Telefónica: One step ahead in E2E QoE assurance

Ensuring end-to-end (E2E) quality of experience (QoE) in mobile broadband networks remains a key challenge for many operators, but Telefónica is one step ahead. Technology Director Cayetano Carbajo shares his insights on this, based on his thorough understanding of network evolution.

According to Cayetano Carbajo, Telefónica has started our E2E QoE project for mobile broadband networks, with a system measuring more than 60 QoE indicators that cover 2G/3G data bearing, mail, web browsing, streaming, and file sharing. The system has an important mediation mechanism that collects information from all network nodes, including the radio network controller (RNC), domain name system (DNS), online charging system (OCS), and serving GPRS support nodes (SGSN); it also collects information from probes. In this mediation mechanism, all information is integrated and sent to a database to do all the correlation needed for measuring the E2E QoE. While ensuring E2E QoE, Telefónica can use the data for big data applications and customer care. Also Telefónica can have information of behavior of devices. For instance, Telefónica can measure E2E latency analysis per device model to find out the average daily percentiles for each device model and its impact on QoE.

By Michael Huang

Ensuring end-to-end (E2E) quality of experience (QoE) in mobile broadband networks remains a key challenge for many operators, but Telefónica is one step ahead. Technology Director Cayetano Carbajo shares his insights on this, based on his thorough understanding of network evolution.

E2E QoE a key differentiator

WinWin: Quality of user experience is becoming a key concern for network operations. Can you share Telefónica’s practices for this? What other network intelligence elements should be introduced?

Cayetano Carbajo: We have many network intelligence elements that can be integrated into our networks. The investment for doing that is not huge (at least not that huge as for the access network), and the value brought to operators is significant. These include managing E2E QoE, integrating customer data into our big data platform, managing the cooperation between the cloud and transport networks, delivering video in an efficient way with ensured quality, and many others.

Among them, I think measuring the E2E QoE is key. Beyond the complex network set of procedures, QoE is what the user sees at the end, and it directly affects user satisfaction level. QoE relates to all elements in the communication chain; no matter what is failing, the Internet, the applications, devices, or operators’ networks, customers’ QoE can be affected. For measuring and reacting to QoE assurance, we are not well prepared yet, as we have to implement new systems, update our internal processes, and enhance capabilities to react for QoE improvements.

We have started our E2E QoE project for mobile broadband networks, with a system measuring more than 60 QoE indicators that cover 2G/3G data bearing, mail, web browsing, streaming, and file sharing. The system has an important mediation mechanism that collects information from all network nodes, including the radio network controller (RNC), domain name system (DNS), online charging system (OCS), and serving GPRS support nodes (SGSN); it also collects information from probes. In this mediation mechanism, all information is integrated and sent to a database to do all the correlation needed for measuring the E2E QoE. While ensuring E2E QoE, we can use the data for big data applications and customer care. Also we can have information of behavior of devices. For instance, we can measure E2E latency analysis per device model to find out the average daily percentiles for each device model and its impact on QoE.

WinWin: When facing OTT players, what unique advantages do you see on telco’s side in terms of technologies and networks? What are Telefónica’s practices here?

Cayetano Carbajo: I think our advantage is that we are in the middle – both OTTs and users need connectivity, which is provided by operators. The relationship towards customers is also unique for us. When customers have problems, they call our customer service centers; when there is a physical person needed, it will be from the operator instead of the OTT. I think the word ‘coopetition’ is a very good summary of what we need to do with OTT, and integrating intelligence into our network (which is our biggest asset) is key for us to compete/cooperate with them. If we add intelligent functionalities that are useful for OTT, they will be using our network in a more complete way and also a safer way for us.

WinWin: How would you suggest operators prepare their overall technology and network strategy for the future when considering OTT coopetition? Will SDN help in this regard?

Cayetano Carbajo: I think something we have learned till now is that we need flexibility, as generally we don’t have any idea of which will be the services and what usage our customers will have of our networks. If we look back five years ago, it was very difficult to predict the majority of services we have today. In the same manner, predicting the services in the next five or ten years is impossible. The rationale here is that we need much flexibility in reacting to the new services and new service requirements, as well as the capacity increase. The second thing is that the technological cycle and service cycle, because of OTT, have been shortened during the last few years, and we need shorter time-to-market (TTM) for our services. Yet with all these things, operators’ characteristics – reliability and trust – should remain in the end.

