LEGOs and modern network performance monitoring
How to make sense of the data chaos in a software-defined and cloud-centric world
Author — Tim Diep
This past weekend, while my wife and I were cleaning up our basement we came across a large pile of LEGO pieces tucked away in a corner closet. We have two boys and they are teenagers now who lost interest in LEGOs years ago. So, not surprisingly, I was given the task to organize the pile so that we can donate them. As I was enjoying my new weekend assignment, it dawned on me that what I was doing with these pieces is exactly what CA’s SDN/NFV monitoring software has been designed to do in the network.
Unsorted LEGO pieces are like the data that network performance monitoring software collects and translates into intuitive visualization and analytics. Like the disorganized LEGOs, the data in modern networks is available in a similarly disorganized way and Software Defined Networking (SDN) and Cloud have introduced even greater magnitudes of complexities and chaos to the network performance data set.
The data available from vendors and open source are unstructured and proprietary. Network monitoring software has to collect the unsorted pile of data, organize it into uniformed structures, store it, and then produce the meaningful visualizations and analytics IT managers needed to maintain operational control and accountability. Each new scheme of virtualization and abstraction adds even more piles of unstructured data sets. Each device, subnet and tenant creates their own data pile.
For example, imagine a small network with hundreds of devices that could have thousands of piles of unsorted data sets. Network performance monitoring software has to collect and unravel this mess in a timely manner so that the topology can show the updated changes, the performance dashboards can show the most current utilization, and the threshold events can be triggered promptly. This is why building reliable software to monitor modern networking is so important.
The analogy doesn’t stop here. Consider that certain LEGO pieces have to uniquely stay together because of color, shape or functions. This is similar to the relationship of attributes in modern networks that advanced monitoring software has to extract and tag relationship attributes on every single piece of data.
Finally, one aspect of networking data that LEGOs are not equivalents to, is dynamicity. Modern network data is constantly morphing because SDN and Cloud networks are continually being changed due to self-service and automation. In the industry today, dynamicity, combined with relationship attributes and the mountains of unsorted network data, will drive the necessity to develop innovative network monitoring software.
Can you relate these network monitoring challenges? Check out this recent report on strategies for successful modern network management in the age of Cloud and SDN.