|Issue:||Asia-Pacific I 2013|
|Topic:||Pipes and policy|
|Title:||Head of Chief Technology Office|
Shai Ophir is head of the CTO office at Starhome. Mr Ophir has more than 23 years of experience in telecom, technical collaborations, funding and IP rights.
Prior to joining Starhome, Mr Ophir worked at Digital Equipment Corporation (DEC) and Amdocs as a software engineer and a development manager.
Shai Ophir holds a B.SC in Mathematics (cum laude) and an M.SC in Computer Science from Clayton University, St. Louis
Smart pipes give subscribers better quality and better privacy protection. They stimulate revenue growth and give operators – with the right tools – the means to gather and analyse data, and create and enforce policy decisions in real-time. The management of LTE/4G networks in real-time will depend upon ‘Big Data Analytics’ and PCRF (Policy Charging Rule Function). PCRF will be a core network component managing subscriber policy and charging rules, directing roaming, allocating bandwidth controlling QoS parameters, and associating charges.
Piping hot, or lukewarm?
Mobile operators are aggressively pursuing smart strategies and new business models in an effort to generate new revenue growth in an ever-changing environment, while maintaining top-quality services for their subscribers.
There has been much debate as to whether mobile operators will have to become a pipe to deliver cost-effective services demanded by subscribers. And, once the plumbing is in place, should they opt for a smart or dumb pipe? Dumb pipes provide only a simple connectivity. Smart pipes, though, can help monetize broadband access in various ways to stimulate revenue growth. Smart pipes give subscribers better quality and better privacy protection.
Policy management’s critical role
Pipes will play a critical role in the mobile value chain. Intelligence will come from smart pipe components to monetize data services and effectively manage network traffic; this is where policy management comes into play.
Policy management, recently implemented in 3G networks, will also be widely used in LTE/4G networks. PCRF (Policy Charging Rule Function) will be an integral part of policy management and a core network component. PCRF manages the policy and charging rules for subscribers, allocating different bandwidths for each, controls QoS (quality of service) parameters, and associates these factors with the appropriate charges.
Charging policies come from the billing system, which maintains all charging plans generated by the operator, and from the CRM, which maintains the personalized subscriber plans. However, no standard or convenient mechanism exists for generating policy rules from the billing/CRM information. The correlation between billing plans and rules is not one-to-one, so there is no simple association between the two.
Mobile operators need sets of policy rules that function in real time. For example, subscriber A purchased 100MB, while subscriber B purchased only 50MB. However, the network can only provide a maximum of 60MB at a certain point in time. To keep both subscribers happy, how should the 60MB be allocated between the two – 40MB/20MB – or perhaps 50MB/10MB?
Numerous parameters can affect the bandwidth allocation, such as the number of subscribers, the current load, the rate plans of each subscriber, and more.
Without A sophisticated mechanism to dynamically adapt the rules, PCRF is only a skeleton for static policy rules.
There is a need to resolve the gap between the billing/CRM systems and the core network systems by developing a process that can automatically build efficient PCRF rules.
A Policy Centre is the answer
A Policy Center would be located between the OSS (Operator Support Systems)/ BSS (Billing Support System) and the core network parts of the mobile network. It would enable the operator to dynamically generate new services in a personalized manner, as well as to automatically generate policy rules for the service. The Policy Centre would be active in generating rules, such as who pays more for which services.
A Policy Centre would typically serve three major roles:
Dynamic service generation would generate new plans with greater flexibility than billing. For example: Customers with a monthly bill of more than $300 would receive a bandwidth upgrade to first class
Dynamic rule generation would create PCRF rules to implement the generated service, using subscribers’ profiles and tariff plans from the billing system. For example, implement PCC rules for all family members, if they have a ‘shared wallet’ billing plan.
A rule mediator for third-party service providers would receive and validate rules offered by third parties. Moreover, it would also check for cross-impact and conflicts between third parties. An example would be enabling third party apps to grant premium access (bandwidth) for specific content, i.e., reverse charging for specific websites (‘1-800- FOR-DATA’).
