As more and more providers move away from the fixed speeds ADSL products, onto the newer fully rate adaptive ones, so to does the number of faults increase.
To many this seems to go against the way the product should work, as it should adapt to current conditions and ensure it is operational. Whilst true, it is only the line conditions at the time of synchronisation that apply, but more importantly, the line is operating as close to its operational limits as much as possible.
Traditional fixed speed services always had cautious line limits set against them and thus had a certain amount of breathing space, but rate adaption leaves little or none.
It is no wonder there can be a large volume of intermittent faults seen from customers moving to the product. For the most part, disconnections should be expected, as the technology isn't designed to be a fully nailed up and robust. It is after all quite often, working over very old copper cabling.
ISPs are often presented with a problem in that they can only test a connection at one particular point in time. This is often futile and it does not say anything for the line history, leaves us in a situation were no fault is found and we can only ask the customer to monitor.
Sometimes, faults can go on for weeks without any sort of fix and in rare cases, months. Over this time, we collect data from the various tests performed and either a pattern emerges, or we notice things that go on to show either way that BT need to be involved, or a simple lack of understanding on the customers part, in terms of the product or their broadband setup.
BT Wholesale are due to launch a new tool to Broadband providers as a means to help combat this problem. Known as the RRT tool, or Reactive Repair Tool, it will allow a provider to look back at the past 14 days of customer service at the following key piece of data.
The speed at which the line synchronises
The amount of signal relative to the noise on the line
The amount of noise on the line, impact by the quality and length
The mean time between errors that occur on the line
The number of seconds the line is active in a day
The number of times the line resyncs per hour
In each case, it provides the minimum values, maximum values and the average seen over the 14 days. Additionally, a computerised assessment is performed of the test results which can help identify a different fault type to that originally reported or being investigated.
In all, this is a huge step forwards. It reduces the number of times data needs to be collected before a trend can be seen, to the point this may be spotted from the start. It wont assist in all cases, but will ensure that focus is applied in the right places.
Adding to this the recent improvements documented in Recent Improvements to IPStream MAX on the comms blog and with more to come, the experiance is set to get better.