By Joachim Horn, Group CTIO, Tele2

LTE is not exclusive to the incumbents. For Tele2, a pan-European operator and challenger that proudly calls itself a “black sheep,” LTE fits in just as well as it allows the delivery of a better user experience at a reduced price. At the Huawei MBB Forum, Tele2’s Group CTIO Joachim Horn shared the telco’s LTE story, which is very positive, but as he points out, challenges remain..

Born to be cheap

I want to introduce you to a commercial Tele2 ran in Sweden when we launched LTE more than a year ago. The black sheep in the spot is our signature animal. We named him Frank. Why a black sheep? Sheep sounds like cheap and we are a low-cost operator. The sheep is black because we see ourselves as a challenger in the market. Finally, the sheep is called Frank because that’s our culture – open, honest and direct, with our customers and also the ecosystem.

When the time to introduce 4G arrived, the question was, “How do we communicate this to our customers?” Customers tend to get confused about what these acronyms all mean. So the idea was to say it is a new generation. How better to introduce a new generation than by giving a baby to Frank? Frank has been known for many, many years. People who know anything about Tele2 know Frank. Our story in the commercial was simple. Frank went to China and miraculously returned with a baby, i.e., the fourth generation. Being a low-cost operator, we focus on what is really important for our customers – no fancy extras, and driving down the costs as much as we can. We do not compromise on the quality that the customer can experience and we compete on that basis. When I share our experiences with LTE and views on some of the features and functions, it is in the context of being a low-cost operator and a challenger.

4G introduction has been positive

Tele2 is active in eleven markets across Europe and Euro-Asia. Our largest market is Russia, followed by Sweden. We have 38 million customers, 9000 employees, and a sales revenue of SEK41 billion (roughly USD6 billion) for 2011.

Although we are a challenger, and we are cost-conscious, this doesn’t mean that we do not do new technologies. We actually like LTE because it is a way to drive down the cost per bit per second, and at the same time, improve customer experience. This fits perfectly with our overall way of working.

We have already built LTE networks in Sweden and the Baltic countries. The most advanced is Sweden where, as of today, we already have 80% population coverage. By the first quarter of next year, our LTE network in Sweden will most probably be the first country in the world to reach 99.5% population coverage. We will achieve this by using 900MHz for coverage, currently with 5MHz of bandwidth, increasing to 10MHz early next year. In addition, we are using 20MHz of 2.6GHz spectrum for capacity.

We are sharing this 2G/4G network with Telenor. Basically we have put our existing 2G networks together, and replaced old base stations with the new SingleRAN technology from Huawei, which allows us to now run LTE and GSM in the same base station through the same radio with the same antennas. We opted for 900MHz as LTE spectrum because we could easily free up some spectrum after merging both GSM networks and it’s a great band to achieve coverage with. I have to admit that there were some concerns that devices would not be so easily available for 900MHz. But now, LTE 900MHz devices and smartphones are becoming available, also from Huawei. Many people do not even know that LTE works at 900MHz, but here we are, and I can say confidently that it actually works very well. In addition to 900MHz and 2.6GHz, we have 800MHz and 1800MHz spectrum which we plan to use starting in 2013.

Our experience with LTE has been really good. 4G has been adopted much faster than 3G. Remember that for the first three or four years after introducing 3G, in most markets, there was hardly any data usage; most operators have experienced this. Contrary to that, at two years after the launch of LTE in Sweden, we are at the data usage level we were at with 3G at seven years post-launch. We have skipped the pre-maturity phase entirely, so to speak. LTE is already working great; it’s working better than 3G.

Practically all data growth in Tele2 in Sweden for the last 12 months has gone to 4G, while 3G data usage has been stagnant, though at a very high level. LTE carries already now, after only two years of operation, the majority of our growth. One reason for this is that obviously most of the dongles which produce huge traffic have moved to 4G. So, it’s not like there would be no growth for 3G anymore, but we expect that for Sweden, in one or two years, all the data growth will be in 4G. This assumption is based on the number of new devices coming to the market, most of which will support LTE. And for our customers, with the better user experience brought by 4G, we will see them moving actively into LTE.

4G introduction has been a very positive experience for us. It’s a very robust technology. We have very, very few issues. 4G does not have the infant mortality problems we had in early technology times; it is very, very stable. 4G also delivers a robust performance in terms of throughput – that’s a major difference from 3G. With 3G, you can enjoy very high throughput, but that’s only at the center of the cell. The moment you move closer and closer to the cell edge, the performance deteriorates quite rapidly. This is not so much a problem for LTE. Also at the cell edge, you enjoy a substantial performance and this makes LTE far superior to 3G. Another feature we like about LTE is the low latency; it’s amazing. Measuring in the field, we typically find 20 or 25 milliseconds of latency; that’s a third of 3G. This makes browsing a lot of fun, and our customers really love it. The very short time to maturity for the network was a real surprise to us.

And last but not the least, there are advantages coming from SingleRAN, which I consider to be the first technology to be true software-defined radio. We are fully enjoying the benefits. For example, we didn’t expect the iPhone 5 to come out supporting only 1800MHz in Europe. It’s a shame, I have to say, but fortunately we have 1800MHz spectrum, and it will take us only a software change to enable LTE in 1800MHz. This is only possible because we use SingleRAN technology.

