Posted by Rick Merrill on 02/14/07 14:31

Gene E. Bloch wrote:
> On 2/09/2007, Gene posted this:
>
> [...]
>
>> I was not aware that the internet protocol had changed, I thought that
>> everything was still in multi-packets, and over different paths. It's
>> probably
>> silly to keep the old security protocol, given that the net is no
>> longer a national
>> security thing. Guess packet verification is no longer really all
>> that necessary.
>> Geesh, I have not seen a parity check in years:-)
>
> [...]
> "Rick Merrill" <rick0.merrill@NOSPAM.gmail.com> wrote in message
> news:Ws2dnb9h_s5DclHYnZ2dnUVZ_uednZ2d@comcast.com...
>
> [...]
>
>>> Well you got the transmission by packets right - but the truth is
>>> that the odds are HUGE that each packet will in fact take the same path.
>
> [...]
>
> Just a few comments here for clarity.
>
> 1. The transmission by packets is NOT for security, Gene. It is because
> it is hard (impossible?) to transmit a whole lot of bits in a huge
> single block without any errors. So you send a lot of small blocks, each
> with error checking data. If the received block is incorrect, the
> receiving stations asks the sender for a retransmission. This has been
> the norm for decades (e.g., xmodem, zmodem).
>
> 2. A corollary of that is that, in the multi-connected net of today,
> these packets don't all have to take the same path to get to the final
> destination. This would happen for sure if one of the intermediate nodes
> failed during a long series of blocks. I have no idea how rare it is for
> one collection of data to get to you by varied paths, but I would tend
> to agree a little with Rick Merrill.


hey, you made my day ;-)


I would just say it's usual for the
> data to all use the same path, but not at all impossible for paths to vary.
>
> 3. I don't chink the Internet Protocol has changed. I *do* think your
> understanding of it is changing :-)
>

[Back to original message]