Professional Documents
Culture Documents
Routing Protocols y H.
H Karl and A.
A Willing.
Willing Protocols and Architectures for Wireless
Sensor Networks. John Wiley & Sons, 2005. (Chapter 11)
y K. Sohraby, D. Minoli, and T. Znati. Wireless Sensor Networks:
Technology, Protocols, and Applications. John Wiley & Sons, 2007.
(Chapter 6)
y E.
E MM. Royer and C.C -K
K. Toh.
Toh A Review of Current Routing Protocols for
Ad-Hoc Mobile Wireless Networks. IEEE Personal Communications.
Vol. 6, No. 2, pp 46–55, April 1999
y X.
X HHong, KK. Xu,
X and dMM. Gerla.
G l Scalable
S l bl R
Routing
ti PProtocols
t l ffor M
Mobile
bil Ad
Hoc Networks. IEEE Network, Vol. 16, No. 4, pp. 11–21, July/August
2006
Introduction... Forwarding…
g
3 4
y WSNs can be deployed to support a wide range of y Whenever a node cannot send its data directly to the
applications, whether they have stationary or mobile sink, it has to rely assistance of intermediate nodes
nodes to forward the data, thus form a multihop network
y Once deployed, the nodes self-organize into a
autonomous wireless ad hoc network, with little or no S
C
maintenance
i A
y The nodes collaborate to perform the task(s) of the
D
application for which they are deployed B
y Their main task is to collect data, process data, forward y Node S wants to send a packet to node D; no direct
the information back to the application; achieving this link, has to pass (forward) the packet to its
link
task efficiently requires an efficient routing protocol to neighbor first
set up paths between the sensors and the data sink
Forwarding…2
g Routing
g tables…
5 6
Routing
gpprotocols… Ad hoc protocols…
p
7 8
y In “wired”
wired networks
networks, the routing protocols are Ad hoc routing
protocols
usually based on the link state or distance vector
algorithms.
y In wireless (mobile, multihop) networks, different Table driven
Source-initiated
On-demand
strategies
st ateg es a
aree required;
equ ed; sshould
ou d be
{ distributed
{ low overhead DSDV WRP AODV DSR LMR ABR
{ self configuring
y Ad hoc routing can be classified as CGSR TORA SSR
y Th
The crucial
i l aspects
t iin WSN
WSNs are y The idea is to eliminate the need of a routing table
{ The overhead is high or
{ Energy efficiency
{ The tables are to be created in the first place!
p
Ù Selecting the most energy
energ efficient routes
Ù Low overhead of table construction
y A parallel is drawn between the distribution of data in a
replicated data base system and epidemics occurring in
{ Stability and dependability of routes
h
human populations
l ti
{ Size of the routing table { Rumor mongering…
{ Resilience Ù Once an update is received, the site propagates this periodically and
randomly to another site
y Unicasting, broadcasting, and multicasting Ù The goal is to spread the updates as fast as possible while minimizing
Randomized forwarding…
g Rumor routing…
g
11 12
y In
I gossip
i bbased
d randomized
d i d fforwarding,
di the
th kkey iis yR
Rumor routingti
the probability with which a node retransmits an { Detect an event and install a few paths in the network
y Take
T k the
th network
t k graph
h y Minimize energy/bit
{ Example: A-B-E-H
y Assign a cost value (reflecting the energy y Maximize network lifetime
consumption)
i ) to each
h li
link
k { Time until the first node fails
y Routing considering available battery capacity
y Use an algorithm that would compute the least-cost { Maximum available batteryy levels
path(s) 4 Ù Example: A-C-F-H
A 2 { Minimum battery cost routing (MBCR)
3
1 Ù Example: A-D-H
1
C
{ Min-max battery capacity routing (MMBCR)
3 2 Ù Only take battery level into account when below a given level
2
D B
1 { Conditional min-max capacity
p y routing
g (CMMBCR)
2
{ Minimize variance in power levels
3 E
Send data from A to H
1 2
4 y Minimum total transmission power routing (MTPR)
4 2 F
H G
2
2
y Attracting
Att ti routes
t bby redirecting
di ti y Instead
I t d off constructing
t ti a single
i l energy-efficient
ffi i t path,
th
y Distance vector routing over topology control focus can be shifted on choosing the best possible
y Maximizing time to first node failure as a flow
path from a set of paths
{ Provides redundancy as a “standby”
problem
{ Goal is to find k paths that do not have either links or nodes in
y Maximizing time to first node failure by max-min common
optimization { Increased robustness
y Maximizing number of messages { Well studied
y Bounding the difference between routing protocols
Multipath
p unicast routing…2
g Multipath
p unicast routing…3
g 3
17 18
y Sequential assignment routing (SAR)… (SAR) y Unicast routing is the broadest research topic in the
{ Based on the observation that the nodes close to the sink are context of ad hoc and wireless sensor networks
likelyy to fail sooner y A variety of other topics considered (but not covered)
{ Construct trees outward from each sink neighbor
include:
y Energy-efficient
gy secondaryy p paths { Routing and topology control
{ Construct “braided” paths where various paths may share { Maximizing data flow for multiple source/destination pairs
nodes { All costs considered
y Si
Simultaneous
lt ttransmission
i i over multiplelti l paths
th { Integration of scheduling and power control
