informa
/
Big Data
Commentary

Storage Should Not be Treated Like an Unloved Part of IT

Because storage needs and technologies have diversified greatly over the past 20 years, it makes sense to enact an IT strategy that coalesces with overall IT architecture planning.

Storage should be a strategic IT concern -- but there are many reasons why it isn’t. First, storage is comparatively cheap -- and easy to add if you have to increase it during a budget year. Second, storage is often included in new systems that IT purchases, so it doesn't have much cost visibility. Third, when IT needs to scale, the scalability argument is for systems and networks. Storage is seldom part of the conversation.

Yet, with data management and digitalization assuming major roles in IT project work, the pressure on storage is immense.

This begins with the versatile (and new) roles that storage is playing in IT architecture.

At its most minute level, you can now find storage on some SoCs (systems on a chip) that are used in edge processing -- or if storage is not directly on the SoC, on chips that contain RAM and standard storage. The most common example of this is the smartphone, but it can also be other edge devices.

From there, storage spreads itself across SAN and NAS networks that distribute its use at the edges of enterprises, and in centralized data centers. In these configurations, storage can appear as either a straight disk or as a file server. Regardless of how it is implemented, storage must be allocated based upon the types of applications in the network it supports.

Finally, there is large-scale storage in data centers and in the cloud -- and a hierarchy of storage that ranges from very slow, cold drives for seldom accessed data to in-memory and solid-state storage that is used to process rapid transactions in high-demand, real-time data environments.

Because storage needs and technologies have diversified greatly over the past 20 years, it makes sense to enact an IT strategy that coalesces with overall IT architecture planning.

“Having a good data storage strategy in place is no longer a luxury for today's organizations,” wrote Kal Rana, Sr. Solutions Architect at vxchange. “Data storing management is more important than ever, largely because the ability to derive actionable insights from data is often a key differentiator for business success. A data storage strategy doesn’t just outline how a company stores its data, but also how it makes use of that information.”

So, let’s see how digitalization, speed of insights, speed of transaction processing and security factor into data storage -- and why storage should be an IT strategic focus. Here are four keys to data storage strategy:

1. Speed versus cost

How well do you know your data usage patterns?

Key to storage and overall IT performance optimization is being able to know which data is used most often in the enterprise, which data demands fast retrieval rates, and which data is seldom or only moderately used.

More expensive solid state and in-memory storage should be reserved for highly critical data that is often and rapidly accessed. At the other extreme, data that is seldom accessed should reside on cheap cold storage drives. Between these two extremes is data that is moderately accessed but that doesn't demand rapid real-time processing. This data can be stored on intermediate storage media that capably access and retrieve data, but that don't have to do it at breakneck speeds.

Being able to organize data and applications so that mission-critical, rapidly processed data is on the most expensive SSD and in-memory storage, but less critical data with lower access requirements resides in cold storage or on slower, more moderate storage, is an important element of storage strategy that guarantees optimal performance at least cost.

2. Infrastructure fit

How well do your storage deployments mesh with your overall IT infrastructure?

With more organizations moving applications (and storage) to the cloud, there are looser controls on how storage is being planned and allocated because provisioning storage in the cloud can literally be done at the touch of a button -- and by anyone.

This invites storage waste (discussed later), and it also has the potential of allocating the wrong types of storage to the wrong applications.

In the data center, there are still misappropriations of storage, but misappropriation is less likely to occur because IT is in a better position to directly supervise storage usage and deployment.

3. Waste elimination

Ten years ago, companies began moving their IT application testbeds to the cloud because it was easier to provision virtual OS servers and storage than the alternative of tying up internal IT resources (like the database group) to perform IT test set-ups in-house.

The man-hour savings in the data center were great, but one unfortunate outcome has been an excess of server and storage resources that remain deployed (and paid for) in the cloud long after application testing is over -- because developers forget to de-provision these resources after testing.

A second storage waste source is the needless storage of data that the enterprise will never use. In some cases, such as retaining documents that could be needed for legal discovery, storing older data is necessary. However, in the majority of cases, IT fails to do its annual reviews with users by meeting to decide which data is necessary to store -- and which can be discarded.

A third area of storage waste is indiscriminately ingesting all incoming data without screening it for irrelevance. Storing IoT data that hasn't been filtered for machine noise is one example. Another example is committing data to storage that hasn't been cleaned and prepared by techniques such as data cleaning and deduplication.

Every minute of every day, according to research by Domo, there are 52,000 people meeting on Microsoft Teams and 280,000 people meeting on Zoom, while WhatsApp users send 41.7 million messages. With data production like that, it’s incumbent on IT to develop data storage strategies that address data safekeeping -- and determine which data should be eliminated altogether.

4. Security

Data breaches and storage compromises are career threats to IT managers and brand threats to companies.

On the storage front, these breaches can occur in two primary ways:

  • Employees within the company store (and share) data on workstation hard drives that IT isn't aware of; or employees walk out with data on thumb drives or other devices, that can be shared;
  • Cloud providers hosting corporate data lack rigid security, and the data gets breached.

For employee data, many companies now store all data on enterprise servers and use thin client machines for employees. Companies also use data tracking software that can detect when an employee attempts to download data onto a thumb drive or other device. This cuts down the ability of employees to store data on their own devices.

The second area of security risk is what happens to data that is stored in the cloud. On an annual basis, security audits from cloud providers should be requested so IT can review cloud security. When cloud contracts that enterprises form with cloud providers are being negotiated, a baseline requirement should be that the cloud provider meet or exceed enterprise security requirements for data and storage.

Finally, for companies using a multi-tenant approach in the cloud where resources are shared with others, it’s critical to assure that applications, data and storage are secured and firewalled at all times from the data, storage and applications of others.

What to Read Next:

New Storage Trends Promise to Help Enterprises Handle a Data Avalanche

Looking at the Cloud in 2021: Growth and Changes

How CIOs Should Manage the Rapid Shift to the Cloud