Big Data Analytics are instrumental
Big data analytics will play a pivotal role in making recommendations, in real time. When building new rules for subscribers, one method would be to search for other subscribers with similar plans. A new rule would be selected from among the rules identified for those subscribers, based on the rule that generated the best utilization of the allocated bandwidth according to the history log.
The system can be a self-learning system, where the association of subscribers to rules would improve over time – as the amount of historical data grows.
Data analytics also takes charging and revenue generation into account. The generated PCRF rules would be tuned to find the optimal connection between the mobile operator’s requirements and the end user needs, so that end users get exactly what they want and pay for exactly what they get.
Large amounts of data must be collected for rule optimization regarding the required level and actual level of subscriber service. There is a need to monitor the level of service at different times and locations and to collect and aggregate all findings.
Analytics can also determine bandwidth requirements at a given time, the amount allocated and the amount actually consumed and whether or not the allocated service was sufficient.
Policy and network steering
The Steering of Roaming (SoR) process is used by the home network to determine the ideal visited network for each subscriber, in real-time (registration time). Its main use would be the allocation of WiFi offload in roaming, when the operator’s network is congested. In overcrowded cities and popular holiday destinations, operators are already experiencing difficulties in meeting the data demands of iPhones, Androids and Blackberries.
A new standard component in the LTE network, LTE ANDSF (Access Network Discovery and Selection Function) adds new functionality to the steering process by combining cellular and WiFi steering at one central decision point. With ANDSF a server instructs the client which network to select in real-time. Mobile operators will be able to use ANDSF logic for WiFi offload, as well as other selection criteria between 3G/4G/WiFi for the local (domestic) network and roaming.
Several operators have implemented WiFi offload; they offload their heavy data subscribers to WiFi hotspots where available and where they have agreements with WiFi service providers.
Client software on the mobile device usually takes care of WiFi selection, but in current 3G networks, manual authentication (a password) may still be necessary.
Billing is another challenge; the end user should be billed for WiFi access by the mobile operator, however since the billing systems are not unified this has proven to be difficult.
These challenges are targeted by the LTE/4G network, where standard network support is added to the network for native seamless and unified control of 4G/WiFi handover.
Mobile data roaming is still expensive, and WiFi is a relatively cheap alternative. Roaming managers need a WiFi solution for their travelling subscribers – one that avoids bill shock and keeps the revenues with their home network.
There is a close synergy between the decisions to steer client to a given service provider and policy management. Steering itself is a policy-driven process that can use subscriber policy profiles to select a network based upon the subscriber’s profile. A heavy data subscriber can be redirected to a high-speed LTE network, while another subscriber might be redirected to a 3G network. Steering of roamers can therefore follow a subscriber’s policy-rule as defined by the PCRF and can take the charging rule for the subscriber into account.
Steering of subscribers to the appropriate access network strongly supports the monetizing of network pipes.
M2M (Machine-to-Machine or embedded mobile) is a new market for mobile operators. The GSMA (GSM Association), the leading forum of mobile operators, has established the Embedded Mobile Initiative to address this new market potential.
Many M2M devices will carry a SIM card that belongs to a specific mobile operator, an MVNO, an enterprise or a M2M service provider. If deployed in another country, the device would be in a state of permanent roaming.
Permanent or continuous roaming might also be built-in by the factory for, say, an always-connected utility meter; this would assure service continuity even when the home mobile network is unavailable. Permanent roaming has an advantage over local subscriptions, as several options are available for network coverage. For instance, if network coverage is lost, the device will select another roaming partner automatically. With local subscriptions, only the home network can serve the device.
M2M devices such as meters and monitoring devices are often installed in places where coverage is weak, such as basements, and in this case, the permanent roaming advantage can be crucial.
Pipes are the future
In a market saturated with competitors, high-end devices and consumer demands for quality of service, it’s no wonder that pipes, smart ones, are becoming the best solution to manage and regulate LTE and WiFi offload. It is after all, the smart alternative.