network

 

Crossroad increased the customers’ pro-activeness by avoiding service disruptions, reducing troubleshooting time and significantly improving the time to restore service in case of incidents.


The Technology

Unomaly continuously consumes unstructured, raw and high volume event and log data – resulting in profiles that enable real-time awareness of changes and unknowns. We analyze that data to deliver insights into the known and unknown.

Challenge

Companies rely more and more on internet connectivity to conduct their business. Our client provides and manages these connections, whether it’s about a simple customer website up to providing bandwidth to support the automatization of complete worldwide processes. To improve their service, they wanted to go a step further by ensuring that problems in their infrastructure are spotted even before they become incidents, hence before customers have business impact.

Approach

To tackle this service innovation a new combination of methodology (anomaly detection) and technology (Unomaly) was used. The complete infrastructure  was monitored for 8 weeks to create a baseline using data analytics and machine learning. After this period, it kept on monitoring the system. When it spots something out-of-the-ordinary it reports a request for action to the support staff.

Results

Although the support staff was a bit hesitant to use the new system at first, it proved itself within a couple of weeks after the baseline was set. Solving a priority one incident was taking a long time because the cause seemed to be routed deep in the system. Taking a look at the Unomaly view of the infrastructure the unexpected root cause was found within three clicks.

Ever since, the system was able to predict even more potential problems and risks. Best case scenario this means a problem is solved in an early stage, worst case the impact is reduced because problem fixing is started earlier.

Crossroad made it possible for the client to provide a higher uptime of their hosted services.