Modern data centers with a large number of physical and virtual servers are complex, and computers are fast, processing hundreds of millions of instructions per second. There are millions of data center metrics per second, too. Making use of all that data is difficult, but big data analytics is now helping. Gartner estimates that IT Operations Analytics (ITOA) will be used by 15 percent of companies in 2017. Vendors that help their customers understand the value of ITOA will find additional opportunities to sell to them.
A New View into Data Center Data
Companies have been trying to understand what's going on in their data centers for a while. ITOA is a kind of IT Operations Management software, which was a $21 billion market in 2014. Gartner expects ITOA applications will be 20% of that market in 2018.
Understanding and managing data center operations previously relied on tools like Data Center Infrastructure Management software and other stand-alone tools that presented information in dashboards but couldn't bring together information from different sources and didn't provide complete information. The configuration databases they relied on were frequently outdated. The data presented could only show operations staff what had happened, leaving them in a reactive mode. In an effort to keep staff aware of events, alerts are triggered frequently—often too frequently, meaning no one pays any attention to them.
ITOA changes the way that operations staff views the data center by monitoring high volumes of data in real time. Input from multiple sources can be correlated along with historical data. This means the analytic processes can identify problems that are developing, before they've occurred. The information also helps speed responses to incidents that do happen.
Predictive Analytics
Companies look for a competitive advantage through analytics, largely through its predictive capabilities. ITOA analytics helps companies achieve an advantage through its ability to predict both demands on IT and the service levels that IT can achieve. Companies can make use of ITOA and understand the data center's ability to meet end-user needs through analysis that can answer questions such as “How will the data center handle a high-load day like Black Friday?” By using predictive analytics this way, data center operations can be managed around genuine business needs.
Capacity Planning Analytics
Predictive analytics that identify the future trend of system behavior provide value for capacity planning. Combining predictive analytics with analysis of different scenarios lets operations managers ensure there's adequate capacity. This modeling and analysis can address changes such as virtualization and consolidation. The analytics can also help with understanding the impact of technical changes, such as the impact of a new virtual machine or physical hardware added into the data center. And they can help assess impacts on security, performance and other areas of concern.
Performance Management Analytics
Whatever capacity is installed and brought online, data center operations staff needs to monitor it and make sure it is operating normally. Traditional monitoring applications detect and report events that have occurred. With new analytics methods, deviations from normal behavior can be detected and reported before they impact performance. The "normal behavior" and "deviation" aren't defined through thresholds, but through analytic methods that identify the baseline and project when acceptable levels will not be met. This early warning allows operations staff to respond to issues before end users experience any degradation of service. Warnings can be correlated to specific components, applications or processes, allowing issues to be prioritized and responded to based on business criticality.
Root Cause Analytics
Because ITOA pulls together data from numerous sources, analytics can correlate that input with an operations issue. By knowing which components—database, server, or network—are associated with a service, the analytics can help staff identify and respond to the cause of any system alert.
Analytics Helps Operations Staff Gain Control
Managing a modern data center means managing a combination of physical and virtual appliances, including non-local devices in the cloud. The challenge is made harder because modern application development favors agile, DevOps methods that implement changes frequently; continuous deployment can mean changes are implemented multiple times per day. ITOA helps operations staff monitor the changes and assess their impact.
Data Center Operations Analytics Is About the Business
The ultimate benefit of analytics in the data center isn't improving data center operations; rather, it's aligning data center operations with business needs. Keep that in mind when talking about data center analytics with business customers so you can help them understand the value to the business.