[log in to unmask] wrote:
> Hi all,
> Just to add to this discussion. The technology required to be able to
> recognise traffic on an application level and adopt policies based on that,
> as well as where trhe traffic request is coming from etc. has been around
> for quite some time. It is called Layer seven switching and has been adopted
> by a number of ISP's. Normally they use it to throttle bandwidth requests
> for high volume uses or for particular traffic types such as music/video.
> One particular American ISP uses this to ensure that new users get good
> bandwidth, they then shuffle loyal customers onto increasingly smaller
> The technology is certainly out there and being used.
Sure, traffic throttling based on ports, single downloads from multiple
servers (such as p2p) and traffic shaping, such as in the case of
encrypted voip packets, is being used today - no great shakes there.
This is fundamentally different to what John is describing though. John
is describing a mechanism for differentiating between http packets that
contain html generated on a server by interrogating a database, and
those containing html that were stored statically on a server. That is,
an application to determine not merely the type of data within the
packet, but the manner in which data the within the packets were generated.