Etisalat: “We went to LTE to be prepared - before the traffic comes”

Disclaimer: “The following interview expresses Dr Hamdy’s personal opinions based on his own experience and published research work. It does not reflect his company’s official policy or directives.”

Etisalat is the largest mobile operator is the UAE, commanding 60 per cent of its local market. As one of the major operators in the region the recent launch of an LTE service across most of the major cities in the county can be viewed as a major statement of its intent to stay ahead of the competition.

Ahead of his key speaking role at the upcoming LTE Middle East and North Africa conference, to be held in Dubai on the 29th-30th April, we catch up with Dr Mohamed Nadder Hamdy, director of capacity planning for Etisalat.

At what stage is your LTE rollout?

We have commercially launched LTE in September [2011] and started selling the devices a few weeks back. We have covered all the main cities with around 1000 base stations on 2.6GHz.

What are the major challenges that an operator will face when deploying an LTE network?

The first challenge is the selection of the spectrum to be used. Most countries have the 2.6GHz bandwidth available, but it has very bad propagation capabilities. It requires significantly more sites that what we used for 1800 or 900MHz. In order to use any of the GSM band, in particular 1800 you need to refarm from the GSM 2G network, which is not that simple. A significant number of users still have 2G handsets, as they are cheaper and don’t consume the battery [as much] as the more advanced smartphones. So that’s a major issue – the spectrum allocation, and selecting the best spectrum to deploy. For us, we have started with 2.6, as we have no choice. We had to deploy LTE as fast as we could and there was nothing else available. It’s fairly costly, in terms of the number of basestations and we feel that in terms of future expansion we should be going for the 1800 band direction.

How was your spectrum auction process managed?

The spectrum was not auctioned – we just applied. It’s a little different here in the UAE, as there are only two operators here and you pay fixed fees for the band annually. So even for the current 2G and 3G bands we are paying annual fees. It’s not a one-time lump sum that we pay.  And similarly the 2.6 band follows the same scheme. We requested an allocation of 20MHz on that band and we just paid the fees.

Why would you characterise the 2.6Ghz band as “very bad”?

What you care about most at this stage is the coverage everywhere – and the distance you can cover with this band is about one third of what you can cover with the 1800. That forces you to deploy a lot more base stations. The penetration for the indoor environment is also much worse that for the other bands. My experience with this band was not encouraging and we would tell other operators that are going into LTE to apply for lower bands.

Because its propagation distance is smaller, it means that you need to put more cells next to each other to get more capacity, [It’s a challenge] …if you are starting a new network and you want to have greater coverage, but you don’t have the traffic this band definitely doesn’t let you do that without a reasonable number of sites.

Are you planning on refarming your network to enable LTE to operate on 1800?

Our dongles do operate at 1800 but the fact is that most of our 2G subscribers are still using 2G handsets. We don’t see that they are going to change it in the near future. We still have a lot of traffic on the 2G bands so it will not be easy to refarm it straight away. To obtain full throughput from LTE you need around 20MHz of bandwidth, so we are taking it in steps for 1800. We might start with 5 or 10MHz and as the traffic on 2G reduces we will evacuate more and so on. I would say this would take approximately two years.

How common are the spectrum issues you are facing are to the rest of the Middle-East?

It is common and it is causing a great deal of ambiguity. It’s a compromise – if you want to start now then you have to use a worse band than if you wait a few years more. For example, the digital dividend that most countries are evacuating by giving up the analogue broadcasts. Even if we are promised that it will be evacuated, we don’t know if it will be in one or two years. So you are under this pressure to wait or you can deploy now with the higher cost of base stations. It’s a dilemma.

What other crucial issues do carriers need to take into account?

Another problem that everybody is facing is the availability of devices. Currently in our market we don’t have handsets at all; they’re all data dongles. The issue is that devices are not all capable of handing over to the previous 2G and 3G networks. This aspect needs a lot of improvement as it’s not very smooth so far. In the US they have some phones but they are CDMA with LTE- they don’t have phones that work on UMTS with LTE.  Of coursethe lack of voice is one of the major challenges so far. It is not available in the initial deployment. We are not offering voice services and this is one of the things that will limit the strength of LTE in the initial stages.

What benefits will VoLTE bring to carriers and consumers?

VoLTE will be of benefit as it will enable the use of smartphones, one of our major revenue streams – because it’s convenient for everyone to have the internet in their pocket. Without voice, no one will be carry an LTE smartphone so it’s extremely important for us.

The scheme we are adopting for VoLTE is the IMS solution, and that is not yet ready. It will take some time. The terminals also need to be IMS ready. A possible solution will be circuit-switched fall back, where the terminal will revert to 3G to make a voice call – but we don’t want to go that route; we want to go directly to the IMS part, which is recommended by the 3GPP and all the industry.

What’s wrong with the circuit-switched fall-back route? After all, it has worked in the US?

Circuit switched fall-back has a lot of call set-up delay. Imagine if your mobile needs to change to 2G or 3G before making or receiving voice calls – it takes maybe 15 seconds or more to search for a cell and switch to it. We are worried it will affect our customer perception.

So the disadvantages are primarily to the end user?

Yes. LTE is characterized by very high spectral efficiency and, much higher capacity so if you are carrying the voice over the LTE then you are relieving the 3G network from voice traffic and it will be faster to refarm the 2G and even dismantle its equipment. While we rely on the 2G and 3G networks to carry the voice we are losing capacity there. It will make it difficult to redeploy into LTE. By the end of 2012 we will have voice enabled handsets and the network will be ready?

Do you think LTE will be easier to monetize than 3G?

We were one of the first to adopt 3G technology in 2002/3 and we didn’t see any traffic pick-up until 2006-7. The reason is that there is no ‘killer-app’. I expect the LTE will be the same. There was no clear application for the 3G – even until today – the only thing that came up was that customers started to get smartphones and YouTube and other websites that require heavy throughput came up and that’s why started people started moving to 3G. The trend is now for smartphones and tablets, but before these devices came we didn’t have a real need. So it’s a matter of applications, a matter of the user need. In a few years we will have HD on YouTube so LTE will pick up. There is no killer app, but it might come.

With that in mind, why did Etisalat need to move to now?

The reason we went to LTE is to be prepared and get the know-how – before the traffic comes. To learn how to implement it, to operate it, before we get the heavy traffic. The data traffic is doubling every year and the major contributor is video which is now 50-60 per cent is video. I don’t forecast huge traffic device from customers as the devices are still expensive, but it’s a good opportunity to learn, to optimise and prepare for when the heavy traffic comes. We are charging almost the same as for 2G and 3G, but the dongle is still expensive so not everyone is not willing to go for that yet. It will come down through economies of scale, and soon the packages will be attractive.

What are you hoping to get out of the LTE MENA conference?

It will be very useful to get to know what the other operators are doing, the lessons learned and how they overcome it. That is the benefit of these sessions – so that others will not have to repeat the mistakes that others have committed.

Dr Mohamed Nadder Hamdy, Director Capacity Planner, Etisalat, UAE will taking part in the LTE Executive Think-Tank on Day One of the LTE Conference, taking place on the 29th-30th April, Westin Mina Seyali, Dubai, UAE. Go to the website to register your interest.