ad hoc and wireless mesh networking - uppsala university€¦ · ad hoc and wireless mesh networks...
Post on 05-Jun-2018
226 Views
Preview:
TRANSCRIPT
Ad Hoc and Wireless MeshNetworking
Laura Marie Feeneylmfeeney@sics.se
Datakommunikation III, HT 2006
Overview
Ad hoc and wireless mesh networksI Ad hoc network (MANet)
I operates independently of networkinfrastructure
I nodes cooperate to provide networkservices
I Mesh networkI supplements network infrastructureI nodes cooperate to provide Internet
access
Ad hoc network
mobile wireless network
capable of autonomous operation
Ad hoc network
mobile wireless network
capable of autonomous operation
I operates without base station
infrastructure
Ad hoc network
mobile wireless network
capable of autonomous operation
I operates without base station
infrastructure
I nodes cooperate to provide connectivity
Ad hoc network
mobile wireless network
capable of autonomous operation
I operates without base station
infrastructure
I nodes cooperate to provide connectivity
I every node is a router (no default
router)
Ad hoc network
mobile wireless network
capable of autonomous operation
I operates without base station
infrastructure
I nodes cooperate to provide connectivity
I every node is a router (no default
router)
ad hoc routing problem
Ad hoc network
mobile wireless network
capable of autonomous operation
Ad hoc network
mobile wireless network
capable of autonomous operation
I operates without centralized
administration
Ad hoc network
mobile wireless network
capable of autonomous operation
I operates without centralized
administration
I nodes cooperate to provide services
Ad hoc network
mobile wireless network
capable of autonomous operation
I operates without centralized
administration
I nodes cooperate to provide services
I address allocation (no DHCP)
Ad hoc network
mobile wireless network
capable of autonomous operation
I operates without centralized
administration
I nodes cooperate to provide services
I address allocation (no DHCP)
I fairness and security
Example
network nodes (mobiles, laptops, PDA’s)
Example
wireless communication
Example
“equivalent” topology
Example
discover multihop route
Example
route failure due to mobility
Example
dynamic route repair
Example
extend infrastructure
Example
How is a MANET different from other
networks?
I Internet
I WLAN/Cellular
I MobileIP
Internet
default router
Internet routing(+ humans)
networkprovider
simple default router at edges
expertise managing core
WLAN/celluar
nodes communicate only with base-station
Mobile IP
correspondent
CN
MNhome
FA
MobileIP allows a node to change its point of
attachment to the network
Mobile IP
CNFA
HA
MN
home agent (HA) tunnels traffic to the node
Applications
I Military/rescue
(no infrastructure)
Applications
I Military/rescue
(no infrastructure)
I Disaster management
(damaged infrastructure)
Applications
I Military/rescue
(no infrastructure)
I Disaster management
(damaged infrastructure)
I Spontaneous networks
(meeting/conference room)
Applications
I Military/rescue
(no infrastructure)
I Disaster management
(damaged infrastructure)
I Spontaneous networks
(meeting/conference room)
I Personal area networks
Applications
I Military/rescue
(no infrastructure)
I Disaster management
(damaged infrastructure)
I Spontaneous networks
(meeting/conference room)
I Personal area networks
I Campus area networks
Applications
I Extend coverage
improve range or capacity in a building
Applications
I Extend coverage
improve range or capacity in a building
I Community mesh networks
share network connectivity with
neighbors
Applications
I Extend coverage
improve range or capacity in a building
I Community mesh networks
share network connectivity with
neighbors
I Developing countries
provide cheap communication
infrastructure
History
Not a new idea
I US DARPA (1970’s)
History
Not a new idea
I US DARPA (1970’s)
I Amateur (ham) radio operators (1970’s)
“packet radio”
History
Not a new idea
I US DARPA (1970’s)
I Amateur (ham) radio operators (1970’s)
“packet radio”
I renewed interest mid-1990’s
History
Not a new idea
I US DARPA (1970’s)
