Possible Mistake 119/78

[Despaxion]

In this paragraph on Page 78/119 I'm not sure but as far as I know, purely because I supply and install Patton VoIP gateways, that VoIP does use UDP in most cases. I might be wrong, thanks for a great book.

So if you’re using Voice over IP (VoIP), for example, you really don’t want to use UDP,
because if the segments arrive out of order (very common in IP networks), they’ll just be passed
up to the next OSI (DoD) layer in whatever order they’re received, resulting in some seriously
garbled data. On the other hand, TCP sequences the segments so they get put back together
in exactly the right order—something UDP just can’t do.

[lammle]

Hi, thank you for writing.
You very well might be right that the VoIP you are using is UDP, and heck, Cisco probably does as well in most cases.

However, to be clear here, I was somewhat "quoting" Cisco for the CCNA objectives.

They are refering to how TCP would work better then UDP in some applications, such as VoIP because of the sequencing.

The CCNA objectives are pretty light when it comes to VoIP, but how TCP and UDP works is more important to understand here.

Thanks for writing.

Cheers!
Todd Lammle

[johnl]

I agree with the original position. The reason you do not want to use TCP is that the latency introduced by the retransmission of failed packets is of greater detriment than the artifact caused by the broken/missing packet. Most real time media streams use RTP (Real Time Protocol) which handles time stamping packets. Most codecs have error concealment strategies such that when they are missing or have received a broken packet they will apply the strategy (most simply repeating the last sample, or sometimes interpolating between samples) rather than cause an buffer underflow. It's disappointing we have to remember the stock Cisco answer rather than them correcting their view.