Analyze This: Difference-making Analytic Capabilities to Elevate the Performance and Efficiency of an SD-WAN

September 03, 2020

By Special Guest
Mike Frane, Vice President Product Management, Windstream Enterprise,

If ever there was a time for businesses to look inwardly for ways to operate more efficiently and maximize their resources, it’s now. And for many businesses, the insight they seek may well reside in the immense amounts of data accessible to them via their SD-WAN.

In a charged and highly uncertain business environment like today’s, where every shred of business insight, every operational decision, every customer interaction and every security threat seemingly carries extra weight, there’s much for enterprises to gain by taking a closer look at the data across their communications ecosystem. That includes data they capture directly from their SD-WAN’s own orchestration, management and monitoring systems, as well as data from sources and services attached to the SD-WAN, such as the core network, LAN, voice, security systems and so on. While there are systems today that provide this aggregation, they typically are used for security forensics, not to improve business operations.

When an organization can pool data from these various sources, then apply advanced analytics tools to it, that’s when the meaningful business and operational insight really begins to flow. As useful as data from an SD-WAN alone can be in optimizing the performance, resilience and security of an enterprise network, it’s only one piece of the overall puzzle. The insight becomes even more powerful when the data comes from a broader continuum of sources, and advanced tools like machine learning and artificial intelligence are applied to it.

When your SD-WAN has capabilities like these working behind the scenes, they can collectively function as an “insight engine,” yielding a consistent stream of potentially valuable information — trends, anomalies, threats, targets for efficiency improvements, etc. Today, businesses are using that engine to gain:

• Aggregated visibility into the health, status and performance of the network. This category of analytics is valuable for the high-level macro view it provides across network locations and circuits. With this aggregate view of the data, an enterprise can drill down into site-, circuit- and location-specific information, with the ability to see, in real-time, bandwidth utilization by application, as well as to identify hot spots in the network, and how the network is performing in terms of jitter, packet loss, latency and the like. All this information enables an enterprise to optimize its network to provide the best possible experience for users (employees and customers) across locations, regardless of the apps or devices they’re using.

• Aggregated visibility into applications. This category of analytics provides a detailed look at applications across the entire network, including client devices, destinations and business policies. It can offer insight into, for example, usage by application traffic priority for any location. It also can provide a real-time look at the top applications at the network level, and at individual locations, for a given period.

This information gives an enterprise valuable insight into the effectiveness of specific business policies and traffic priorities that have been enacted across its network and at specific sites.

Here’s an area where, based on what the data suggests about business policies and traffic priorities, an enterprise could configure the network with certain automated actions that automatically block an unfamiliar social media app or peer-to-peer app when it first appears on the network, or that sends network administrators an automated alert whenever a new app is identified on the network, so they can evaluate the app, then decide if they need to establish a business policy around that app. This enables an enterprise to limit the consumption of non-business-critical apps.

• Greater visibility into bandwidth spiking instances. As part of their overall application visibility, it’s important that organizations have access to enterprise-wide reporting on availability, usage and percentage of upstream and downstream traffic across all clients at each location. As part of that reporting, they need the ability to isolate bandwidth spike occurrences for investigation: When did it happen? Which apps were running when it happened? How exactly were circuits being utilized during the spike?

Analytics can provide quick, timely answers to each of these questions, speeding the troubleshooting process. It also gives the enterprise clearer insight into the impact of business policies and traffic priorities on the network. Are they having the desired effect? How could they be adjusted to optimize bandwidth utilization?

• Predictive capabilities to optimize application traffic. Here’s where artificial intelligence-powered analytics capabilities come into play, reading in between the lines of all the data to predict how the network — and its constituent applications and systems — will behave. So if, for example, the predictive analytics tool you’re using tells you to expect a short-term incident of particularly high latency, you then can activate a backup cellular connection to address the expected issue before it causes problems across the network.

As useful as analytics capabilities like these can be to enterprises with SD-WAN, it’s important that they be paired with tools that help the data to tell its story — specifically, dashboards that enable decision-makers at the enterprise level and at the network level to easily visualize and manipulate data from various perspectives and in different contexts.

The power of dashboards lies in their simplicity and flexibility. Essentially they provide a unified environment — a single pane of glass — through which to view and manage a network. The graphics a dashboard produces can make macro and granular trends in the data easy to digest and understand, even for decision-makers who aren’t data scientists. These dashboards function as the control tower for an SD-WAN network, enabling adjustable and customizable views to gain multiple coherent perspectives on disparate data, with the ability to zoom, scroll, pan and filter. These dashboards are situationally aware, pulling in the appropriate information to deal with the anomaly, issue or opportunity at hand and leaving the extraneous information out.

Beyond viewing, dashboards also can enable action based on the story the data is telling. So, for example, a network administrator could adjust a business rule that applies to a specific location, or a priority that applies to a specific application, based on usage data that’s been observed at that location.

In today’s volatile business environment, capabilities like these that enable your organization to make timely network adjustments and capture new efficiencies have never mattered more.

Mike Frane is vice president of product management, SD-WAN, for Windstream (News - Alert) Enterprise.

Edited by Maurice Nagle