On Tue, Jul 16, 2013 at 6:40 PM, Joe Touch <[email protected]> wrote:
> Hi, Spencer,
>
>
> On 7/16/2013 1:52 PM, Spencer Dawkins wrote:
>
>> A Transport AD should have a broad understanding of core end-to-end
>> transport topics such as congestion control, congestion signaling and
>> congestion management, control loops and hysterisis, flow control, and
>> middleboxes such as NATs and firewalls.
>>
>
> As I said, this is insufficient. IMO, it focuses far too much on the
> congestion issue and not enough on reliability, connection issues, network
> layer interactions, and application layer interactions.
>
Yes. I must agree with Joe again, having written in my blog about the
misconceptions that many networking engineers and researchers have about
applications last week.
>
> I would suggest rewording it to avoid repeated reference to congestion,
> and include these other issues:
>
>
> A Transport AD should have a broad understanding of core end-to-end
> transport topics such as congestion, control loops and hysteresis, flow
> control, reliability issues, connection issues, and interactions with the
> network layer, the application layer, and middleboxes.
>
There are two words missing entirely that are key:
Queuing and, even more important, Latency
- Jim
