ra2 studio - Fotolia
Before we can understand SD-WAN analytics, we first need to explain software-defined WAN. Generally speaking, SD-WAN is a WAN architecture that has gained significant traction within enterprise IT over the past few years. The framework intelligently optimizes the forwarding of mission-critical packets across two or more pathways, such as MPLS or the public internet, to the same destination on the other side.
It accomplishes this by continuously monitoring each path to the final destination, based on factors such as packet drops, latency, jitter and round-trip times. At the same time, packets are identified and classified based on what network administrators have configured to be time-sensitive or mission-critical data. The SD-WAN will then forward packets deemed most critical over the link identified as most optimal, while shifting lower-class packets to traverse suboptimal links.
SD-WAN analytics takes the raw network data generated by networking components and combs through it to deduce how well the network is performing, among other factors. For the most part, analysis and conclusions are formed with AI, which can then be used to monitor, troubleshoot and improve performance across a WAN.
How SD-WAN tools get their information
SD-WAN analytics taps into a variety of sources to obtain performance data, among them Simple Network Management Protocol metrics, logs, NetFlow/IP Flow Information Export and DNS data. More modern network analytics platforms -- including most enterprise-grade SD-WAN products -- also collect health probe statistics and streaming telemetry data. Both are real-time sources of granular data that are used to provide a far better view of what's actually happening on the WAN from a performance perspective.
In some cases, an SD-WAN platform can use the analyzed data to automatically adjust the WAN to improve application performance. In other cases, it may be up to a human administrator to manually deploy the changes that must be made after examining the results of the analysis. In either situation, SD-WAN analytics is a beneficial tool administrators can employ to optimize end-user performance across expensive WAN connections.
Dig Deeper on Software-defined WAN (SD-WAN)
Related Q&A from Andrew Froehlich
While VLANs are a type of subnet, they have their own unique capabilities and characteristics that differentiate them from subnets. However, the OSI ... Continue Reading
Broadband and Wi-Fi are common technical terms often used interchangeably -- although they shouldn't be. Both provide connectivity, but they do so in... Continue Reading
Migrating to UCaaS doesn't always mean an organization can keep its PSTN or SIP carrier. But some providers are taking a 'bring you own carrier' ... Continue Reading