Dongle users dingled

Saying your 3 data dongle works great in the centre of london and the home counties, is not really saying anything.

I can't imagine there being single telco's whos dongles don't work in those areas.

The real test is the rest of the country, and my experience of 3 is that its completely crap.

I use my 3 dongle most weekdays - great service now the cells have been upgraded, even in the centre of a steel structure building. I get between 0.8-1.5 Mbs. Location City Of London, EC2M.

I also use it on the train from London Bridge all the way to Surrey. Works great for me, even on the train.

Its patently obvious when you use the mobile data networks that the bandwidth is oversold.

The statements made in this article are supported by real life experience of fast and then slow connections at peak times, rather like some less than wired scrupulous broadband operators suffered from a few years back.

The main reason why bandwidth is oversold in the UK at least is to try and re-coup the grossly huge sums of money the telco operators paid a few years back for 3G licenses lining the UK governments pocket in the process. So it seems our government screwed up mobile broadband as well as the other litany of disasters so self evident in their copy book.

@Matthew
you commented on Three's mobile network being upgraded for the extra control channels, where exactly were you when you managed to actually get a Three signal?

Ok, I agree that the server may get a bit backed up... but the real reason to drop the dongles is outright obvious...

Check the prices and bandwidth cap between dongles and smartphones.

While the smartphones usually have no cap on usage the dongles charge for every little bit downloaded...

It's economical the reason, not technical...

The telco's fixed the bandwidth problem ages ago. When the iPhone first came out, pretty much every telco read the writing on the wall and fixed their backhaul to the cell towers.

Uptil then, many towers only had a 2Mb/s connection, as that would allow around 250 simultaneous phone calls - a limit many towers never came close to. The telco's actually implemented plans (which uptil then had been sitting on the shelf labeled "Good ideas") and ran fibre to at least the base stations. Most towers are connected to the base stations by microwave links - these were upgraded to 10Mb/s, which is adequate for most 3G use. So the bandwidth problem was solved some time ago.

What the network designers got wrong was average session time. Basically, they decided that there would be two typical use cases - some one who connects to check there email and do a little web browsing, with a connect time of around 15mins and someone - such as iPhone users - who would be permanently connected, with mail updates happening every few minutes and web browsing or applications accessing web sites pretty much continuously. The data networks were built to handle these sort of loads with capacity to spare.

The problem was that while the network engineers had solved the issues they had thought about, they had overlooked major changes to how mobile phones send and receive data.

Every time a phone or dongle connects, the network has to perform a whole series of actions. It has to authenticate the phone and decide whether they have a data plan and have sufficent credit to allow connection. It has to assign an IP address to the phone. These and other tasks are performed by specialised servers. In addition, the various 3G protocols require special radio channels to perform call setup. The number of control channels a cell tower provides depends on how many calls are expected at any one time.

What the engineers go wrong was how many times a phone would connect to the network and therefore what sort of load these servers have to handle. The number of control channels was also inadequate and could not handle the load from smart phones.

The problem was the engineers who designed the 3G radios in the phones were getting a little clever. They were worried about battery lifetime, and one of the tricks they had developed was to only turn on the 3G radio when it actually in use. This means for data that instead of connecting and perhaps holding the connection for many hours, if not days, they only hold the connection for seconds.

Note that dongles don't cause any of these problems. They setup a connection and hold it for a long time. It has been the shift from using dongles to using smart phones either for direct data access or for 'tethering' a PC which has caused all the problems for the telcos.

When the iPhone wants to access the network, it turns on the 3G radio, establishes the connections, download the email or web page it wants, and then drops the connection and turns off the 3G radio. Total time - around 1 or 2 seconds. A few seconds later it will repeat the process for the next data access. This rapid turning on and off of the radio saves the phones batteries but causes the control channels to be clogged and the radius and dhcp servers to be overloaded.

So, what the telcos need to do, and some (three?) have started to do, is retune the cell towers channel settings to provide extra control channels and upgrade the specialised servers to prevent bottle necks during call establishment.