DAB sounds worse than FM

Planning for the future of digital radio
 
Quality of BBC Internet radio streams to overtake DAB

Home
 
DAB/DAB+
 
Digital 
Satellite
Freeview
 
DRM
 
Internet 
Radio
Tech
 
Software
Downloads
DAB 
Samples
Newsletter
 
Contact
 Us
Introduction to DAB
Incompetent adoption of DAB
When will FM be switched off?
BBC DAB Multiplex
Digital Radio Bit Rates
Wasted DAB Capacity
DAB Around the World
Design of DAB
DAB vs DAB+ technology
T-DMB vs DAB+
Coverage Maps
DAB Summary
 
DAB Radios
DAB CD Portable Stereos
DAB Personal Radios
DAB Micro Systems
DAB Clock Radios
Digital radio via satellite
Satellite Receivers
UK satellite radio bit rates
UK satellite HDTV bit rates
UK satellite TV bit rates
Internet Radio
Broadband Internet Radio
Audio Advice
Aerials
MPEG Audio Coding
Bit Rate vs Audio Quality
MP2 vs AAC+
Audio Processing
FEC Coding
OTA software upgrades
COFDM
Analogue vs Digital Radio
Bandwidth
RF Carriers
Sampling
RF Antennas
Links
         
 

A way to improve speeds on iPlayer downloads


6th March 2008

I've had an idea on how to improve the speeds on the download version of the BBC iPlayer, which has just posted a new record number of daily users with 660,000 people using the iPlayer on 24th February, which is a pretty phenomenal rate of growth considering it was only launched a couple of months ago, and which in my opinion vindicates both mine and more importantly Fru Hazlitt's view that the future of digital radio lies on the Internet, but I'll be posting a long article on that shortly with my post-mortem on GCap dropping its DAB bombshell a couple of weeks ago.

Back to the iPlayer. The ratio of people using the streaming to the download version of the iPlayer is currently 8 to 1, and having used both, it's thoroughly understandable why that is, which is that it takes too long to download a programme file, especially at peak times -- at off-peak times it can be very fast. The problem is that the download version of the iPlayer uses a P2P network, where users receive small chunks of the programme files from other users' computers who have that file on their hard disk, and eventually once all the chunks of data have been downloaded the file is playable. But the P2P version is currently in a chicken and egg situation where because the iPlayer has only just launched there's only a relatively small number of 'peers', i.e. people with that file on their computers to download from, and that is having the effect of reducing the download speeds so it's taking a long time to download files. So a way to improve upon this situation woudl be for the BBC itself to store the programme files on a relatively large number of servers which are assigned different IP address, so they would look to all intents and purposes like normal users' computers, and this would artificially increase the number of peers that users can download the programme files from, thus increasing the download speeds.

The BBC would save a lot of bandwidth using this method, because the streaming version uses unicast distribution, which means that the streams all have to be delivered in parallel and in real-time, so the peak bandwidth demands will be huge (the streams are using bit rates of 550 kbps each and as mentioned above, there's up to 660,000 people using the iPlayer per day, with the large majority using it in the evening), whereas if the BBC provided 'artificial' peers to 'seed' the P2P network a population of other peers would then emerge, and the BBC wouldn't have to shoulder the whole bandwidth demand itself like it has to on the streaming side. Also, over time as more people are using the download version the BBC could reduce the number of artificial peers they use to seed the network because there should eventually be a sufficiently high number of peers carrying each programme file, so they would be able to save even more bandwidth.

The ISPs would also benefit from this, because the download version eliminates the real-time requirement that the streams have, which would reduce their peak bandwidth requirement, because packets can be delayed on the download version as there's no requirement for them to arrive at the user's computer by a certain time limit as there is on the streams.

Ultimately, the P2P version will come into its own, and those that have written it off will have to eat their words, because eventually there will be a large number of people using the download version, so download speeds will improve substantially due to the increased number of peers holding the programme files, which should in turn lead to more people using the download version because the lengthy download times issue will have been vastly improved. The download speeds will also improve substantially due to increases in people's broadband connection speeds as the UK migrates from typically ADSL to ADSL2+ over the next few years. Don't get me wrong, I think the majority of people will use the streams because there's absolutely no wait and it saves having to install a piece of software, but I think the 8:1 ratio at the moment will change substantially over time, or at least it could change substantially if the download speeds improve markedly in the near future, and if not then a lot of people may try the download version but the download speeds put them off and they'll go back to streaming and never try the download version again. As it's in the interests of the BBC for as many people to use the P2P version as possible because it would vastly reduce the amount of bandwidth it requires compared to streaming, the ball's in their court...