I spent much of last week sorting out a network and ADSL connection in a small rural Buddhist hilltop monastery in the Northeast of England. This should have been a straightforward job, but the exercise proved that connecting the rural population is as much of a challenge in the developed West as in developing parts of Asia.
The ADSL connections were a case in point on how the traditional way of selling best effort, means wildly different user experiences depending on how far you are from the exchange. Checking with BT, the post code said we can expect between 1 Mbps to 2 Mbps, and this is at the same “best effort” price as someone in town would be paying for 15 to 20 Mbps. That is not a good start. Still, at least the area has ADSL but 1 Mbps is not quite broadband in this day and age.
Worse, it soon was apparent that the connection was extremely flaky. After hours of troubleshooting and removing all other devices, a call to BT confirmed that the exchange had a problem and that it would be fixed by the next working day.
One wonders how much connecting this group of houses at the top of a hill would have cost. There are about four or five homes in the area at best, so one exchange for four to five accounts paying £18 (€21) a month hardly makes economic sense. Hence the exchange was situated further away and thus relegated to around 1 Mbps.
Once the new exchange was installed, the raw line speed went up to almost 5 Mbps, much faster than predicted, with speedtest.net showing real-world downloads at just short of 2 Mbps. All of this highlights the broken model of wired internet in rural areas, but the wireless part was not much better.
How about, I suggested, using a 3G connection instead, given the very low speeds and terrible reliability of copper? Well, despite the UK being a first world country, the data speeds on T-Mobile/Orange were quite an eye-opener for all the wrong reasons. Apparently the area does not even have 2.5G EDGE coverage, let alone 3G. With my phone showing a long-forgotten “G” icon on the top right, it was clear that this was old-school 2G GPRS.
Much has been said of the need for a cellco to have a 2G network to fall back to, but in this case, it is useful only for voice. The lack of even EDGE speeds made it totally useless for tethering to a laptop and even on a phone, apps or web browsing required the patience of a saint (which, given that it was a Buddhist monastery complete with meditation hall, was probably a divine hint). The “unlimited” data package that we had was worth naught given the lack of decent speed.
Hutchinson 3 proved more useful. There was no indoor coverage at all, but perching the phone precariously on a high window and setting it to hotspot mode yielded just over 1 Mbps, only slightly less than ADSL.
After a few days of fixing up LAN cables, swapping routers and switches and getting some serious hand exercise re-crimping RJ-45 cables, I left with a realization that the only model that had the incentive to provide access to this small rural area was by H3G. The company does not have any unlimited packages and thus there is an incentive to get people to use more data and to provide faster infrastructure.
T-Mobile provides an unlimited (fair use) package and thus it is in their interests to have people use as little as possible. Hence, providing just GPRS coverage is enough, why provide 3G coverage if it will result in no extra revenue? Outside of London and large cities, EDGE is interesting in its absence with networks either on HSPA 3G or GPRS.
And as for BT’s Openreach copper? Well, they are probably losing money given the sheer lack of houses in the area, but at least they were committed to their USO targets and fixed the flaky exchange within a day. But at what cost?
Wireless broadband via 4G LTE cannot come soon enough. When the UK issued its 4G trial licenses in the Cornwall in the Southwest earlier this year, the target was communities just like this, [which] lack the economies of scale for copper. BT said that the technology was planned to give the final 10% of the population coverage with at least 2 Mbps speeds - hardly LTE, but given that neither 3’s 21 Mbps HSPA+ network nor the BT copper ADSL actually got to 2 Mbps in real world use, it is a decent goal to aim for.