Challenges remain

There are also some challenges that I see with LTE. Critical is the lack of voice support. While we have introduced a fascinating new technology for data, we seem to believe that we can leave voice behind. I think that’s a big mistake. This has been one of my pet topics for discussion with industry players over the last year. It’s about Voice-over-LTE (VoLTE).

Whether we like it or not, despite the fact that we have invested billions in the last ten years to grow the data business, we didn’t manage to make data a better source of revenues and profit than voice. And for the foreseeable future, voice will remain a major contributor to our revenue and profits. Therefore, it is indispensable that we get voice to the next level. But in the last several years, I think little investment has been done in voice, while a lot of investments have flown to a lot of new data services, of which a good deal didn’t turn out to be successful.

We need VoLTE now. That’s the only way I believe telcos will be able to compete against the OTT high-quality services such as Skype and FaceTime. There are some technical requirements which I think are important. First, voice quality should not only be the same as 2G and 3G, as this quality is already not competitive anymore against the OTT services; it must be better. Handover is another one. Whoever knows the abbreviation SRVCC (Single Radio Voice Call Continuity) knows that it takes at least 300 milliseconds for a handover in a best- case scenario. While I’m aware that this is already an improvement compared to earlier versions, I personally believe that this is still not acceptable from the customer perspective. We have to reduce handover complexity to make it faster. In addition, we must define roaming for VoLTE, so for instance, when I go from Sweden to Canada, I should be able to use my VoLTE service there.

There are different possibilities for making a voice call in LTE – VoLTE, CSFB (Circuit Switched Fallback), and OTT (VoIP). CSFB is what we use today in Europe. Every time I want to make a voice call, my phone goes back to 2G or 3G and then establishes a circuit switched call. This gives us the same carrier-grade quality in voice as we are used to in 2G/3G. But if you look at rich media, CSFB cannot support simultaneous 4G data and voice. This is a step back! In 3G, we can have data and voice in parallel, but in LTE, we can’t do this when using CSFB.

VoLTE enables different levels of quality, and this is really important. It will be the first carrier-grade VoIP service on any cellular network. And there is a big opportunity for integration of VoLTE with other applications, because VoLTE is an All-IP service which can easily be merged into rich media services. Through that, we can create a new level of seamless and superior customer experience.

So we believe VoLTE must have top priority on every operator’s LTE introduction list. I have a lot of discussions with fellows and friends from other operators, where I recommend to not think of VoLTE as “we can do that later.” I think that, as an operator, we will not be able to recover from such a delay.

LTE-A improvements

We did some trials together with Huawei on LTE-Advanced, and achieved some impressive speeds of 290Mbps thanks to carrier aggregation. We like it because it will help further reduce costs per megabit per second. This enables us to give better services to our customers at a more favorable price. In addition, if you don’t have 20MHz but 10MHz on 800MHz and 10MHz on 900MHz, you can combine them and enjoy the maximum throughput. This is really a cool feature, and I personally believe it will work fine.

Our priorities for LTE-Advanced features, besides carrier aggregation, are better QoS integration (as I mentioned earlier) and relay nodes. Differing from repeaters, relay nodes can basically extend your coverage geographically, using in-band LTE backhauling. We are also interested in uplink MIMO, up to 4x4, but there are some other issues not yet resolved.

I’m personally not yet convinced regarding the buzz around HetNet. This doesn’t mean that I don’t believe in microcells or their structure at all. The major reason for my hesitation is the fact that backhaul of small cells has yet to be solved, and nobody can explain to me so far how this can be done on a large scale with low cost.

Coming back to QoS, we know that we use it for two things. We can give a customer a reliable service quality with SLA, and run tiered pricing plans, while the other thing that we do is protect the network from unnecessary load, such as spam or other user experience disturbances. Architecturally, QoS control is being done today somewhere in the core network, behind the EPC (evolved packet core). There are lots of bumps on the wire until the traffic hits this intelligence and the system says, “I don’t like that traffic, so please don’t allow it or please prioritize it down.” The problem with this is that the traffic has run through the base station already, utilizing our most precious resource, which is spectrum.

I believe that we need to change this. We need to modify LTE-Advanced to move the QoS intelligence to the edge, to the scheduler of a base station, making it aware of what is happening, while making sure that we protect our valuable spectrum resources from unnecessary traffic. By doing this, we can expect a better user experience and higher efficiency. The scheduler, in my opinion, is the perfect place to make sure that the network is best utilized. Technically, it should be possible; it just needs to be done.

In conclusion, the next generation is here; it’s LTE. It is growing fast and delivering robust performance at lower cost per bit and a superior customer experience. But, there are some issues as well, particularly with the very important voice services. Early VoLTE availability will be critical. We expect LTE-Advanced to bring important advantages, but we should be clear on what’s really relevant for us, and focus on that to make sure it’s there soon. In particular, quality of service needs to be improved substantially and critical telco applications need to be made available fast. Here, I would request that we work together in the ecosystem so that we can push these items forward.