I think SDN is a fashion now and theoretically it sounds very good. We would also like to have the advantages of SDN in our networks today; unfortunately, it is not mature enough yet. What we are doing is trying to use SDN in the data centers in a way which is more feasible or easier to implement. Also we are cooperating and innovating with a lot of partners to bring SDN as soon as we can to the transport networks.

Building future-oriented networks

WinWin: Currently, we see Telefónica increasing its efforts in M2M. Would you please describe Telefónica’s strategy for network access? How will it efficiently accommodate more diverse applications?

Cayetano Carbajo: We have a very well-defined access strategy for both fixed and mobile. In the case of mobile, it is clear that we are moving to LTE. We are deploying LTE, , in our operations, which will bring a lot of benefits to the access, not only in terms of capacity increase but also a new architecture more suitable for data services. In the case of fixed access, our bet is fiber. We are deploying fiber to the home (FTTH) in many countries, including our two major markets of Spain and Brazil.

For the compatibility between application and access, I think in general the access has to be agnostic of applications, and any application could run on the access. Of course you have to broaden your access and improve it for dealing with the requirements for applications. There are applications such as M2M that will pose significant challenges in the access. In coming years, the traffic will be significant, so challenges have yet to come, but we need to prepare for them.

WinWin: LTE is bringing increased challenges to backhaul. Can you talk a bit about the progress of your backhaul construction? Based on your current infrastructure, what are the viable solutions you are looking at in this regard?

Cayetano Carbajo: It is clear that backhaul will be a bottleneck for LTE networks. One solution is to bring the fiber near the base station sites, meaning building first an aggregation network and then taking fiber directly to the base stations where there is a lot of traffic. Another one is microwave evolution to IP technologies that can increase significantly the capacity. Also we need to take into account a lot of functionalities that are needed for LTE, including both frequency and phase synchronization, as well as the way to aggregate the transport routes from different sites.

In LTE, we also have to consider the whole backhaul chain. Having a very good radio access and not the right backhaul, or having very good radio access and backhaul network, with some functionalities missing, such as buffering sites in the radio links, may ruin your efforts, as the final customer experience will not be as good as expected.

WinWin: With data traffic booming, what is Telefónica’s strategy for the transport and backbone layers? What are your current priorities, and what will you be looking at in the next the two to three years?

Cayetano Carbajo: What we are doing now is upgrading the transport network capacity. We see that 100G is finally at a rational cost for commercial deployment, and we are implementing it in our operations and this significantly increases our network capacity. We are not deploying 400G or 1T right now, instead we would plan it for the next five or six years when the cost of the technologies is reduced to a point that is rational for deployment. Maybe we can expect limited beyond-100G applications in certain links, but the cost is yet to be reasonable.

For the next two or three years, our plan is to work towards convergence. We don’t have a quite clear view of IP and optical convergence yet, and I think multi-vendor support is a key requirement, especially in the control plane. I think the path computation element (PCE), as well as SDN are expected to be working at in the next 2-3 years.

Another trend could be integrating the optical elements in many other network nodes, forming a kind of DWDM everywhere. We will also continue evolving our photonic meshes, with some strategies already implemented, and we have seen significant improvement in network management and redundancy. I think in the coming two or three years, we will be expanding our capability in this field, and finally we can probably have contentionless in the photonic mesh networks.

WinWin: Would you please talk a bit about your mobile core network construction? How are you planning to evolve your packet core networks?

Cayetano Carbajo: We are evolving our core networks to add more capacity. We are also thinking about integrating some external functionalities into the packet core networks, such as video optimization and architecture simplification. The core network should also accommodate multiple technologies such as GSM, UMTS, and LTE. Redundancies such as the pooling of mobility management entity (MME) and SGSN are also very important for us. In addition, the protection mechanism for signaling storms in MME is critical, as we have learned some lessons from serious node congestion due to these storms.