Re: [PCN] PCN edge behaviour experiment

[Tom Taylor <[e-mail]>]

Good points. That means the experiment doesn't have to run very long to 
meet the first two objectives. I do have one concern that if you 
introduce artificial traffic then maybe you don't get the same outcome 
that real traffic gives you. I figure it's inevitable that the benefits 
of PCN with real traffic in a real network will be less than our 
simulations show, if only because real networks tend to be over-provided 
with capacity.

On 20/03/2012 4:52 PM, Michael Menth wrote:
> Hi Tom,
>
>
> Am 20.03.2012 21:18, schrieb Tom Taylor:
>> I am making what I trust will be the final revisions to the edge
>> behaviour documents in response to IESG comments. Amongst other
>> things, this is the text I propose to add in the introduction to
>> justify the Experimental status of the documents. The text will be the
>> same for CL and for SM, following on Ruediger's observation that both
>> behaviours are valid in different contexts. Comments are welcomed.
>>
>> ---
>>
>> This document describes an experimental edge node behaviour to
>> implement PCN in a network. The experiment may be run in a network in
>> which a substantial proportion of the traffic carried is in the form
>> of inelastic flows and where admission control of micro-flows is
>> applied at the edge. For the effects of PCN to be observable, at least
>> some links of the network should be running near or at capacity.
> Better: "the aggregate rate of admitted flows on some links should come
> close to the bandwidths of these links."
>
>> The amount of effort required to prepare the network for the
>> experiment (see Section 5.1) may constrain the size of network to
>> which it is applied. The purposes of the experiment are:
>>
>> - to validate the specification of the CL [SM] edge behaviour;
>>
>> - to evaluate the effectiveness of the CL [SM] edge behaviour in
>> preserving quality of service for admitted flows; and
>>
>> - to evaluate PCN's potential for reducing the amount of capital and
>> operational costs in comparison to alternative methods of assuring
>> quality of service.
>>
>> For the first two objectives, the experiment should run long enough
>> for the network to experience sharp peaks of traffic in at least some
>> directions.
> These peaks could be intentionally caused for the sake of the experiment
> so that the effect of admission control is visible.
>
>> It would also be desirable to observe PCN performance in the face of
>> failures in the network. A period in the order of a month or two in
>> busy season may be enough. The third objective is more difficult, and
>> could require observation over a period long enough for traffic demand
>> to grow to the point where additional capacity must be provisioned at
>> some points in the network.
> Also failures could be intentionally caused for the sake of the
> experiment. Capacity shortage on backup paths could also be planned for
> so that the effect of flow termination is visible.
>
> Best wishes,
>
> Michael
>
>> _______________________________________________
>> PCN mailing list
>> [e-mail]
>> https://www.ietf.org/mailman/listinfo/pcn
>