OPEN THOUUGHT
DON SAMBANDARAKSA
Nokia-Siemens dropped me a nice press release the other day on Dtac's "environmentally friendly" base station upgrade and how it had placed an order for 1,500 FlexiEDGE base stations that cut power consumption by 70 per cent. A nice gesture, but would it not be better to invest in 1,500 new W-CDMA stations that use even less power?
Yes, I am sure if the RMS Titanic were in operation today, we would be thinking about fitting more efficient coal-powered steam engines to her. Life is so much easier with the modern aeroplane and 3G in equal measures, I feel.
Only scrolling down, we see that Nokia-Siemens' latest generation of EDGE base stations need 800 watts of power, while their W-CDMA base stations take only 500 watts with the next generations to hit 650 and 300 watts, respectively.
As every data centre manager knows, for every watt of power used at the chip level, an additional two watts of power are needed to remove the heat and keep the data centre from turning into an oven. Most modern base stations are no different, with the bulk of the installation being an air-conditioned mini data centre in a shipping container at the base of the mast.
Only, with even today's W-CDMA 3G installations, it does not have to be that way. In Cambodia, QB's Ericsson base system uses tiny (relatively speaking) cabinets with ambient temperature air-cooled components, with dust filters of course, and yes, QB's CTO assures me that they work fine under load during a hot Cambodian summer.
Once power consumption is lowered to a point that allows air-cooling, a lot more power can be saved.
Those calculations are even before density needs to be considered. With a limited amount of bandwidth per antenna, more data consumption will need a greater number of overlapping cells to provide users with access to YouTube (if their state-owned telecom company does not ban it, like in some under-developed countries) and other forms of rich media. The same is true, even more so, of 4G.
I often get asked what 4G is, and why would anyone need it. To the first question, I say that 4G is a different radio modulation system, Orthogonal Frequency Division Modulation, features more than one antenna (multiple in, multiple out, or MIMO) and is based on software defined radio.
On the latter question, I tell them that on the one hand it is much faster while also being much, much more efficient.
W-CDMA has a spectral efficiency of 0.8 bits per second per hertz. OFDM can reach up to 3.5bps/hertz though early WiMAX implementations (which, like 4G LTE uses OFDM) delivers around 2.5 in the real world.
What this means to the user is that for the same amount of frequency, a telco can give four times higher data rates or fit in four times as many users for the same cost. One of the biggest costs today is frequency, or rather the telecom licences, and thus being able to use frequency more efficiently can, depending on the licensing regime, save costs.
Which brings us back to an interesting issue today that is affecting Thailand. In the Nokia-Siemens press release, the fine print clearly stated that Dtac operates 800 and 1,800 frequency bands under a licence by CAT.
Which brings us to 3G. CAT is effectively saying that since it "owns" the spectrum for AMPS (and consequently CDMA), it can be re-farmed for WCDMA 850.
But if capacity quadruples every generation, a merely sufficient amount of bandwidth in 1G days could easily be 16 times what would have been needed for the original purpose two generations ago. Or 256 times going from 1G analogue to 4G OFDM. Rule of thumb thinking of course, and the truth is that older carriers are narrower than W-CDMA and Wi-Fi, but the logic is still there.
The same is also true of television and radio. Thailand is still arguing (and setting up) UHF analogue stations to help The Powers That Be brainwash the population with propaganda. But DVB-H digital TV is much more efficient. Digital radio (DAB) can fit an almost infinite number channels in, and with much greater clarity too. Plus, there is all that spectrum left over for analogue systems that begs to be re-farmed.
The question that whoever is in charge should be asking themselves is whether WCDMA 850 should be CAT's and if it is in the country's best interests to effectively grant it a perpetual licence, perhaps wasting bandwith and sub-optimally allocating frequencies.
Again, the awakening came from Cambodia. QB's CTO John Kjellemo said that he was eyeing a combined 850/2,100 network for Cambodia. The 2,100 band would be suitable for small cells in densely-populated cities such as Phnom Penh, and the 850 super cells with up to 120km reach would be perfect for bridging the digital divide and providing Internet access to Cambodia's rural population.
So, dear NBTC, would it not be a good idea to sort out the mess by consolidating frequencies and somehow ending these perpetual leases and allocating them in a much more efficient way? How about four telco licences, each with a combined LF/HF set for rural and urban coverage? 850 and 2,100, or 900 and 2,100, or perhaps even 850 and 1,900 for WCDMA, which could take advantage of US handsets quite nicely.
Anyone with an 850-only licence, as Dtac seems to be getting, will struggle with affordable handsets. Anyone with a 2,100-only licence will struggle with coverage and the need to pepper the landscape with lots and lots of small cells (which means more base stations and more power and more network complexity), as is the case in the West today.
Giving out an 850 licence today and then letting the telcos squabble and fight for a 2,100 licence later may be a "free market" mentality, but sometimes a bit of proper planning can help a country develop, provide universal access and provide Internet access to those who are today on the other side of the digital divide.
Not doing anything and not taking a lead is leading us down a convoluted regulatory path that is by far one of the messiest the world has seen.
Prev
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Next
