Need to improve network efficiency, optimize operations, and deliver a better customer experience? Look deeper into the performance data generated by your network, services, and applications.
Hidden in the performance data generated by your network, services, and applications is the fuel you need to improve network efficiency, optimize operations, and deliver a better customer experience. There is untapped value in this data.
First, to get value from analytics, you need high-quality data. This means ensuring data is collected in a timely and reliable way. The data also must be clean, accurate, and delivered in a standard, usable format.
Active performance monitoring data is based on Two-Way Active Measurement Protocol (TWAMP), an RFC-5357 standards-based measurement that enables continuous monitoring across heterogenous networks. It has several key advantageous characteristics, summarized here:
How Accedian Skylight uses TWAMP
Accedian’s network and application performance monitoring (NAPM) solution, Skylight, goes beyond TWAMP standard metrics—basic delay, packet loss, and jitter—to provide a rich set of KPIs and quality of service measurements. These include one-way metrics, packet re-ordering, duplication, packet loss bursts, IP ToS reporting, mean opinion score (MOS), etc. This granularity and precision helps identify previously invisible issues impacting performance (and user experience).
For example, with Skylight sessions can be measured not only at 5 and 10-second intervals, but all the way down to sub-second intervals to detect micro-impairments. To make troubleshooting and root-cause analysis a lot more effective, sampling frequency can be increased to uncover issues, going from a few packets per second to hundreds.
But, data alone only reveals part of the picture.
Metadata extracts value from performance data sessions
Here’s a scenario that probably sounds familiar: you set up TWAMP sessions to collect precise, granular KPIs on your network performance. You notice that these measurements show packet loss and/or excessive delay and jitter. What do you do with this information? How do you start drawing conclusions about how your network is performing? How can you tell what is working well and what’s not? How do you make the right decisions to improve performance?
First, organize the information in a meaningful way to analyze it. Doing this in real-time makes it possible to understand what is impacting QoS. Looking at historical data is a good way to understand the root cause of recurring issues. Using metadata—a set of data that describes and gives information about performance monitoring sessions—makes this very effective.
Metadata are pieces of information you can add or modify dynamically and attach to performance data. Some possibilities:
- Class of service (voice, data, signaling)
- Region (South, North, East, West)
- Type of link sessions are using (leased line, microwave, fiber)
- Technology (4G, 3G, DSLAM, Ethernet)
- Topology information, such as the traversed routers
This metadata can be dynamically tagged to a TWAMP session by importing it via APIs into the Skylight performance analytics platform. You can also set up automated processes to collect new metadata as it changes. Reporting can benefit from this as you can create different dashboards to show performance by region or class of service—voice over LTE (VoLTE), for example. Skylight’s web interface is flexible, with customized reporting and visualization to support a large number of users and workflows.
Once you have set up metadata tags, it is easy to start analyzing and filtering data to find correlations and reach conclusions.
Example: use metadata tags to detect mobile backhaul network performance issues
Step 1: You notice delay and packet loss on specific mobile backhaul downlinks, and want to first look at whether peak delay is affecting all sessions or only some classes of service. Using a CoS metadata tag, you can look at voice and data sessions separately and determine if delay is impacting a certain type of traffic or all sessions.
Step 2: Next, you want to determine which regions are affected affected by packet loss. Using regional metadata tags, you can see different regions (in the example below, these are areas in Spain) and how delay and packet loss differs by region. You could then filter further by VoLTE vs. data sessions. Or you could look at which cities are experiencing the worst delays for voice and or data. It is also possible to add customer metadata tags to determine the impact of degradation on customer experience and enterprise service level agreements (SLAs).