Transcript of "GDS International - Next - Generation - Telecommunications - Summit - Europe - 8"
Financial Benefits of RAN Aware Policy Management and Video Optimization It is well known that mobile broadband data is growing very rapidly and that it poses several challenges to operators to deal with. Operators are challenged with keeping up with the traffic growth, most of it from video applications, and with implementing a sustainable revenue capture model that supports network infrastructure investment. Typical questions pertaining to the management and optimization of mobile broadband networks relate to real time visibility, either directly or indirectly: 1. What is the full capacity of the network per cell area? Relative to this full capacity how is the network operating? 2. Across the different resource pools (radio, signaling and bandwidth) how is the network loaded? 3. Can you get more capacity out of a network by evening out the load? 4. Can you safely offer premium services in real-‐time and rely on the network to come through on the promise? If so, can this be a driver for incremental revenue growth? 5. What applications are subscribers using and which of these pose the biggest challenge to network? 6. Based on information about which applications subscribers are using, where they are using them and when, is it possible to devise new services that can help business growth and customer satisfaction? 7. Is it possible to focus video compression on cells with congestion issues? 8. Can video optimization focus on users in areas with certain radio conditions? 9. Can video optimization differentiate between different types of users and handsets? The CommProve BCN RAN solution is aimed at helping operators deal with the complexities of managing and optimizing mobile broadband data networks. Such networks are loaded very unevenly as users move around and change their behaviors and usage patterns. This dynamic environment calls for an equally dynamic and responsive real time approach to optimization and revenue capture. Currently, networks are managed with only limited insight into quality of experience (QoE) and actual network traffic levels relative to capacity. This makes it difficult if not impossible to properly handle radio access network (RAN) congestion and it does not allow for introduction of real-‐time premium services. Policy control solutions aim to address these shortcomings. With traditional policy control solutions there is an interplay between three elements: 1) deep packet inspection and policy enforcement, 2) PCRF or the rules function and 3) billing systems with subscriber specific information and plans. Traditionally, deep packet inspection is done in the core of the network (Gn or Gi interface) from which information about who the users are and what they are doing 1
is gathered. This information is sometimes augmented with information from the Gx interface (Radius protocol) adding a level of location awareness but not in real-‐time. This information is fed to the rules function (PCRF). The rules function also gets per subscriber information in terms of pre-‐paid plans etc. from the billing system. Based on this, the PCRF determines which rules on a per subscriber basis should be enforced. The enforcement is done by the DPI solution. Similarly, solutions are deployed at the Gi or Gn interface focusing in on video optimization as the majority of the mobile broadband growth comes from a variety of video applications. For such solutions optimization, both in terms of improved video quality and bandwidth compression, works well across all users and geographical areas, but if optimization needs to be done for specific users, or by differentiating between cell areas depending on their congestion levels, for example, there are certain shortcomings. These concepts only work well when users aren’t mobile and as long as they stay close to base stations. What is clearly lacking is a set of subscriber specific information. The policy manager or the video optimization solution knows nothing about handset received power levels or interference power levels. It’s only possible to view users who can get on the network, and to manage bandwidth resources in the core of the network but there is no insight into signaling resources or radio resources and last but not least there is no real-‐time information about the location of each subscriber. Therefore, policies are enforced with no real-‐time visibility to subscriber location or movement, and with no visibility to the subscriber impact on other subscribers in terms of network interference, leading in many cases to throttling of the wrong user. As an example, if one user is close to a base station but consuming a lot of bandwidth the policy manager is likely to make this individual a candidate for throttling. If at the same time another user is far from the base station, even if he is consuming less bandwidth, he may be a much better candidate for throttling due to introduction of service impacting network interference from the higher power levels in connection with the communication. Additionally, when policies are enforced or video is optimized it is done in a push like fashion with no feedback loop from the RAN with QoE information about what the policies did to the satisfaction of the subscribers. Finally, with no information about service conditions at the location of each of the subscribers, operators are not able to offer real-‐time premium services as it’s just too risky. BCN RAN was introduced to address these issues head-‐on. The BCN RAN solution collects and processes in real-‐time, measurements from different interfaces of the RAN. DPI measurements and RAN quality measurements are both collected 24x7 network-‐wide. Interface monitoring includes Iub and IuPS interfaces. DPI measurements provide information concerning application and network level delays, retransmissions, peak and average throughput, application type breakdown, both at 2
user and cell level granularity. The cell level granularity is essential to understanding if, how and where the load generated by the user is affecting the network availability for other users. This information is generated leveraging BCN RAN’s correlation capabilities; user mobility is tracked and the actual cells serving the user are saved in a call detail record (CDR), along with all relevant signaling information (Call Set-‐up/Tear down, RAB Assignment/Release, Failure Events, Inter-‐RAT Handover events). With the Iub interface monitoring included, the signaling procedures of the radio interface can be logged (HDSPA/HSUPA transactions, radio bearer set up/reconfigurations, soft/hard handovers, transitions between Common Channels and Dedicated Channels) and correlated with channel quality measurements (Received Signal Code Power, Ec/N0, Transmitted Carrier Power and Received Signal Strength Indicator). The additional measurements from the Iub interface allow the Policy Manager to differentiate between the users based on the effect their data consumption has had on the congested cell resources; a data intensive user close to the antenna and in good visibility conditions is much lighter on the radio network than a stationary lower bandwidth indoor user in a severe interference condition. If air interface resources are congested, differentiating between these two types of users is vital. Throttling the latter user is much more effective in addressing the congestion. On the other hand, if the bottleneck is located at the backhaul level, it is the first user who should be throttled as he or she is consuming more of the transport resources. CommProve’s BCN RAN is the only solution capable of capturing and making sense of this type of information as well as base station power allocated to each user, which in some cases can be a more scarce resource than bandwidth. BCN RAN’s real-‐time correlated CDRs provide the Policy Manager with all the information needed to make the Policy Control decision based not only on the user data consumption information but, most notably, on how the user data consumption is affecting the network resources. The Policy Control decision can be taken in a much more effective way, minimizing the perceived effect on the user QoE and maximizing the utilization of the network resources. Finally, with BCN RAN the Policy Manager becomes QoE aware in the sense that after policies are enforced BCN RAN can tell the Policy Manager what it did to QoE and before any area is deemed to be in congestion in the first place BCN RAN can use a QoE based congestion measurement to tell the Policy Manager that congestion has occurred. The user specific QoE information can also be used to enable premium services in real-‐time for incremental revenue capture from mobile broadband data. All of the BCN RAN functionality is offered in a stand-‐alone point solution or it can be implemented as a software application on the CommProve NetLedge network monitoring system. 3
In order to quantify the financial benefits of BCN RAN, specific network information is needed. The financial benefits of BCN RAN fall into three categories: 1. With BCN RAN it is possible to load networks with more traffic and as such defer CAPEX expenditures on network infrastructure expansions. Often cells are loaded at less than 25% capacity but operators do not have access to this type of information and they dont have the tools to even things out 2. With real-‐time visibility of QoE for subscribers operators can better manage customer satisfaction and subsequently address any churn issues 3. Also based on real-‐time visibility of QoE, BCN RAN is an enabler for offering real-‐time premium services and as such for incremental revenue growth. How this is done depends on what services the operator is interested in providing. Additionally with the per subscriber insight into applications used, their location, their QoE and how they load the network, operators gain insight into which services to offer. This can also be used to estimate the merit of marketing campaigns by comparing subscriber behavior before campaigns are launched with how this develops in the hours and days after launch. 4