Network Analytics has always been a challenge of many folds for telecom companies. From interworking across network elements from different vendors to analyzing the sheer volume of data generated, there isn't a single telecom vendor offering a concrete solution that caters to the entire spectrum of telecom infrastructure. While Operators have dozens of options for every core network nodes, there isn't a single analytics vendor that boosts of a pre-integrated, out-of-box solution for analytics. At present, most of this market is catered by solution integration and consulting vendors.
Most of these challenges can be categorized into four broad areas - Multi-vendor environment, Lack of Standardization, Limited Usability (localized statistics) and Absence of Data Monetization models.
It's fairly common for an Operator to have multiple vendors that together enable the end to end telecom infrastructure. Sometimes this complexity gets magnified with having multiple vendors providing same network elements in different OpCos. Such complexity is possible because of the high degree of standardization in Telecom. Unfortunately, 3GPP has always considered OAM topics outside its scope and specific to individual operator needs. This decision has triggered most vendors to couple an OAM NE as a sidecar (with the core network element) to avoid additional integration points. With all the KPIs, logs and statistics, a mesh of OAM infrastructure is fairly limited in providing an end to end network view.
Then we have, millions of TBs of call and event records, whose usability is limited to regulatory compliance or troubleshooting. Operator's vision with records has been limited to identifying the rouge network element. This has severely limited the attention analytics has received from vendors as well. Most vendors have localized the logging and statistics that focuses more on successful call handling and focuses little on meta-contents of the call itself. The localization is so severe that no vendor in the industry uses these KPIs even to enforce licensing and mostly opt for hardware limitations to trigger reordering or network expansions. Finally, there's a lack of monetization model that has only further weakened the case.
With 5G, some of these issues are finally addressed.
There's a Network Data Analytics Function (NWDAF) that is the central entity that standardizes record collection. Service-based Architecture proposed in 5G brings the telecom industry challenge far closer to the internet world and enables NWDAF to offer these insights to policy function and other internet applications. With most phone apps already taking consent from end users to use metadata information such as location to improve customer service, telecom network can follow the path and hugely benefit by effectively utilizing network analytics. Advance topics like predictive analysis and real-time machine learning can finally give the push the industry needs.
Catenna proposes an innovative analytics infrastructure based on open-source technologies that are not limited to network health or customer experience but sets the stage for Operator's next generation Data Monetization Platform.
#AutonomousNetwork blog series addresses Catenna's proposal for an innovative analytics infrastructure. We look forward to your comments and inputs.