  • In recent years, there have been a number of technology advancements that facilitate capacity utilization and efficiency. These technologies can be put into three main categories: capacity optimization, storage virtualization with intelligent provisioning, and resource aggregation.

Capacity optimization refers to using compression, single instancing, and data de-duplication to reduce the data footprint and thus improve utilization of capacity. Optimization technologies have been primarily used for backup and archiving, but there are products on the market that can also reduce your primary file-based data set. If the acquisition cost of such technologies is less then the acquisition of more storage capacity, then there is value to be realized.

Storage virtualization delivers true value through features such as thin provisioning, which enables organizations to delay the acquisition of hardware to the time it is actually required. Thin provisioning alone, however, won’t deliver a big cost savings since capacity is often over-provisioned to increase spindle count for performance; block-based virtualization that can optimize capacity use and deliver performance would be necessary to get true value. Since hardware costs continue to decrease on a per-gigabyte basis year over year, delaying purchases can provide significant cost savings in the long run as well as free up budget in the short term.

Server virtualization has done a lot for data center consolidation and efficiencies. In the file-based storage world, data is still stored on single-instance file servers or NAS devices. In many organizations, this poses a major challenge. Some systems end up being under-utilized, while others are |arely handling the load. Administrators end up spending a significant amount of time load-balancing users, migrating data, and provisioning new storage hardware.

Unifying or consolidating file-based storage systems under a single umbrella or file system that can load balance and scale seamlessly can go a long way to increasing utilization, improving performance, and lowering management overhead. This is achieved through the deployment of distributed file systems or NAS devices that present a single file system across multiple server nodes. Each node represents performance increases, as well as potentially more addressable capacity.

  • One of the more interesting developments in the storage market has been the creation of Storage-as-a-Service (SaaS) based on cloud storage architectures. Using such services means having the ability to receive storage “on the fly.” Organizations that either have seasonal data or are looking for options in deploying storage resources could consider cloud storage as a way to augment their existing infrastructure. Data that is static or archived, as well as backups, can be stored in the highly-available storage cloud. By utilizing public cloud storage, fewer capital expenditures are required and capacity costs can be delayed. Paying for only what you consume also results in more free space in the data center.

End users will continue to struggle to deliver what business demands with smaller budgets, and vendors will be looking for ways to stay afloat with decreasing orders. Such an environment creates an opportu- nity for both sides of the table. End users have an opportunity to examine how they are handling storage demands today and what they can do to achieve greater efficiencies and accomplish more with less. Vendors also have an opportunity to differentiate their offerings with improved utilization, increased efficiencies, and investment protection.

Some may be able to take advantage of this opportunity with existing products, while others will have to seek partnerships or acquisitions. This is not the way we typically want to institute change, but on the other end of this economic downturn waits a more efficient and smarter way of managing file-based information and its underlying infrastructure.

NOEMI GREYZDORF is a research manager at IDC (www.idc.com).

This article was originally published on February 01, 2009