I Amateur (ham) radio operators (1970’s)
“packet radio”
I renewed interest mid-1990’s
I IETF MANET working group (1997)
History
Not a new idea
I US DARPA (1970’s)
I Amateur (ham) radio operators (1970’s)
“packet radio”
I renewed interest mid-1990’s
I IETF MANET working group (1997)
I startup companies (2000–)
Ad hoc routing
Outline
I challenges
I design choices
I protocol example
I wireless communication
Challenges
I distributed state in unreliable
environment
Challenges
I distributed state in unreliable
environment
I changing topology
I limited communication capacity
I limited battery capacity
Challenges
I distributed state in unreliable
environment
I changing topology
I limited communication capacity
I limited battery capacityI wireless communication
I variable link qualityI non-symetric linksI interference and collisions
CriteriaI effectiveness
I convergence/recoveryI scalability (number of nodes, density)
CriteriaI effectiveness
I convergence/recoveryI scalability (number of nodes, density)
I performanceI data throughputI route latency (delay)I route optimality
(hops/stability/diversity)I overhead cost
(packets/bandwidth/energy)
Alphabet Soup
many proposed protocols:
AODV CEDAR ABR FSR
TORA GSR OLSR LANMAR
ZRP LAR DSR OSPF++
RDMAR CBRP DSDV WRP
TBRPF CGSR GPSR HSR
Design choices
protocols divided into a few main categories
I on-demand (reactive)
I table driven (proactive)
I hierarchical/cluster-based
I geographic/position-based
Reactive Routing
find routes as needed (on demand)
I advantagesI no overhead maintaining unused routes
I disadvantagesI high route latencyI optimization?
Proactive routing
table-based, more similar to conventional
routing
I advantagesI low route latencyI state information
I disadvantagesI high overhead (periodic table updates)I route accuracy depends on updates
AODV (DYMO)
Ad hoc On demand Distance Vector
Perkins et.al.
AODV (DYMO)
Ad hoc On demand Distance Vector
Perkins et.al.
conventional distance vector
I nodes exchange distance information (to
all nodes) with their neighbors
I periodic exchange and immediate
update for changes
I routing table selects shortest path
AODV (DYMO)
Ad hoc On demand Distance Vector
Perkins et.al.
conventional distance vector
I nodes exchange distance information (to
all nodes) with their neighborsI periodic exchange and immediate
update for changesI routing table selects shortest path
exchange a lot of information that is never
used
AODV (DYMO)
on-demand variant of conventional distance
vector
AODV (DYMO)
on-demand variant of conventional distance
vector
I route request (RREQ) packet is flooded
through the network
AODV (DYMO)
on-demand variant of conventional distance
vector
I route request (RREQ) packet is flooded
through the network
I route discovery creates (temporary)
reverse paths back to the source
AODV (DYMO)
on-demand variant of conventional distance
vector
I route request (RREQ) packet is flooded
through the network
I route discovery creates (temporary)
reverse paths back to the source
I route reply (RREP) turns a valid reverse
path into a route
AODV (DYMO)
handling topology change
AODV (DYMO)
handling topology change
I link failure causes route error (RERR)
I destination managed sequence number
ensures loop freedom
AODV (DYMO)
handling topology change
I link failure causes route error (RERR)
I destination managed sequence number
ensures loop freedom
AODV is RFC 3562 (experimental)
DYMO is IETF Internet draft
AODV (simplified)
7
1
2
6
11
13103
5
12
4
8
914
route from node 1 to node 14
AODV (RREQ)
1479
8
4
12
5
3 10 13
11
6
1
2
14?1 hop
broadcast flooding of route request
wireless multicast advantage
AODV (RREQ)
5
1hop
41−>14: via 1
8
1hop
914
1−>14: via 1
3
6
7
111hop
13
1
101−>14: via 1
212
node from which RREQ was received defines
reverse path to source
AODV (RREQ)
14?
2 hops
2
6
1
14
11
13103
5
12
4
8
97
2 hops
2 hops
14?
14?
RREQ is flooded though the network
AODV (RREQ)
2hop
6
14
1
1310
2
3
5
12
4
8
97
1−>14: via 3
1−>14 (via 1)
1−>14 (via 1)
1 hop
1 hop
1−>14 (via 1)
1 hop
1−>14: via 3
1−>14 (via 5)2hop
2hop
2hop
1−>14: via 2
11
reverse paths are recorded
AODV (RREQ)
14?