{ Send several copies of a packet over multiple node-disjoint { Routing and link quality
paths { R ti and
Routing d lif
lifetime
ti guarantees
t
y Random selection of the next upstream hop { Routing for one-shot queries
y While
Whil transmitting
t itti d data
t ffrom one node
d tto another,
th
nodes may also need to perform a broadcast
operation Broadcast Multicast
y Efficient broadcasting Æ to restrict the set of
f
forwarding
di nodesd as much h as possible,
ibl while
hil all
ll th
the
nodes receive data Source-based Shared core-
Mesh
t
tree b d ttree
based
y Multicasting
li i h has a similar
i il goall
y The
Th id
idea iis tto construct,
t t forf each
h source, a ttree, y A greedy heuristic—Shortest
heuristic Shortest Path Tree
rooted at the given source, that contains all the { Calculate the shortest (cheapest) path to each destination and
overlayy these p
paths onto a tree
destinations for this source,
source as well as other nodes
y Broadcast using minimum cost spanning tree—
that are needed
Prim’s algorithm
g
{ Try to find a tree for which the sum of all link costs is
{ Start with a tree consisting of the source node and, in |V|-1
minimum (Steiner tree)
steps, add one node
{ Try to find the minimum cost to each destination (shortest
(shortest- { For every next node,
node construct a shortest path to some other
path tree) node already on the tree
Steiner tree Shortest path tree
src
2
dst 2 src
2
dst 2 y Other Steiner tree approximations
y Broadcasting or multicasting with a finite set of
2 1 2 1
p
powers
dst 1 dst 1
Wireless advantage…
g Broadcast incremental p
power ((BIP)…
)
23 24
yE
Exploiting
l iti wireless
i l multicast
lti t advantage
d t y The
Th challenge
h ll is
i to
t find
fi d the
th ““core””
{ Multicast Incremental Power (MIP) y The problem can then be solved with a source-based
yE
Embedded
b dd d wireless
i l multicast
li advantage
d tree algorithm
l i h withi h the
h core as the
h source
{ Transforming existing trees y One approach is “merge point formation”, where a
y A distributed, position based approach merge point for a tree with a few sinks is to be found
Mesh-based protocols…
p Other broadcast and multicast approaches…
pp
29 30
{ The mesh (a sub-graph of the original graph) has to contain all y Optimal solutions for tree networks
{ Collecting and distributing data
sources andd destinations
d i i and
d provide
id at least
l one path
h ffrom
y Time to complete a multicast
each source to each destination { Time is important, too…
{ Redundancy in a mesh can enable shorter paths than in a core- y Data replacement
based tree, based on the forwarding procedure { Caching
Geographic
g p routing…
g Position-based routing
g basics…
31 32
intermediate nodes can be used to assist the routing Æ { Nearest with forward progress
position-based routing { Directional routing
{ It is necessary
y ((for manyy applications)
pp ) to address p
physical
y { Restricted flooding
locations (e.g., any node in a given region) Æ geocasting { Hmmm. Dead ends...
y For wireless sensor networks the latter is more y Basic idea to get out of a dead end: Put right hand to
important the
h wall,
ll follow
f ll the
h wall
ll
{ Nodes are interchangeable and distinguishable only by their { Does not work if on some inner wall—will walk in circles
external aspects, the location service is usually not necessary { N d some additional
Need dditi l rules
l tto d
detect
t t such
h circles
i l
Ù Send the packets around the face using the right-hand rule
Greedyy p
perimeter stateless routing
g ((GPSR)) Greedyy p
perimeter stateless routing
g ((GPSR))
33 34
y When
Wh stuck
t k iin a d
dead-end,
d d one way off escaping
i iis tto
keep the right hand to the wall and keep walking...
{ Practically,
P ti ll b backtrack
kt k ththe packet
k t outt off th
the d
dead
d end,
d counter-
t
clockwise around the obstacle
y Similar approach will allow the discovery of a node
closer to the destination in WSNs
y In GPSR,
GPSR a packet is greedil
greedily for
forwarded,
arded until
ntil an
obstacle is reached at, when the algorithm switches
to “perimeter
perimeter routing
routing” (essentially,
(essentially sending a packet
around a plane using the right-hand rule)
A GPSR example…
p Other p
position-based routing…
g
35 36
y Route
R t packet
k t from
f node
d A tto node
d Z y ID-based
ID b d routing,
ti hi
hierarchies
hi
y Randomized forwarding and adaptive node activity
Leave face (GeRaF))
(G
routing
E I y Geographic routing without positions (GEM)
B H K
F
Z
A D
Enter
E t
face J L
routing G
C
Geocasting…
g Further issues in geocasting…
g g
37 38
y Sending data to a subset of nodes that are located in an y Impact of localization errors
indicated region (an example of multicasting!) { Impractical assumption in real systems, that all the nodes know their
location
{ Geographically
g p y restricted flooding
g
Ù Location based multicast y Location
i services
i
| Static zone, adaptive zone, adaptive distance { Mapping node IDs to node locations; more important in ad hoc networks
yS
Sources off mobility
bilit iin WSN
WSNs: So rce
Source
{ Data
D t sinks
i k
Sink moves
{ Observed event(s) downward
Source
Source
Sink
moves
upward