Business support systems (BSS)—specifically, charging and policy functions—hold the key to monetizing 5G investments for telecommunications companies and their partners.
However, for this key to unlock new revenue for communication service providers (CSPs), there are a couple of things that need to happen:
- BSS applications need to evolve to enable the agile creation and launch of new products, and,
- In this evolution, these applications need to change the way they handle data.
Why am I talking about this?
At Volt Active Data, where I drive product vision to meet the ever-more-stringent performance and latency SLAs of 5G telecom data usage, we’re all about making data work for telcos rather than against them.
In other words, we’re all about the second item above: changing how applications use data.
Must Read: What Is Event Driven Architecture
To support game-changing BSS applications in the age of 5G, there are certain features your data platform must have to allow your applications to use data in a way that optimizes 5G monetization, while preventing fraud and other revenue leakage.
1. Immediate consistency
Immediate consistency means your data and the decisions it drives are accurate and correct at all times.
Quite often, solution providers and software vendors turn to point solutions that are perhaps open-source but compromise on correctness to meet scale.
This is not necessary. There are technologies that can scale and also guarantee immediate data consistency.
There are two primary benefits of immediate consistency:
- Data accuracy: When it comes to charging, accuracy is quintessential. Your data platform must be able to make usage-based changes to account balances in real time to ensure subscribers’ accounts correctly reflect the current state taking into account the latest activity. Without accuracy, CSPs end up approximating, which usually means rendering services the customers haven’t paid for or not rendering services the customers have paid for.
- No data loss: When a hardware or software failure happens, immediate consistency ensures you do not lose data while eventually consistent systems end up losing data. In systems like Charging (CHF) and Account Balance Management (ABMF), losing data is simply not acceptable since it involves subscribers’ money and account balance.
Must Read: Best Data Science Platform
With 5G’s service-based architecture and the requirement for low-footprint environments, such as an edge environment where infrastructure availability is limited, the data platform must be cloud-native.
And I’m not just talking about being able to start and stop clusters, but all operations. It’s vital to keep the cost of operations—measured in money, personnel, and time—very low. This means using full automation to save precious time. Thanks to technologies like Kubernetes, enterprises can now easily manage cluster availability with far less human intervention than if they were using traditional data centers.
One can find a plethora of options on the CNCF cloud interactive landscape chart, but a CSP software solution builder should ensure its data platform’s “cloud native-ness” does indeed meet the rigors of a 5G telecom deployment environment
The data platform must be able to perform and scale to meet the needs of multiple network functions so that there is one consistent truth across these functions.
You must avoid monolithic solutions at all costs and go with distributed technology platforms that are cloud-native to avoid the headaches of scale management creeping into application logic (think of application-driven data sharding).
To ensure maximum ROI, your data platform must be able to use as much of the hardware resources available and allocated to it – including CPU, memory, and network bandwidth—applying them to add real business value rather than just architectural overhead.
Must Read: Bss Evolving to Support 5g
4. Low latency
The platform must provide the lowest latency path possible from data ingestion to data-driven intelligent action, which could take the form, for example, fraud prevention for revenue assurance.
Low latency is meaningless if you have several low-latency layers coming together to form a high-latency assembly.
Consider the steps involved in an intelligent, event-driven business decision (ie, any time a customer interacts with an application or a system as part of a process).
For every event, your data platform must:
- Ingest the data
- Store the data
- Aggregate the data for the sake of a KPI (which could be something as simple as balance usage to something as complex as a subscriber’s buying propensity score or even a fraudulence score)
- Measure this KPI against what is considered the normal (i.e. when all is well)
- Detect any deviations and run through a rule set to determine the appropriate action
- Take the appropriate action
To truly deliver low-latency business automation, your data platform must be able to perform all of this with low latency—ie, within 10 milliseconds.
A less ambitious system could potentially achieve this process by assembling together a variety of technologies such as an ingestion layer, a storage layer, a processing layer and a ruleset manager, but at what cost and what loss of latency?
CSPs vying to lead the way in the age of 5G need a unified platform that can address the entire data lifecycle in one quick, seamless motion. The infrastructure savings alone justify staying away from the so-called “free” open source solutions, and that’s not even taking into consideration personnel costs or the price of getting to market late.
Finally, the high performance and low latency must be consistent and predictable to ensure SLAs are not compromised.
5G-based applications are going to benefit immensely from both human subscribers and industrial IoT (i.e. machine subscribers). One of them grows at the rate of human population growth and the other grows at the rate of enterprise innovation.
Your data platform must be able to scale at the pace of IoT while still being able to meet the high performance demands of enterprises without any degradation in latency SLAs.
Volt Active Data was designed specifically to do this.
The architectural choices we’ve made ensure low latency and scalability with high predictability. In other words, we allow CSPs to build real-time applications that take real-time data-driven decisions and actions without compromising on data accuracy.
If you would like to learn more about Volt Active Data and how we can help your solution meet your CSP customers’ demands, please reach out to schedule a demo.