25 Sep 2015

Better Spectrum Allocation

It won’t provoke a lot of disagreement to say that India’s cellular spectrum allocation needs improvement. I have some ideas to that end.

First, we need LTE service on all the frequency bands that phones support. Just one phone, the iPhone 6s supports 23 LTE bands:

LTE (Bands 1, 2, 3, 4, 5, 7, 8, 12, 13, 17, 18, 19, 20, 25, 26, 27, 28, 29, 30)
TD-LTE (Bands 38, 39, 40, 41)

India has LTE service on one band (2300 Mhz), and we’re considering starting service on a second band (1800 Mhz). Cramming everyone onto two bands isn’t going to work, anymore than a highway with just one lane for both directions to share is. We need LTE on all bands that any phone in India supports. Even if your phone doesn’t support a particular band, you should be happy about having LTE on that band, because other people will use that band, leaving the bands your phone does support freer for you to use. Imagine how much better things would be if each band has 1/23th the users it has today. Imagine converting the one-lane highway mentioned above into a 23-lane highway.

This will also make it easier to buy phones. You’ll be able to buy any phone anywhere in the world, say if it’s cheaper there, and use it with LTE in India. We’ll also be able to import any and all LTE phones into India and have them work on LTE. As opposed to the status quo where we probably have crores of LTE phones that don’t support TD-LTE (the kind that’s available in India) and so fall back to 3G. This is a waste on both ends: the LTE function of the phone goes waste, and the LTE network gets lower usage than it could, increasing the pressure on 3G.

Second, future auctions or leases of spectrums should be accompanied by the condition that they be used only for LTE or future technologies like LTE Advanced, not old technologies like 3G or 2G.

Third, the emissions norms should also be set in a scientific manner rather than fear-based. Or, they should be set to match developed countries, rather than being roughly one-tenth of what they use in developed countries. If the tower can’t transmit a signal with sufficient power, no wonder calls will drop and the Internet won’t work.

Fourth, the networks should be run by third-parties, not the cellular companies like Airtel or Vodafone. That way, they can serve customers of all carriers. When each carrier sets up a network of its own with LTE and with no or few gaps in coverage, it’s redundant infrastructure, at the cost of billions of dollars for each carrier. It makes as little sense as building different roads for Maruti and Hyundai cars. To put it differently, if you’re an Airtel customer, and your phone doesn’t get a good signal from any Airtel tower, but does get a good signal from a Vodafone tower, wouldn’t you like your phone to connect to Vodafone? Or would you like your call to drop or your Internet to stop working? Or, what if you’re a Vodafone subscriber, and Vodafone has only a 3G network, and you’re in range of an Airtel 4G tower? Wouldn’t you then want your phone to use 4G [1]?

The only way this kind of shared infrastructure can work is if the networks are all owned by third-parties that are legally bound to work in a fair and non-discriminatory manner. Then there won’t be any concern about playing favourites. That is, Vodafone wouldn’t want to use Airtel’s network, because they’ll be worried that the terms will be to their detriment, such as a high fee, or other restrictions. No company wants to put their fate in their competitors’ hands. When it’s a third-party that has no incentive in seeing any particular customer of theirs benefit or lose, all carriers can use them. This can work only if the network is owned and run by a different company, is legally bound to be fair and non-discriminatory, and does not own a stake in any of their customers (or vice-versa).

The network companies can sell bandwidth not just to Airtel, Vodafone, etc, but also companies like Google (disclosure: my employer) and Microsoft. After all, I’m already paying Google for storing my files in Google Drive, and it’s illogical that I have to pay again to access them. Or have a 2TB quota and not be able to use any more than 1GB, my cellular fair use limit. As long as the bandwidth is sold to everyone at the same prices and with the same conditions, this is great.

So, all future sales or leases of spectrum should be to these neutral third-parties. And free. The government shouldn’t try to extract tens of billions of dollars from companies, leaving them with no money to actually deploy a network on the spectrum they’ve purchased. Or having it take years to deploy newer technologies, with the result that most Indians are on 2G, not 3G, to say nothing of 4G. This is a shame. Spectrum is a public good. So, it should be used to benefit the public, rather than the government. How might we do this? Companies should acquire spectrum on demand. They should be able to specify a point on a map, as a lat-long pair, and ask to build a tower there. Then, they should be allowed to use all the bands at that site that don’t conflict with nearby sites. And this would be at no cost. But the company has to build a tower and start serving customers within three months, or they’ll lose the right to those frequencies at that site.

So, instead of spending tens of billions of dollars acquiring spectrum in slow, bureaucratic processes, companies should be able to acquire spectrum on demand, just when they want to build a tower there. And they would be given that for spectrum free, as long it’s used only for LTE and newer technologies, and they agree to serve users of all carriers.

In summary, spectrum should be given free to companies that will promise to be fair and non-discriminatory, and run only LTE on that frequency. And we should have LTE service on all bands phones in India support.

[1] This is similar to roaming, but roaming is a last-ditch option, while I’m advocating it to be the only option.

No comments:

Post a Comment