-
The
Safari
architecture
(Du
et
al.
2008)
also
provides
a
self-organizing
network
hierar-
chy and
a
routing
protocol
with
similarity
to
the
landmark
approaches.
Such
landmarks,
called
drums
in
Safari,
use
a
self-election
algorithm
(a
distributed
algorithm
with
no
cen-
tralized
coordination)
to
form
subnets,
called
cells
and
supercells
.
At
the
lowest
level
of
the
cell
hierarchy
(level
0), each
individual
node
forms
its
own cell.
At
level 1, Safari
de?nes
fundamental cells
, that is, cells that contain
multiple
individual
nodes,
but
no
other
cells.
Higher-level
cells
are then composed of multiple
smaller cells at lower levels. Each
drum
periodically
broadcasts
beacon
messages
within
well-
de?ned
limited scopes in the network. These beacons aid
the
hierarchy formation, give nodes an
indication of their position in
the
network topology, and provide routing information
toward the
drum
’
s
cell.
Safari
uses
a
hybrid
approach
for
routing,
that
is,
routing within cells is
based on the reactive DSR protocols, while
a proactive
routing
approach
is used to compute
routes
to more
distant
nodes.
The
inter-cell
communication
relies
on
a
destination
node
’
s hierarchical address
and on the beacon records
stored on
each node. Basically,
inter-cell
communication
follows
the
Network Layer
183
reverse
path
of
the beacons
issued
by
the
cells
’
drums.
The
main
advantage
of
Safari
over
other
approaches
is
its
scalability,
due
to
the
hybrid
routing
approach
and
the
hierarchical
addressing
scheme.
7.8
Location-
Based
Routing
Location-based
or
geographic
routing
can
be
used
in
networks
where
sensor
nodes
are
able
to
determine
their position using a variety of localization
systems and algorithms (see Chapter
10 for examples of localization
techniques). Instead of topological connectivity
information, sensors
use geographic
information to make forwarding decisions. In
unicast location-based routing, packets
are sent directly to a single
destination, which is
identi?ed
by its location.
That is, a sender must be
aware not
only of
its own
location,
but also the location of the des-
tination. This
location
can
be
obtained
either
via
querying
(e.g.,
?
ooding
a
query
to
request
a
response
from
the
destination
containing
its
location)
or
a
location
broker
,
that
is,
a
ser-
vice
that
maps
node
identities
to
locations.
In
broadcast
or
multicast
location-based
routing
approaches,
the
same
packet
must
be
disseminated
to multiple
destinations.
Multicast pro- tocols
take advantage of the known destination
locations to minimize resource
consumption by reducing redundant
links.
The
identity
of
a
sensor
node
is
typically
less
important
than
its
location,
that
is,
data may
be
disseminated
to
all
nodes
that
lie
within
a
certain
geographic
region.
This
approach
is
called
geocasting
and can, for example, be used to
diffuse queries to
speci?c
regions of interest instead of
?
ooding the entire network,
signi?cantly
reducing both bandwidth and energy
requirements. Once a
packet
reaches
the desired
region,
it
must
be either
dissemi- nated
(multicast)
to
all
nodes
within
this
region or transmitted
to at least one
node within this region (
anycast
).
Typically,
location-based routing protocols require that
every node in the network
knows
its own
geographic
location and the identities
and locations of its one-hop neighbors (e.g.,
obtained
via periodic beacon messages).
The destination is expressed either as the
location of a node (instead
of a unique
address)
or a geographic
region.
Compared
to other
routing solutions,
an advantage
of
location-
based routing is that only geographic information
is needed for forwarding decisions and
it
is
not
necessary
to
maintain
routing
tables
or
to
establish end-to-end
paths
between
sources
and
destinations,
eliminating
the
need
for
control
packets
(apart
from
the
beacon
messages
among
neighbors).
-
-
-
-
-
-
-
-
-
上一篇:无肉不欢:各种肉类英语怎么说
下一篇:简单常用食物类英语单词翻译:肉类及海鲜