with Kris Rinne, EVP of Network Technologies at AT&T Labs
AT&T (NYSE:T) has been rapidly expanding its LTE network to reach its goal of having 150 million POPs covered by year-end, which will be the half-way mark for its ultimate goal of having LTE cover 300 million POPs by the end of 2014. Kris Rinne, AT&T Labs' executive vice president of network technologies, is at the forefront of that charge as she is responsible for the company's network architecture, service platforms, radio access roadmap and initial implementation.
FierceBroadbandWireless Editor-in-Chief Sue Marek met with Rinne at last week's 4G World conference in Chicago to talk about the company's LTE deployment, including its testing of carrier aggregation, deployment of small cells and more.
FierceBroadbandWireless: You have talked about the ability to do carrier aggregation in the future, but you haven't provided a time frame. When do you expect to have that capability?
Kris Rinne: We haven't given a time frame but we would expect to be in a test mode in 2013. It's getting carrier aggregation from the infrastructure side and then getting it in our network so we can start certifying the devices.
FierceBroadbandWireless: Can you explain the device part of carrier aggregation?
Rinne: In our case we would look at combining the 700 MHz spectrum and AWS spectrum in carrier aggregation--that is one of our first combinations. Even though you have the devices with radios for both the 700 MHz and the AWS spectrum it doesn't have the software as part of the chipset to aggregate that as a single carrier. That's how you get that content across both those spectrum bands. You need the software to collect the data that is being transmitted across both and collect that as a single, common stream.
There will also be some optimization work because we will have different propagation characteristics. If you are seeing those different bands with different signal strengths, how do you deal with that?
FierceBroadbandWireless: Almost every operator is talking about implementing carrier aggregation at some point. So besides the radios in the devices, you will also need this software in the chipset. That seems like a lot of different elements.
Rinne: If, for example, 700 MHz and AWS spectrum is our first carrier aggregation, and then 700 MHz, 1900 MHz is our second aggregation, then the next step is not as much work. But you are right, different countries are going to be asking for different band preferences and priorities. Hopefully once we have the base capabilities out there, making it work in different combinations will be more streamlined.
FierceBroadbandWireless: You have consistently talked about VoLTE happening in 2013. How big of deal it is to integrate E911 and SMS with VoLTE? Do these leverage your IMS architecture?
Rinne: The E911 is separate. It is not leveraging the IMS architecture. The way our A-GPS works is over the circuit-switched channels, and now we need it to work over the LTE infrastructure so that requires taking software in the devices again and upgrading those controllers related to that. The information it is going to get on the location is formatted a different way, so upgrading that infrastructure is part of it.
The testing is pretty critical to make sure you have the same quality in that different protocol.
The SMS part is having an element tied to your SMSCs [short message service centers] that knows to send an SMS on an IP-basis when a customer is in an LTE coverage area vs. when a customer is not and then to send it over the traditional SMS channel.
I view these as basics in getting the VoLTE upgrade out there--the E911 integration and the SMS upgrades and the QoS implementation. Then RCS [rich communications services] is the increased feature functionality that I can add on top of VoLTE now that I have leveraged that basic functionality.
FierceBroadbandWireless: Can you talk about small cells a bit more. You talked about trialing metrocells. Can you provide more details?
Rinne: I consider metrocells to be a small cell that handles 32 to 64 simultaneous users. Femtocells handle up to four simultaneous users. We are focusing on the UMTS/HSPA piece of this. In 2013 we will add the LTE portion to that.
FierceBroadbandWireless: Are the metrocells easier to deploy than femtocells?
Rinne: The zoning issues are not eliminated, but it is more straightforward because you are not trying to construct another tower. You are leveraging a light pole or the side of a building. Some of the backhaul challenges are there. We are looking at how we can use more consumer-grade backhaul. You have the U-Verse or broadband capabilities out there. It doesn't have the same latency requirements or lack of jitter than I would require in my macro network to support handovers etc. How can you leverage consumer grade broadband in a residential area or office or a park? That is one challenge.
FierceBroadbandWireless: Everyone is talking about cost per bit. Is that the metric that you must look at when you make decisions with the network?
Rinne: I have always been challenged with looking at the cost per bit in my network. We just didn't talk about it as much. As we are continuing to scale, it is something we are driving in the ecosystem. Over time we need to have this kind of efficiency. You have to address it on the spectrum side and also how you deal with that back into the network.