[c-nsp] to shape or not to shape

[Mikael Abrahamsson]

On Sat, 9 Oct 2010, Roger Wiklund wrote:

> So, as we shape, as long as we have buffers, we will never see any tail 
> drops, as we will just delay the packets until we send it, correct?

Buffers are not infinite, so you might still see tail drops.

> If this e1 is 100% utilized, we will get tail drops when the buffers are 
> full.
>
> So my question now, what if the shape the e1 to 1984, we will still
> have the full speed, but we shape, and thus avoid tail drop, and just
> delay the packets instead. I'm thinking we avoid TCP restarts etc etc.
> pros/cons, or am I wrong about the whole thing? :)

Tail drop is basically dropping a packet because the buffer/queue is full.

The only difference between your qos policy and fifo is that fifo will 
drop when the single queue is full, your qos policy has several queues and 
thus will "tail drop" differently depending on what queue the packet is 
in.

So when you apply "qos" to your e1 then it won't do fifo anymore but 
instead it will use your policy to decide what packets to handle in what 
way.

-- 
Mikael Abrahamsson    email: swmike at swm.pp.se