I realize I may be crossing the lines into things like the Rating Function and the ABMF, etc. but, figure that the same concepts apply. You might want to do things that go beyond what 3GPP specs account for, but then you have to find creative ways to make them work and still ‘trick’ the 3GPP architecture into doing what you need it to do. Imagine something like allowing for account overdrawing on a prepaid balance, where the actual balance is Zero. So the PCRF wants to say ‘STOP’, but basically your billing system is going to fool the OCS into believing there’s a balance there to draw against, which is actually just an overdraw balance or limit. So, the billing system in the background knows the true state of things, but the 3GPP system never gets exposed to that (I assumed this specific capability is not built into the spec; admittedly I was not 100% certain).
This is a very interesting area and close to my heart. In general, the customer’s account can be considered partly as one or more financial or service-specific balances, updated ideally in real-time and against which thresholds may be configured by the operator or the customer themselves. Reaching these thresholds immediately triggers events such as user notifications, promotions, service limitation (e.g. Fair Usage) or payment events. This removes the hard historical distinction between postpaid and prepaid and is more consistent with the flexibility demanded by future charging models. So to come back to your question, real-time charging in combination with flexible thresholding provide the flexibility to implement pretty much any account behavior – prepaid, postpaid, hybrid and doubtless a few we didn’t think of yet.
But your question also touches on an issue central to the OCS/PCRF architecture debate. Specifically, where is the intelligence? As the volume and variety of real-time service interactions increases (notifications, QoS control, fair usage limits etc.), there needs to be a coherent model for implementation of the intelligence that controls these interactions. For example, some operators have devolved data usage counting and business intelligence for fair usage throttling totally to the PCRF. The result is that both the OCS and the PCRF now share responsibility for real-time decisions that impact customer experience. Even for simple services, this distribution of intelligence creates complex data and control synchronization challenges. But these challenges can rapidly multiply as offers become more innovative including, for example, sharing, multi-device support and hierarchies. This emergent complexity is another important consideration in the arguments for a unified OCS/PCRF architecture.
In general, I believe we are starting to see a change in the way standards are considered and applied in the BSS space. While absolutely critical for radio, access and core network domains, there is a danger that excessive focus on standards for the BSS domain can pull focus away from the business imperative of delivering solutions that create immediate and sustainable business value in a massively disrupted market. I see evidence of this change particularly in the increasing use of rapid Proof of Concepts (PoCs) early on in a tender process.
So, now I turn to the audience. Clearly this is a controversial topic. Before we dig into the battle around whether or not to unify policy and charging, I invite you – our readers –to share your thoughts (publicly here via comment or Tweet us @BillingViews, or privately via email) to sound off on these opinions. (No personal shots at Nigel though; he’s done us a service.) So,what’s your stance on the value of 3GPP specs with specific regard to policy, charging, and rating? And – does it make sense to unify policy and charging and/or to unify the OCS and PCRF on a common platform?