More devices are generating more data, making it harder to manage the networks, but IT service managers want it to be as easy and as automated as possible.
According to Venkatakrishnan, technologies such as machine learning can be applied to the available data in order to automate network management. This simultaneously improves the customer/user experience and IT productivity.
CommScope's approach uses patented technology including anomaly detection that can distinguish real issues from routine events.
For example, a wireless network at a railway station is likely to experience a large number of associations as a train arrives, and a lot of dissociations as it departs. That's normal behaviour for this environment.
While there will be broadly similar patterns in eg office or classroom environments, there will be a different ebb and flow.
CommScope's Ruckus Analytics system collects data from network switches every 90 seconds and analyses it in a few minutes in order to distinguish true anomalies from general trends, seasonal peaks and occasional spikes.
It groups anomalies into events, and events into incidents, then prioritises incidents (eg, according to the number of affected users) so that technicians deal with the most important matters first. It also applies machine learning to help determine the root cause of an incident, and recommends actions that should lead to a resolution.
Those recommended actions – which are typically configuration changes such as adjusting an access point's power level – can be executed automatically, though it is common practice to have the system gain human approval before proceeding.
The system then checks whether the actions had the desired effect, and learns from experience.
This approach helps with service validation, Venkatakrishnan says. For example, before users are allowed to connect to a new access point, the system can use the access point's virtual client to automatically check association, dissociation and other functions.
This way, network managers can be sure that each part of the network is functioning before it is made available to users.
It also allows automated checking that the network continues to meet operational standards, eg, that a users can join the network within five seconds.
Network monitoring is more about refining various settings as opposed to fixing acute problems. The system knows how a particular device should behave – for instance, a particular type of access point should be able to accommodate a known number of devices with a particular throughput and utilisation – so if these expectations aren't met it will make adkustments.
Where problems do occur, "it helps IT guys to quickly look where the issue is." he says.
Ruckus Analytics reduces the time needed to identify an issue, determine the cause, and fix the problem by 50 to 70%, according to Venkatakrishnan.
The system can integrate with IT service management tools such as ServiceNow to file help desk tickets when human action is required, and to allow service validation tests to be performed remotely.