[Winpcap-users] Re: Re: Re: Delay and jitter

[Guy Harris]

Antonio Ocampo wrote:
> I'm trying to measure one way delay.

Again, is that application-to-application (meaning that the time the 
packet is sent is the time that the application on the sending machine 
sends it, and the time the packet is received is the time the 
application on the receiving machine sees it), or 
network-adapter-to-network-adapter (meaning that the time the packet is 
sent is either the time the first bit of the packet or the last bit of 
the packet is put onto the network, and the time the packet is received 
is either the time the first bit of the packet or the last bit of the 
packet is received by the receiving machine's network adapter), or 
something else?

The first two are different; the first is much easier than the second. 
The second might be impossible without using specialized network 
adapters on both machines, and specialized software with those adapters 
- and I don't know whether those even exist.

All you'd get with libpcap/WinPcap (or, I suspect, any *other* capture 
software) would be a value somewhere between the 
application-to-application time and the 
network-adapter-to-network-adapter time, which would, for example, 
include a significant amount of networking stack delay.