In today’s world, ISPs and broadband operators are confronted with an increasing number of customers with connectivity issues exacerbated by the likewise increasing number of wireless devices in homes and the greater demand for bandwidth. Real-time analytics platforms, like AXON Predict, improve ISP and broadband operator’s abilities to respond to and conquer connectivity issues.
Service provider device management systems like TR-069 platforms are designed to authenticate, provision, and manage connected customer devices, but they can’t meet this challenge alone. Such platforms cannot provide real-time and/or detailed customer premises equipment (CPE) information, nor can they report on third-party or unmanaged devices. This impairs the operator’s ability to gain insight into the source of connectivity problems and often results in higher call volume, longer support interactions, and more truck rolls (>$100/per issue).
The AXON Predict analytics platform helps operators improve their services through detailed, individualized, and global device insight and anomaly detection. Utilizing the Atom edge analytics engine, AXON Predict users can remotely resolve issues in a fraction of the time it takes with traditional techniques.
AXON Predict manages and monitors distributed Atom analytics engines. This engine continuously analyzes, learns, and compares the current behaviors and performance of any CPE device to its past; and can be scaled from a single device to thousands of devices out in the field. It identifies potential issues or behaviors and, where possible, takes corrective action automatically, thereby eliminating the need for support calls and truck rolls.
How AXON Predict Works
Let’s run through an ISP customer support example for an imaginary operator to illustrate AXON Predict in action.
The biggest driver for support calls to service providers are Wi-Fi-related issues. When a customer calls a service provider, the complaint tends to be one of the following:
● “I can’t connect to the Internet”
● “My Internet is slow”
If there is no WAN connection, a TR-069 platform can verify the issue. But if the WAN connection is present, there are either device- or network-related issues. This is where AXON Predict shines.
When a customer makes a support call, the support technician locates the customer account information through their existing CRM system (AXON Predict directly ties into CRM systems and presents the main “report card” page). This highlights the real-time nature of AXON Predict by providing immediate visualization of the status and real-time data tracked and logged from the gateway. As we know, there are different requirements for Tier 1 support versus Tier2/3 support. The parameters in a report card page are therefore fully configurable and customized to each user, and views specific to target user groups can be saved as preset dashboards.
In this example, the real-time data being tracked includes temperature, memory usage, and CPU usage for the gateway.
A Gateway Performance panel will show gateway static parameters, including model number and firmware version, as well as dynamic parameters, such as number and types of connected devices. It also has a Performance Issues detected in the last two hours button, which links to the gateway’s Health and Performance dashboard.
A Connected Devices panel displays details for the devices connected to the gateway. It shows all the devices and RSSI signals logged in various time increments. A WiFi Anomalies and Issues detected in the last two hours button links to a WIFI Connectivity Anomaly Board.
The report card page provides a quick snapshot of the network quality, gateway health, device(s) status, and—most importantly—performance and Wi-Fi connectivity issues detected and logged by AXON Predict. The support technician can easily view the current and historical conditions, as well as any detected anomalies, to quickly get to the root cause.
Let’s dig deeper.
A Tier 2/3 technician would click a “Performance Issues…” button on the report card page to view the gateway’s Health and Performance dashboard.
A graph will display all performance issues were logged while monitoring the Utilization and Temperature of the Gateway, and another graph will show the magnified version of each performance issue and the parameter values configured to trigger it. Graphs can be panned and zoomed to provide quick, side-by-side comparisons of performance for different times of day, such as the segment of time when an anomaly is detected compared to a longer period. The platform’s flexibility enables a technician to monitor any parameters, associated patterns, or thresholds configured to trigger alerts or alarms.
In our example, say that a temperature spike has been detected (it would be visually highlighted in orange on the performance graph). This would prompt customer support to provide a temperature-related solution (determining if the device is enclosed, stacked, or covered with other household items that are preventing air circulation and cooling).
A Smart Sight chart located below the graphs will dynamically update based on the “Learned Averages” and “Current” time windows selected, and calculates a comparison of the metrics monitored within both windows (“Difference”), providing instant analysis of performance deviation.
A “WiFi Anomalies…” button on the report card page will open a WIFI Connectivity Anomaly Board for an overview of all Wi-Fi issues AXON Predict’s Atom engine is detecting across all connected devices, as well as details for each issue.
Using the displayed information, a support technician can determine when and how often an anomaly is detected and the actions needed to resolve it (such as sending notifications to operations engineers or rebooting the gateway when anomalies are detected). A notification will contain a URL link to the related dashboard, quickly enabling the operations engineer to easily analyze and troubleshoot the detected anomaly.
More Support for Support
AXON Predict essentially provides support for ISP and broadband support operations, enabling technicians to quickly identify the root causes for performance and connectivity issues detected via the Atom analytics engine. Future support capability examples include additional action options, such as restarting specific interfaces, throttling devices that are hogging bandwidth, and so forth. The eventual goal with AXON Predict is to let the platform conquer most connectivity issues—it will detect trouble and autonomously take self-healing actions with minimal service provider involvement.