From Mohamed Awang Lah
I read with great interest – and concern – a recent report on the 5G rollout and the stakes being offered by Digital Nasional Berhad (DNB) to the telecommunication companies.
The most interesting part was the “threat” to open the market. Does it mean foreign players (with large funds in hand) will be allowed to come in?
There is no doubt that 5G is important for the nation. It should have been made available much earlier.
Technology is not the issue. But we have been a laggard not only in 5G services, but in high-speed broadband in general.
This is not due to any lack of effort on our part but it has more to do with having the wrong strategy and priority.
Whatever it is, there are some fundamental issues that must be clear:
a) DNB plans to cover 80% of populated areas by 2024. But what exactly constitutes “populated areas”? How many sq km?
At one time, I did ask the statistics department the same question, but it did not have an answer.
The size will determine the number of towers and, hence, the eventual cost. The number of towers also depends on the spectrum.
Higher spectrum frequency means shorter radius of coverage and more towers are needed to cover the same areas.
How about rural areas? Will DNB use the 700MHz spectrum with a larger coverage area per tower (hence, a lower number of towers) but at a slower speed, like 4G? Are we giving second grade 5G for the rural population?
b) The price of 20 sen per gigabyte (GB) has been repeated many times. But what was the basis?
We know DNB plans to lease bandwidth to 5G operators at RM30,000 per gigabit per second (Gbps) per month.
Theoretically, with one Gbps, we can transmit one GB of data in eight seconds. That means we can transmit 10,800GB per day at a cost of RM1,000 or 9.26 sen per GB.
Obviously, the link will not be used 100% all the time. With 50% utilisation, the unit cost would be doubled.
If it is 10%, it would be 92.6 sen per GB. It depends on the usage behaviour of the users. It varies with locations, too.
So, how was RM30,000 per Gbps per month translated to 20 sen per GB? What are the assumptions?
Fibre optic infrastructure
We seem to approach various initiatives in isolation rather than finding a common factor, such as infrastructure, which can be shared for different initiatives.
When we talk about 5G, we tend to ignore 4G and wired broadband services.
We forget or ignore the need for a common fibre optic infrastructure which can be used for all. 5G and 4G will not work properly without fibre.
Why don’t we start with fibre first, then use the fibre to improve 4G and, at the same time, rollout 5G services in certain areas.
We don’t need to build 5G for the whole country – it would be too expensive.
The same fibre network can be used for FTTx, which can be used for schools, universities, residential and commercial properties, as well as other premises.
With fibre, we mean dark (unlighted) fibre, not bandwidth through fibre.
Unit rate is per link or per km, not per Gbps. It is up to the service providers to light up the fibre to deliver any bandwidth.
It is the way to future proof the data transmission cost.
One of the reasons for our laggardness is the lack of competition in retail services. We tend to award contracts for last-mile services.
In a way, DNB is providing last-mile service for 5G. It is the umbrella for retail service providers. It is a monopoly.
Competition among retail mobile providers will then be stifled. What we need is full competition at the last-mile so that end-users will have more real choices.
Certain bands of spectrum can still be shared, other bands can be exclusive.
It is fine to provide wholesale for shared infrastructure such as fibre, towers and poles. With economies of scale, the unit cost can be reduced significantly for the benefit of the end-users.
I hope we are not adopting the wrong strategy again.
Mohamed Awang Lah has been involved in internet services since its inception in the mid-1980s. Before his retirement, he was the CEO of Jaring Communications Sdn Bhd.
The views expressed are those of the writer and do not necessarily reflect those of FMT.