Rereading the original post, it looks as if Ahmed may have tunneled
himself into a position of 'NS-2 is the answer, what is the
question?'.
Most of the posts have, in the spirit of MANET, provided layer 3
answers. But perhaps a review of 'mobility' is useful to remind
ourselves that there's more to the problem. Just what does mobility
mean to _you_?
At layer 1, mobility means platform mobility and is measured in metrics
like doppler.
At layer 2, mobility usually means handoffs. As in passing a cellphone
connection from one pylon to another as the cellphone physically moves.
Similarly, the location database on the backside needs to know where an
on-hook cellphone is in order to direct an incoming call (same issues
for a campus WiFi setup with multiple WiFi splotches. The key here is
homogeneity -- the mobility characteristics do not affect router tables.
At layer 3, mobility connotes MANET ... as most of the responders
tacitly assumed. Routing table volatility. But one can argue that OSPF
and DHCP both deal with mobility too.
At layer 4/5, 'mobility' may not be the correct term of engagement, but
Ahmed used 'multicast' in his query and it is certainly valid. The
brute force observation here is that wired internet is provisioned at Gs
while as soon as you hit a radio, the r-WANs (MANET or otherwise) are
provisioned at low Ms ... on a good day. Multicast is the only real
offset to the physics-imposed bandwidth limitations. This observation
makes it a bit difficult to divorce multicast issues from mobility ones
in reality (although we do need to do that for protocol development
reasons).
And at layers 6/7 we have few applications that actually use the
multicast. This bullet for completeness -- we don't have a complete
solution without.
The reminder is that MANET is properly disciplined to a layer 3
solution. That's good. But it's somewhat less than a complete
discussion of mobility.
| Article list | Name | Date |
| Click Here | mohamed abas |
||
| Click Here | jacquet |
||
| Click Here | Dearlove, Christopher (UK) |
||
| Click Here | Charles E. Perkins |
||
| Click Here | jacquet |
||
| Click Here | Chi-Anh La |
||
| Click Here | Dearlove, Christopher (UK) |
||
| Click Here | Nils Aschenbruck |
||
| Click Here | Stephan Bohacek |
||
| Currently Viewing : This Article | Rex Buddenberg | 2010-02-08 17:13:33 | |
| Click Here | Dearlove, Christopher (UK) |