Ways that Network Performance Management Has Evolved

As data center and network infrastructures grow increasingly complex, network monitoring has had to adapt. Network performance management (NPM) and application performance management (APM) have evolved from simple CPU memory, storage, and network packet monitoring. SDxCentral’s recent report “2017 Next-Gen Software-Defined Infrastructure Assurance” discusses the ways in which APM and NPM have evolved to meet the needs of these new networks.

One of the primary changes highlighted in the report is the introduction of big data, analytics, and the cloud in APM and NPM.

According to the report, “One of the biggest challenges we’ve had in performance management is maintaining historical data for trending and prediction.” Raw data from data centers and networks was able to be stored for only one day or week. Therefore capacity planning was not entirely reliable, and these aggregated approaches would miss certain micro-bursts of traffic that cause increased latency, packet loss, and more problems for the network.

Now, as network traffic continues to grow — and bandwidth use also grows exponentially — APM and NPM are leveraging elastic cloud resources and analytics tools to address these challenges. The APM and NPM technologies allow data to be constantly streamed up into the cloud and sort through the data for monitoring and management.

In the past, the use of raw data meant that typical performance management data was provided in large amounts for each managed object. End-users received only device and infrastructure metrics. And APM and NPM products each presented data differently. But with analytics, big data platforms, and the cloud, APM and NPM vendors can, according to the SDxCentral report, “abstract the underlying details and present application-, service-, and even business critical KPIs.”

As this data becomes more accessible, APM and NPM platforms are becoming more programmable. Thus, end-users can adapt the platforms to their specific needs.

Another related element of change is closed-loop automation. Companies are looking to automate the way that their network infrastructures scale and heal, creating closed-loop systems. As such, NPM and APM vendors need to provide easily-consumable data. Then, applying machine learning and artificial intelligence in a closed-loop system can enable the network to automatically mitigate performance issues and provide insights.

The emergence of converged NPM and APM platforms has also led to a number of changes. The convergence itself helps achieve greater visibility into all areas of the network by combining disparate data across domains. Previously, enterprises had to deploy and maintain multiple APM and NPM systems with specific focuses. The consolidation has also reduced the need for a large quantity of domain-specific tools.