Re: Rose Tea

An accurate, simple explanation would have said there are two parts to an address, 
that the 
first part identifies a network, the second part identifies a computer or other system 
attached to the network, and that there's nothing to do with geography in the way 
network 
addresses are structured.

What was incorrect about your "simple" explanation was that you added details that are 
not 
generally true.  In this case the details were four levels of hierarchy for allocating 
addresses, when in fact the standards and routers only deal with two.  Some network 
administrators might make up extra hierarchy as a tool for allocating allocate 
addresses 
that way, and perhaps it was done that way in the university where you worked, but 
there is 
no standard rule that says that network addresses have to be allocated that way, and 
any 
bookkeeping/allocation scheme that a administrator builds on top of the standard 
two-part 
structure isn't seen or used by the router.  If you had said that it was an example of 
how 
network addresses could be allocated, that other ways were possible, and that 
allocation is 
different from the two-part structure that the router sees, that would have been 
correct, if 
not necessarily simple.  But you were telling the group that addresses are necessarily 
allocated and structured in a certain way, in four parts, and that's just not the 
case.  In 
fact, one of the key elements in the design of the IPv4 address space is that it is 
completely flat, not at all hierarchical as you portray.  Internet routing protocols 
would 
work entirely differently than they do if addresses were structured rather than flat.  

Here are two examples of the consequences of the IP address space being flat.  They 
both 
contradict your explanation.  First, two network addresses may begin with the same 
octet or 
octets, and yet be completely unrelated to each other - on networks that are thousands 
of 
miles apart and owned by different organizations.  OTOH, two network addresses may 
start 
with completely different three-octet sequences, yet both be owned by the same 
organization, 
representing adjacent subnets and connected via ports on a particular router.

The networking personel where you were working computer support may have been telling 
you 
how they did things... a particular example... or they may not have been as 
knowledgable as 
one would hope.  My background is more than 25 years as a network technologist.  I 
have 
contributed to the design of protocols used today in the Internet.  I have also been a 
product manager for routers for a major vendor (no, not Cisco)... and specified the IP 
protocol support down to the nittiest grittiest detail.  The development engineers 
would 
have let me know it if I had anything wrong...  The press would have been quick to 
point out 
egregious errors like not understanding IP addressing in their reviews...   Customers 
would 
have complained mightily.  What I say about tea may be only my opinion, but when it 
comes to 
the basics of IP networking I think I am very firm ground.

Debbie
-- 
Anti-spam advisory: The email address used to post this article is a throw-away 
address.  It 
will be invalidated and replaced with another if and when it is found by spammers.