Broadband News
News, views and analysis
Analysis: Broadband Quality Study
11 Oct 2009 | 14.18 Europe/London
The Global Broadband Quality Study produced by Oxford's Said Business School and the University of Oviedo generated a number of headlines, mainly around what place in the "league table" was occupied by the UK. The detailed presentations of the report contain further interesting material which we have been digesting.
It should be pointed out that the Study, sponsored by Cisco, is not published in the form of a detailed academic report but as a set of presentation slides. We cannot therefore be sure of the methodology used for example to decide which speedtests were from rural users and which were from a given city.
The UK shows up well in the "Digital Divide" analysis, with a very small difference between the broadband quality score for cities and areas outside cities. The UK average score is 30.77 with "areas outside cities" scoring over 30 with similar values to Birmingham, Glasgow and London. We have some reservations about the separation of data by geographical area, but on the face of it the UK data looks fairly uniform - unlike the situation in Russia and the Baltics where cities score much higher, or Sweden and Iceland where areas outside cities score higher.
The cost of Next Generation Access is often cited as a reason for the UK's lack of progress towards high speed low latency fibre systems, so it is interesting to see the broadband scores compared with GDP in the report. On this measure the UK sits in 22nd place between Austria and Spain and above the US, Canada and Australia with a relative score of 84 BQS per 100,000 GDP per capita. This would suggest that the UK is doing OK relative to its wealth but is certainly well behind the high performing countries like Portugal, Sweden, Estonia, Slovakia in Europe and the inevitable Japan and South Korea in the Far East. These countries achieve 150 to 240 BQS per 100,000 GDP per capital and are clearly spending more of their wealth on broadband relative to the UK.
Boris Johnson may like to ponder that London comes in 107th in the world's top cities by broadband quality, whereas Birmingham and Glasgow just edge into the top 100.
Download speed is the common obsession of league tables but this study also takes into account upstream speeds and latency (ping time). The UK is 18th in the latency league table, well ahead of the US (24), France(28) and Germany(37).
On upload speed the UK looks to be 40th, although it is possible to confuse UK, USA and UAE on the presentation slide graph. This is an area where the UK usually lags, with the bulk of ADSL connections offering less than 0.5M upstream and little take-up of the faster business grade services. BT's FTTC and further tweaks to VM's cable services should see this figure improve in future.
We would like to see some clarity on how a user is allocated to a location, for example does a rural user of a city based ISP get categorized under that city ? We should also point out that the data is derived from speed test sites where users have themselves elected to go and run tests - does Sweden's download speed average of nearly 15 Mbits/s accurately reflect the average for the population or is it skewed by the 10% of Swedes with fibre connections spending a lot of time admiring their results on speed testers ?
It should be pointed out that the Study, sponsored by Cisco, is not published in the form of a detailed academic report but as a set of presentation slides. We cannot therefore be sure of the methodology used for example to decide which speedtests were from rural users and which were from a given city.
The UK shows up well in the "Digital Divide" analysis, with a very small difference between the broadband quality score for cities and areas outside cities. The UK average score is 30.77 with "areas outside cities" scoring over 30 with similar values to Birmingham, Glasgow and London. We have some reservations about the separation of data by geographical area, but on the face of it the UK data looks fairly uniform - unlike the situation in Russia and the Baltics where cities score much higher, or Sweden and Iceland where areas outside cities score higher.
The cost of Next Generation Access is often cited as a reason for the UK's lack of progress towards high speed low latency fibre systems, so it is interesting to see the broadband scores compared with GDP in the report. On this measure the UK sits in 22nd place between Austria and Spain and above the US, Canada and Australia with a relative score of 84 BQS per 100,000 GDP per capita. This would suggest that the UK is doing OK relative to its wealth but is certainly well behind the high performing countries like Portugal, Sweden, Estonia, Slovakia in Europe and the inevitable Japan and South Korea in the Far East. These countries achieve 150 to 240 BQS per 100,000 GDP per capital and are clearly spending more of their wealth on broadband relative to the UK.
Boris Johnson may like to ponder that London comes in 107th in the world's top cities by broadband quality, whereas Birmingham and Glasgow just edge into the top 100.
Download speed is the common obsession of league tables but this study also takes into account upstream speeds and latency (ping time). The UK is 18th in the latency league table, well ahead of the US (24), France(28) and Germany(37).
On upload speed the UK looks to be 40th, although it is possible to confuse UK, USA and UAE on the presentation slide graph. This is an area where the UK usually lags, with the bulk of ADSL connections offering less than 0.5M upstream and little take-up of the faster business grade services. BT's FTTC and further tweaks to VM's cable services should see this figure improve in future.
We would like to see some clarity on how a user is allocated to a location, for example does a rural user of a city based ISP get categorized under that city ? We should also point out that the data is derived from speed test sites where users have themselves elected to go and run tests - does Sweden's download speed average of nearly 15 Mbits/s accurately reflect the average for the population or is it skewed by the 10% of Swedes with fibre connections spending a lot of time admiring their results on speed testers ?