How big data is challenging mobile service mediation
The traditional approach to service charging involved each network event being equal to a chargeable unit that was priced individually. But this paradigm has changed. The new approach involves tiered service bundles and periodic allowances. This change has been the result of a combination of evolving consumer behaviors and technology advances that have resulted in a rapid increase of usage data being generated by communications networks. These volumes manifest themselves as the challenge we are coming to recognize as big data.
A corollary of the market evolution noted above is a transformation of operators' business support system (BSS) environments. Service providers are having to rethink how they collect and use usage data. Processes for policy control, billing, business analytics and more depend on the same source data. But distributing the collection and counting of service usage leads is not the answer. That approach results in duplication of processes, and thus (eventually), an unsustainable situation.
As s a result, operators are growing to accept the need for a unified usage management solution to manage and control how customers use their network assets. This should both increase customer satisfaction and reduce cost by allowing the service provider to use its physical assets more effectively.
Today, consumers of mobile data services demand personalized price plans, generally based on tiered service bundles. Thus, the old approach based on individually priced network events has been replaced by periodic allowances. Enforcing policies as well as keeping subscribers informed requires the capability to count usage in a model that reflects each specific user account.
The shift in how services are charged and delivered has coincided with technological advances that allow mobile data speeds comparable to those of fixed line and cable networks. At the same time, the number of usage events produced in the service providers’ network will continue to accelerate with the current rapid uptake of mobile data services and the introduction of new types of connected devices.
The rapid volume-increase is putting a strain on existing BSS infrastructure for several reasons:
- There is a lack of cost efficiency to constantly upgrade systems capacity
- The large number of systems that require usage data from core, access and IT networks leads to duplication of processes and inconsistencies
- The complexity of having multiple point-to-point integration solutions becomes unsustainable
- Software licenses are often based on transaction volumes
Reducing the load on BSS systems is a necessity, and this can only be done if consolidation is performed early in the processing chain. But consolidation also means loss of information due to a lower level of granularity. This loss of information can be mitigated by performing usage counting before consolidation
In a mobile network, there are typically online transactions between network gateway nodes and online charging or mediation systems for charging purposes. However, these volumes only represent a fraction of the usage data from the network, which primarily comes in the shape of records that are transferred at discrete intervals. It is common for network operators to set these intervals to 10 to 15 minutes and batch process data at off-peak hours.
For end-user communication, this is not an acceptable solution. Customers demand accurate information about their usage through, for example, SMS, IVR, smartphone applications or USSD. Sending these notifications and alerts in a timely manner requires that usage data is continually collected at short intervals. With mobile data speeds of eight to 20 Mbit/s, reaching a 1GB allowance within minutes is a practical reality. The risk of exceeding allowances and thereby triggering overage charges or service limiting actions without warning has thereby increased.
Today, slow processing and long reporting intervals result in lost opportunities to drive revenue through timely offerings of migration to new tariff plans. The lack of timeliness in processing usage data constitutes a business risk, and the result is churn and overutilization.
For timely information to the end-users, much shorter intervals are required. While enabling increased customer satisfaction, this also means a significant increase in usage data volumes that are propagated in BSS and OSS systems. Therefore, usage data must be counted, filtered and consolidated early in the processing chain to reduce the number of single events that are forwarded for further post-processing procedures.
As cost reduction also becomes a major driver for a more efficient BSS environment, so does the need to decrease reliance on monolithic and complex systems. Because of this, operators need solutions based on modular and flexible components. Market forces drive down cost in a modular architecture, as each module can be individually commoditized. In a modular BSS environment, mediation combined with counting and policy rule functionality forms an ideal platform for usage management, as it enables service providers to offset the cost for handling increasing data volumes.
Billing and policy control are not the only cases for which usage management functionality is needed. Service assurance mediation includes creating a customer-centric view of the customer experience and enables effective root-cause analysis. The core network does not provide enough information to perform this; it must be consolidated with access network information to provide a complete end-to-end view of the user experience. The capacity requirements for this type of processing are radically different from billing mediation, as the number of records is counted in millions per second.
The conclusion, and the reality, is that mediation isn’t just mediation anymore. The mediation platform now triggers a chain of events that renders it much more significant than simply a data feed for billing.