Blog | AppLogic Networks

Telcos have no lack of data. OSS data, BSS data, diagnostics data, speed test data, etc. But having all this data doesn’t mean that every different telco department has the data they need to answer all their questions or solve all their problems. Often the missing link is not network-based data, which is readily available from a number of different sources, but customer-based data, which can be hard to come by in timely fashion. That’s where AppLogic Offnet can help.

Before we dive in, let’s back up a step. One of the most commonly used data sources for understanding network traffic is NetFlow data -- metadata about IP traffic flows that can provide insights into network traffic patterns, usage, and potential issues. This is obviously useful for managing the network as a whole, but getting down to the customer level is cumbersome, to say the least. Storing actual packets, even for a sample set of subscribers, is just too unwieldy to be of practical use.

But storing flows, possibly in a sampled way, can give a reasonably good representation of the network traffic. Typically, a telco’s planning and operations teams would use this to figure out where to expand more capacity, or what new content to add to the content delivery network (CDN), or who to expand peering with.

Let’s look at what’s actually in this data:

• Subscriber identity (MSISDN or similar)
• Device Identifier (mobile only (IMEI))
• Network Location (ECI, CellId, MCC/MNC, Gateway, CMTS, BRAS, DSLAM etc)
• Start Time & End time
• Volume In, Volume Out
• Source & Destination IP – v4 or v6 addresses used in the flow
• Source & Destination Ports
• IP Protocol

Now, looking at the fields above, it’s easy to see how someone in planning organization can aggregate usage by destination IP, maybe couple it with the BGP information, and go work on capacity upgrades at the edge of the network.

Also, a customer care agent working on reconciling a bill shock case where a subscriber travelled to Bahamas and racked up a $6000 bill and “Didn’t even have roaming on!” could use this data quite effectively to sum up the volume hour-by-hour and counter that claim.

But what about more sophisticated questions, like those from Marketing or Product Management? Questions that could uncover potential revenue opportunities, like:

• What are the most popular games people play online?
• How many subscribers have more than three OTT video subscriptions?
• What percentage of the subscriber base has an Amazon Alexa or Apple Homekit? Is this more popular in cities or in rural areas?
• How much video per day does a typical user do?
• What is the fastest growing app in terms of popularity?

Could that be done with the data above? Maybe. But what would it take? You would have to take the destination IP address and try to reverse DNS and WHOIS and try to map the results somehow into what the application was. And there are thousands of popular apps, and the long tail is millions. And a lot of them are hosted on Cloudflare or AWS or GCP, so neither the WHOIS nor the DNS route will take you very far.

The reality is, you can tell the Autonomous System (AS) of the destination, and you can tell who owns the IP address, but you really don’t have a chance to map it to an application. And if you can’t map it to an application, you can’t answer any of the questions above.

And this is not true for just Marketing. Churn prevention, Product Management, Operations, Planning, Legal, and Executive Management all have questions that could and would be answered from your flow data if it had an “Application” field.

And now we’re back to where AppLogic Offnet comes in.

AppLogic Offnet is a product library that adds Application and Application Category fields to your flow data. It knows exactly what IP, server hostnames, protocol and port usage and behaviour represents a particular app and will enhance your existing data with this information, at rest or in the data pipeline.

AppLogic Offnet is highly accurate because it’s based on massive automation systems using these apps, as well as AppLogic Networks’ DPI-based deployments worldwide learning what IP, ports, server hostnames, protocols and behaviours are generated by each app, and then continually feeding this data into the AppLogic Offnet product, just like a threat feed or an anti-virus update.

Speaking of threats, another thing you can add to this data is whether the flow represents an indicator of compromise. Basically, meaning is it command and control traffic for a known threat indicating that the user device has been compromised by spyware or malware?

You can also add Lat/Long information based on what we call Geo-IP, which is an IP database with geo-positioning data.

All of these additional fields make your existing data incredibly more useful. The big data team can now field questions from a lot more departments and with a lot less effort.

These three new pieces of information – the application, the geo-position and the threats – make your network data at least three times more valuable to the business.

By answering more questions, and answering them faster and better, the organization will simply make better decisions!

And that’s how AppLogic Networks makes networks better.

To learn more about AppLogic Offnet, or to request a demo, contact us here.


And follow us on LinkedIn, Facebook, Instagram, and BlueSky.