Friday, January 29, 2010

'Net neutrality: towards a co-regulatory solution' released by Bloomsbury

Its on a Creative Commons download- and the Preface thanks all of you as readers, and many of you by name - so thank you:-) 
If you enjoy reading the bits you print or read onscreen, there is an introductory offer on the hardback, either contact me or look on the Bloomsbury website.
Happy reading - I will post the gentler or most revelatory reviews and link obscurely to the rest...
STOP PRESS: if you wait to download until Monday, you can get the cover and see it in all its glory.


Rob Frieden said...

Hello All:

Professor Marsden had made a big, I mean major contribution to the literature on a complex topic. Thanks for making the manuscript available for reading.

Rob Frieden said...

Professor Marsden has made a significant contribution to the literature and to the debate on network neutrality. This issue has had a U.S emphasis, but it applies to the global Internet

Mark Taylor said...

Your section on QoS seems a little light. It strikes me there is a great deal of debate to be had here. I would suggest that the overwhelming majority of QoS schemes implemented today make absolutely no difference to the quality levels experienced between the various QoS tiers. And that is because most networks are engineered with bandwidth appropriate for the demand that the network operator is experiencing. Capacity planning is well understood by network operators and taken very seriously.

So why the talk of QoS? Partly it is by people that do not understand the relative cost of implementing QoS (not a trivial cost) compared to an appropriately sized network. Some people have referred to this as maintaining an over-supply of bandwidth. I disagree. It is the appropriate level of bandwidth for the demand. And it is a cheaper alternative - in all parts of the network; backbone, middle mile, last mile.

The danger with simply assuming that QoS is needed is that it may be used not to prioritise packets that needs, say, lower latency. But to degrade packets for application owners do not pay.

And it seems to me that is such an important issue that far more debate should be had on the merits and costs of QoS versus bandwidth.