How to Save $1 Million on Enterprise Storage

How to Save $1 Million on Enterprise Storage

Posted in tech

Big money, big promises? You don’t have to take our word for it.

Savings sound great in blog posts and social media, but they are far better when they actually hit your bottom line. To show you just how much control you can have over how much storage costs you, and to help you give CXOs and application owners the cost of meeting application requirements, we created the Data Profiler. The Data Profiler tool lets you examine the effect of your data management policies, making assessments using your actual storage infrastructure. With a slider, you can see the savings impact of tiering data that’s been idle long enough to call it cold.

With petabytes of data being managed at most large enterprises, the savings possible with DataSphere easily scale into the millions. We look at an example in this blog, but with no DataSphere install required to run the Data Profiler, it’s easy for you to see how much DataSphere can save your company. Just drop us a line at deepdive@primarydata.com to schedule a meeting.

Tiering Data by Access Time

As explained above, the Data Profiler can operate independently as a standalone application or within Primary Data’s DataSphere platform. When run as a standalone app, the Data Profiler passively reads metadata from the target(s) to model. Importantly, this process only reads file metadata; it does not touch the clients or the actual data.

A common way to tier data in DataSphere is based on the last time data was accessed (atime), changed (ctime), or modified (mtime). For our example scenario, let’s start with a high-performance NAS system with 1.7PB of data on it. If storage for this system costs $5/GB, the total system cost is about $8.5 million. Looking at our available resources, we find we have two lower cost storage systems: an object storage system that costs .50/GB and a low-cost filer that costs $3/GB.

The Data Profiler reads the system’s metadata to determine when each file was last accessed, changed, or modified. Next, we add the two lower-cost tiers, along with their storage costs. Finally, we set up our policies using a simple slider bar. We tell the Data Profiler that we want to keep all data that’s been accessed within the last two weeks on the production system, all data that’s been accessed between two weeks and one year on the filer, and all data older than one year on the object storage system.

The Data Profiler calculates the savings from the cost of the original production system, minus the difference between the reclaimed production capacity and the cost of the lower tier storage and the cost of the DataSphere license. The formula for savings is determined by noting the baseline system cost, then subtracting (reclaimed production capacity minus costs of lower cost storage) and the cost of the DataSphere license.

For our example scenario, let’s assume that data use is extremely high and that 50% of data has been accessed in the last week, and that the other 50% is evenly split between being accessed within 2 weeks and one year and longer than one year ago. Note that for most environments, this scenario is quite conservative, as IBM found: “approximately 75 percent of the data stored is typically inactive, rarely accessed by any user, process or application. An estimated 90 percent of all data access requests are serviced by new data—usually data that is less than a year old.”

The screen shot below illustrates that tiering with these parameters to storage at the specified costs would result in savings of over $2.6 million dollars!


Figure 1. Adding object and capacity NAS tiers to a 1.7PB system can easily deliver $2.6 million dollars in savings!

Tiering can be used to add performance, as well as lower costs. The Data Profiler can also model costs for adding performance tiers to serve applications and data that need serious speed. If we add an all flash array tier at $8/GB to our example for data accessed in the last hour, we still achieve savings of $1.8 million dollars.


Figure 2. An efficiently used $800,000 flash tier improves service levels, while still saving $1.8 million dollars.

If you’d like to walk through each step of this example, watch our demonstration video that shows the Data Profiler in action. To learn how much tiering with DataSphere can save your enterprise, with your unique tiering parameters and storage costs, connect with us at deepdive@primarydata.com to schedule a meeting or demo.



Contact Form

Channel Partner