On Sat, Mar 28, 2009 at 1:23 AM, Tony Li <[email protected]> wrote:
> locator     A locator is a name that has topological sensitivity and
>            must change if the point of attachment changes.

Hi Tony,

That soft-sells it a little. A locator specifies an entity's current
attachment within the network topology.

Types of locators include:

* Local Locator - a set of names and/or numbers which specifies the
point of attachment of a host, interface or service within the network
topology of the local administrative zone.

* Remote Locator - a set of names and/or numbers which specifies the
point of attachment of an administrative zone within the network
topology of the Internet.


Since I've tied these definitions to "administrative zones," I'd
better define that term too:

* Administrative zone -  a portion of the network in which a single
entity holds some combination of access, authority and ownership which
grants said entity full control over that portion of the network.

That's a little fuzzy, so let me clarify it with some examples:

1. The computers on Tony's network have static IP addresses and
connect to an ISP. Tony's network is a separate administrative zone
from the ISP's because the ISP has neither access to change Tony's
network settings nor ownership of Tony's equipment.

2. Brian's computer gets a dynamic IP address from his cable Internet
provider. Brian's computer is within the ISP's administrative zone
because the ISP has the access and authority to change Brian's IP
address at its discretion.

3. Tony's wife has has her own computer. It get's a DHCP address from
Tony's computer. Her computer is part of Tony's administrative zone.

4. Tony's wife has a static IP address and she forbids Tony from
touching her computer. Her computer is a separate administrative zone
from Tonys because Tony lacks access and authority to change her
computer's IP address.

5. XYZ corp has an email team, a web team and a network team which
operate independently under the CTO. XYZ corp is a single
administrative zone because XYZ corp owns the whole thing and the CTO
can compel changes to the network across all three teams at his
discretion.

6. XYZ corp outsources it's spam filtering to ABC corp. ABC corp is a
separate administrative zone from XYZ corp. Though XYZ's contract
presumably allows it to have ABC change it's network to match XYZ's,
no one at XYZ has the access, ownership or authority to directly
change ABC's network.



> identifier  An identifier is the name of an endpoint.  It has no
>            topological sensitivity.  That is, the identifier will not
>            change, even if the endpoint changes its attachment within
>            the topology.  Identifiers may have other properties, such
>            as the scope of their uniqueness (global or local) and the
>            probability of their uniqueness (absolute or statistical).


Types of identifiers include:

* Host identifier - a name or number expected to be attached to a
specific computer for it's entire service life.

Example: "minax" is my web server.

* Globally unique host identifier - Same plus is unique within the
scope of the Internet.

Example: "minax.dirtside.com" is my web server.

* Service identifier - a set of names and/or numbers which designate a
service provided by a host or cluster of hosts. A service identifier
is expected to remain constant for as long as the service is offered
to the Internet and must be unique within the scope of the Internet.

Example: www.whitehouse.net protocol http on IPv4 TCP port 80

* Transport identifier - a set of names and/or numbers which designate
a particular transport connection or associated set of packets between
an initiator and a service. A transport identifier is expected to
remain constant for the duration of the connection and must be unique
within the scope of the two communicating endpoints.

Example: the combination of source and destination IP address plus
source and destination port is a TCP transport identifier.

* Session identifier - a set of names and/or numbers which associate a
single communications session which occurs over multiple invocations
of one or more transport protocols. A session identifier must be
unique within the scope of the associated user and service.

Example: a web login cookie.


To clarify some of the above, I think I should also define two new terms:

* Initiator - the computer which starts a communication session
between itself and a service on the internet. Computers on the
Internet don't just happen to bump into each other. One end or the
other initiates that communication.

Example: A PC with a web browser.

* Service - some presumably useful communication provided by an
Internet-connected computer upon receiving a request from an
initiator.

Example: The results of the various URLs under http://www.cnn.com/


The purpose of routing is, of course, to move packets from the
initiator to a computer providing the service and then back to the
initiator.


> address     An address is a name that is both a locator and an
>            identifier.

An address is a number used by the networking protocols to implement
one or more types of locators and identifiers.

In deployed IPv4, the "IP address" is overloaded with the following semantics:
1. The remote locator
2. Part of the local locator (the MAC address also plays a role in the
local locator)
3. The host identifier
4. Possibly part of the service identifier. Depends on how the
application interacts with the DNS. For example, because of "DNS
pinning" the IPv4 address is bound to an http or https service
identifier for the duration of each web session, even if the session
is anonymous (not identified). On the other hand, if you telnet to a
hostname, then the IP address is not part of the service identifier
because the service name is translated to a host identifier after
which the service identifier plays no further role.
5. Part of the transport identifier (the other parts are the remote IP
address and the local and report protocol ports)


Finally, we should define (or at least clarify) two additional terms:

* Route - The path from the current entity through the network
topology to the destination locator. Alternately: the next hop in said
path.

* Map - The association between an identifier and the current set of
locators at which the identifier is found.


Regards,
Bill Herrin


-- 
William D. Herrin ................ [email protected]  [email protected]
3005 Crane Dr. ...................... Web: <http://bill.herrin.us/>
Falls Church, VA 22042-3004
_______________________________________________
rrg mailing list
[email protected]
http://www.irtf.org/mailman/listinfo/rrg

Reply via email to