ietf
[Top] [All Lists]

Re: "mobile" orthogonal to wide-area wireless

2000-10-18 22:30:03


All of this and a great deal more is discussed in various old books,
such as:

- Internetwork Mobility - The CDPD Approach
  Taylor, Waung and Banan
  Prentice Hall
  1996
  ISBN: 0-13-209693-5

Hope this helps.

...Mohsen

On Wed, 18 Oct 2000 23:04:39 -0400, Keith Moore 
<moore(_at_)cs(_dot_)utk(_dot_)edu> said:

  >> >I would rather have one address for a wireless WAN interface and
  >> >another address for a wireless LAN interface -- which seems to be
  >> >doable today -- much more than I want to wait for the solution
  >> >with "Mobile IP" address traversal to become commercially available.
  >> 
  >> No, what you want is a serial number. Or at least what you describe 
  >> is a serial number.  An address is very different.

  Keith> the word "address" is used in so many different ways that any
  Keith> argument of the form "x is/is not an address" tends to be 
  Keith> nothing more than an argument about which definition of the
  Keith> word "will be master" (to quote Humpty Dumpty)

  Keith> users don't care about whether their mobile device has one address
  Keith> that follows it everywhere or whether it changes addresses as
  Keith> it moves.  however, depending on their needs, they might care about
  Keith> their applications continuing to stay connected while they're mobile.
  Keith> they might also care about running applications that are both mobile
  Keith> and "always on".

  Keith> any network stack that supports the latter two kinds of applications
  Keith> will almost certainly employ (at least) two different sets of 
  Keith> things that could be called "addresses" - one which is stable
  Keith> even while the device is mobile, and another which is less stable.
  Keith> whether those two addresses look alike or different, whether
  Keith> the technology used is "mobile IP" or something else , and
  Keith> the level of the protocol stack at which the indirection occurs
  Keith> - these are implementation choices.

  Keith> of course, some implementation choices work better than others,
  Keith> especially when it comes to interoperating with the wired Internet,
  Keith> or in being able to support existing applications, or in being 
  Keith> able to switch from one communications medium to another.  but 
  Keith> the implementation choice can quite reasonably be different
  Keith> depending on the particular characteristics of the device and
  Keith> its communications media, and also on the needs of its users.

  Keith> Ketih