[CALUG] DoD/OSI Layer 2, 3 and 4 in the real world -- WAS: open ports

Rajiv Gunja opn.src.rocks at gmail.com
Thu Sep 15 17:54:53 EDT 2011


Ok Mr Bryan the IEEE guru, please read what I wrote.

----
Network driver does not care if the packet is from a telnet application or
KDE application or VNC or java program, its function is to deliver the said
data to the end client and guarantee (tcp)that the packet reaches the
destination without corruption.
----

In that above statement, am I talking about UDP or stateless packet?

-GGR

--
Rajiv G Gunja
Blog: http://ossrocks.blogspot.com


On Thu, Sep 15, 2011 at 17:28, Bryan J Smith <b.j.smith at ieee.org> wrote:

> Thanx, I was hoping I'd find someone who agreed.
>
> There are so many, common solutions aspects of exchange that do not.  Ping
> is a type of ICMP (IP protocol 1).  The type is contained in the header of
> the IP packet (layer 3) itself, and does not use a transport.  There are
> several others involved with PXEBoot'ing, which is very, very important when
> you're trying to netboot systems.  And I can state countless other, "real
> world" examples too.  And that's before I dive into security aspects.
>
> Understand I dove into this because I run into individuals at clients that
> not only think _exactly_ like this, but that one individual will undermine
> my trying to help the rest of their team.  It's one thing to get the dumb
> stare like many do when we walk into Radio Shack, that's understandable as I
> don't expect everyone to know everything, but only where their experience
> has been.  But it's not about not following, it's about _undermining_ the
> knowledge some someone is trying to share it, and help you.
>
> I've been involved with everything from IEEE subcomittees to government
> installations to 5 figures of nodes in computing and trading grids.  There
> is nothing more destructive than people talking where they don't have
> experience, and stating incorrect information.  It starts with
> over-simplification, but that's usually not bad.  Most people do web, maybe
> remote shells, and you're talking TCP segments passing back and forth
> between source and destination ports.  But I saw where it was going, and my
> assumptions proved right.  They quickly became factually incorrect.
>
> You don't have to be doing real-time, Infiniband and other exchanges to not
> have transports using ports.  There are many, common IP protocols and
> framing exchange, even some higher-level applications, that do not use
> transports with ports.  As I used as an example above, troubleshooting
> PXEboot is one area where you really need to be familiar with this,
> especially.  But I could dive into many other, real world cases.
>
>
>
>
> ----- Original Message -----
> From: Joel J. <tcepsa at gmail.com>
> Sent: Thursday, September 15, 2011 5:11 PM
>
> GGR> Yes, every application that talks over the network or within the
> OS itself has a port, it may be a temporary port, but it is
> GGR> present. In fact when 1 app talks to another app within the same
> server/pc, a port is opened for communication. It may not be
> GGR> seen on the network, but it is still present.
>
> I'm going to have to go with Brian on this one.  For example, I
> present the humble ping application.  It communicates with another
> computer across a network, yet does not use a port.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.unknownlamer.org/pipermail/calug/attachments/20110915/d58b126c/attachment.htm 


More information about the CALUG mailing list