Pace “Detecting Network Problems Before Users Do” by Stefano Gridelli of NetBeez, 13 June 2017
In a perfect world, water networks have 100% uptime and adequate pressure. In reality, customers experience loss of pressure, water quality issues, and outages. In under-invested networks, this is even more frequent.
Yet, providing a good customer experience is possible. Detecting problems before customers do is not a fortune teller’s trick, but something that can be achieved with the right tools and processes in place. So how can Network Engineers achieve Zen for their networks?
Here are three main points that are critical to ensure a great service to your network users:
Detect performance issues before they turn into outages
A monitoring solution that actively verifies network performance can quickly alert if degradations occur. This system continuously compares the monthly historical average of pressure and flow, which can be taken as baseline of network performance, to the 15 minute or 1 hour value. As soon as the short term average deviates from the baseline, an alert is triggered.
For example, the following graph clearly demonstrates how a pressure surge can be detected before it leads to a complete network outage at a remote location:
Validate network configuration changes
Many network changes are performed outside of regular business hours to reduce the risk of affecting customers. Yet many network outages are still caused by poorly planned configuration changes. The typical scenario is where the Network Engineer applies a change at 2 a.m. At 8 a.m., when the first users start flushing toilets and running showers, they find they don’t work because of the network change that was made a few hours before.
A way to avoid these unneeded outages, is by properly validating network changes. This requires, for example, making sure that all remote locations are performing as they did before the change was made. This task can be easily done with a lift and shift logger located at remote locations. However, this manual check is not easy to implement in a network with dozens, hundreds, or thousands of remote locations. It can only be accomplished with a permanently deployed set of loggers and a network monitoring and testing solution.
Use higher resolution logging
Higher resolution logging from permanently deployed standard loggers enables you to quickly detect transients. A logger that averages across 15 minutes can’t get the same results as one that polls at 10Hz. The sooner a transient is detected, the sooner it can be troubleshooted and fixed.
I hope this short article provided some useful tips on how to provide an excellent customer experience, detecting problems quickly, and reducing service downtime.
i2O goes live today with its new iNet solution. One of the key areas of functionality is detecting network problems as early as possible, and long before they affect customers. Thereby reducing costs, avoiding penalties, and maintaining service and reputation.