Fraud Prevention in the 5G Era

Detecting & Blocking Fraudulent Transactions with Intelligent Operational Insights

Fraud has been an ongoing problem for Communication Service Providers (CSPs) and their customers for decades, with a ever evolving cat and mouse game being played out between CSPs and fraudsters. Unfortunately telecom fraud claims billions of innocent victims across the globe. The cost of fraud to CSPs in 2017 was nearly $30 Billion. What is even more concerning however is that with 5G; the magnitude of the fraud problem will rise hundredfold. The surge in the amount of connected devices and sensors will lead to an exponential rise not just in the volume and velocity of data streaming in, but also significantly increase the level of complexity associated with processing the data to identify and prevent scams. Traditional fraud prevention techniques implemented by CSPs will just not be viable anymore.

CSPs have to constantly contend with hundreds of distinct fraudulent schemes; the ever changing schemes, along with their high frequency, globally dispersed origins, and multiple layers of machine and network generated anonymity makes these schemes extremely hard to detect and prevent.

It is important to note that when it comes to fraud, in-event detection and prevention is far more desirable than post-event reconciliation. Once the fraud occurs charges are levied either on the CSP or the subscriber and revenue is lost. At 5G scale and speed however, in-event fraud detection and prevention is easier said than done.

New Considerations for the 5G Era

So how can CSPs effectively manage the sheer volume, velocity, variety and complexity of fraud threats in the 5G era? They need to:

Analyze data available from multiple channels

Fraudsters try to veil attacks with complexity. Traditional fraud analysis involved analyzing CDR records, CDRs however lack crucial information and are generated after the call is complete (when it is too late to stop fraud). Signalling System No. 7 (SS7) however contains additional information that may be crucial to fraud prevention such as: subscriber location, inter switch call attempts, request for subscriber information, and more. External radio probes or automated subscribers can also be used in conjunction with SS7 to detect fraud.

Automatically monitor and prevent of fraudulent transactions

In addition to monitoring and identifying fraud correctly (without blocking legitimate subscriber activity) it is crucial to block the fraudulent calls / texts, etc before they are placed. The database needs to be intelligent and agile enough to detect changing fraud threats and counter threats.

Implement real-time Intelligence (Machine Learning) & Complex Event Processing

To block fraudulent transactions and users, the underlying database needs to analyze thousands of attributes such as subscriber behavior, geolocation, device information, transaction type, etc in real-time. Using built-in machine learning algorithms these attributes are compared to correct behavior and statistical anomalies are identified and blocked in real-time / in-event, prior to the transaction occurring.

All of the above at 5G scale

5G is expected to deliver microservices such as IoT at massive scale. The database engine must be able ingest data from multiple channels and automatically apply thousands of rules with built-in machine learning in milliseconds to prevent fraud, while processing millions of operations/second.

The VoltDB Database

VoltDB is the only database that offers all the functionality required to fulfill all of 5G’s data processing requirements. VoltDB can seamlessly:

• Scales to millions of events per second.
• Ingest and analyze fast streaming data for real-time decision making based on complex Machine Learning rules, while simultaneously querying historical data to continuously improve the model.
• Import complex models built by data scientists through PMML, PFA, and other standards ensuring quick implementation and consistency.
• Real-time queries at high throughput with an in-memory, distributed architecture.
• Seamless deployment on large-scale distributed cloud environments.
• Core based data and processing distribution.
• Orchestrate VoltDB with Kubernetes; to easily spin-up new instances and scale-out dynamically, and monitor running clusters.
• Guaranteed ACID compliance, ensures the consistency and integrity of your data.
• Complex event processing, to seamlessly normalize data streams in mediation.
• Unparalleled high performance with sub millisecond response times, enabled by a distributed scale out architecture. Fast edge analytics supports the stringent QoS metrics dictated by 5G. Several 3rd party TAS applications can run concurrently without any performance degradation.
• Dynamic service delivery leading to high agility and reduced TCO. VoltDB is cloud native, so real-time applications can run smoothly in the cloud.
• Resiliency, high availability, active active disaster recovery are fundamental features of VoltDB, that are not compromised for applications running in the cloud.
• Proven security technology such as TLS & Kerberos built-in, along with auditing and monitoring tools to verify usage, and identify attacks on system.