In part one of our blog series we discussed real-time monitoring’s position in the data center management chain, while part two explained how to use operational data to dramatically drive down asset TCO through effective capacity planning and strategic asset management.

In today’s concluding blog, we explain how predictive analysis can leverage the above even further to raise a business’ competitive advantage while delivering significant savings.

As previously discussed, real-time monitoring, asset management and capacity planning offer substantial benefits in themselves, but the true potential of these technologies lies in the insights that can be gleaned through detailed analysis of the data collected.

Machine learning technologies can understand and interpret the complex connections between millions of pieces of data collected over time in a way that humans cannot grasp. Within the data center this allows an operator to create automated policies that integrate into management software for more autonomous, optimized operations. This decreases the risk of human error, improves productivity and creates a more strategic data center.

Essentially, predictive analytics moves data center operations from a reactive to a proactive mode.

The business outcomes are particularly compelling when scaled globally. If an organization has implemented real-time capacity planning and monitoring capabilities across the enterprise, predictive analysis could facilitate shifting workloads between facilities on an intermittent basis.

The ability to dynamically shift loads safely between data centers not only defers capital expenditures in upgrades or new facilities, but also allows the delivery of other efficiency initiatives like on-site renewable power generation, “follow the moon” strategies, or the participation in utility demand-response programs that offer significant financial rebates.

Focusing on a specific data set like longitudinal insight demonstrates further advantages. Longitudinal data gathered through monitoring can be analyzed to identify trends as well as predict issues. For example:

• If CRAC units have a characteristic pattern of declining performance before they fail, the enterprise knows exactly when to replace a unit that’s following the same pattern.

• Applications could have a particular cooling and power use signature under one set of circumstances and another when conditions change. Predictive analytics allows enterprises to understand the relationships among environmental conditions, work patterns and compute activity over time, allowing them to derive rules and develop models that assist in decision-making.

• Another benefit is the more accurate measurement of the true cost of providing a service. If a business can identify precisely what application is running on a given server at a specific time, track the power and cooling used to support that compute load, and incorporate data on other resources required (e.g., staff, bandwidth, additional equipment), it can calculate precisely what to charge a specific business unit or client for that service. This is particularly valuable for colocation providers and managed services organizations searching for new value-added services and revenue streams.

The above is transformative in terms of business planning. Workloads can be prioritized and costs allocated based on demand and availability. Organizations can develop industry- specific productivity metrics that give users visibility into the efficiency of their infrastructure and allow them to fully understand and exploit the capacity of the data center to further business growth.

One such example is eBay’s Digital Service Efficiency (DSE) dashboard. eBay’s DSE quantifies data center effectiveness in the context of business key performance indicators to help the company track its progress in four target areas: performance, cost, environmental impact, and revenue. Metrics such as these highlight the importance of data center efficiency in meeting business goals.

The Keystone of the Modern Business

In our always-on world, customers expect a data center to be continuously available and companies invest millions of dollars each year to meet those expectations. Unfortunately, inefficiencies mean that data center availability often comes at an unnecessarily high price, both in terms of capital and operational costs, and environmental impact.

To succeed in today’s competitive business climate, enterprises must implement techniques and technologies that enable them to maintain continuous availability while optimizing efficiency. The secret to achieving both goals is real-time monitoring and management of the data center environment, and the predictive analysis of the data collected to enable a more integrated, autonomous data center which informs decision making throughout the organization.

A company armed with this combination of information and technology is able to fully understand and exploit the capacity of its data center to further business growth.