3 hops
3 hops 14
3
7
1
6
3 hops
1310
5
12
4
8
9
3 hops
11
2 14?
14?
14?
unreliable broadcast
destination managed sequence number, ID
prevent looping
AODV (RREQ)
3hop
79
8
4
12
5
3 10 13
1
14
6
11
1−>14: via 3
1−>14 (via 1)
1−>14 (via 1)
1 hop
1 hop
1−>14 (via 1)
1 hop
1−>14: via 3
1−>14 (via 5)2hop
2hop
2hop
1−>14: via 22hop
1−>14: via 83hop
1−>14: via 83hop
1−>14: via 7
2
reverse paths are recorded
AODV (RREQ)
4 hops
4 hops2
14
3
7
1
6
11
1310
5
12
4
8
9
4 hops14?
14?
14?
broadcast collision problem (jitter)
AODV (RREQ)
4hop
6
14
1
2
13103
5
12
4
8
97
1−>14: via 3
1−>14 (via 1)
1−>14 (via 1)
1 hop
1 hop
1−>14 (via 1)
1 hop
1−>14: via 3
1−>14 (via 5)2hop
2hop
2hop
1−>14: via 22hop
1−>14: via 8
1−>14: via 83hop
1−>14: via 73hop
3hop
1−>14 via 10
1−>14 via 104hop
11
reverse paths are recorded
AODV (RREQ)
14?
3
7 14
13
1
6
11
10
5
12
4
8
9 5 hops
5 hops14?
2
broadcast flooding is very expensive
AODV (RREQ)
5 hops
79
8
4
12
5
3 10 13
1
14
6
11
1−>14: via 3
1−>14 (via 1)
1−>14 (via 1)
1 hop
1 hop
1−>14 (via 1)
1 hop
1−>14: via 3
1−>14 (via 5)2hop
2hop
2hop
1−>14: via 22hop
1−>14: via 8
1−>14: via 83hop
1−>14: via 73hop
3hop
1−>14 via 10
1−>14 via 104hop
4hop
14!2
RREQ arrives at the destination
two routes are discovered
AODV (RREP)
14
7
3
14
1
6
11
10
2
5
12
4
8
9
13
destination sends unicast RREP (sets
sequence number)
“activate” reverse path
AODV (RREP)
14
27
3
14
1
6
11
10
5
12
4
8
9
13
destination sends unicast RREP (sets
sequence number - not shown)
AODV (RREP)
11
6
1
9
3
7
1314
14
14 via 13
214
8
4
12
5
10
RREP messages “activates” reverse path
AODV (RREP)
2
5
12
14
4
148
14
79
13
14
via 13
1
3
6
1114
10
via 10
RREP messages “activates” reverse path
AODV (RREP)
via 8
9
8
4
12
5
10
11
6
1
14
3
7
1314
14
14 via 13
via 101414
2
RREP reaches source – route discovery
complete
AODV (RREP)
via 5
9
8
4
12
5
10
11
6
1
14
3
7
1314
14
14 via 13
14 via 10via 814
2
source adopts destination sequence number
AODV
via 10
9
8
4
12
5
10
11
6
1
14
3
7
1314
14 via 5 via 814
14 via 13
14
2
traffic flows along forward route
reverse path information times out
AODV(RERR)
via 10
9
8
4
12
5
10
11
6
1
14
3
7
1314
14 via 5 via 814
14 via 13
14
2
link failure
how to be sure?
AODV(RERR)
10
14
5
12
via 5
4 14
8
via 13
9
3
14
7
via 8
13
1
6
14
11
2
14
node generates route error message (RERR)
AODV(RERR)
via 13
9
8
4
12
5
10
11
6
1
14
3
7
1314
14 via 5
14
2
forwarded to source
AODV(RERR)
4
12
5
10
11
6
1
14
8
7
1314 via 13
14 via 1014?
2
14
9
3
initiate new route discovery (RREQ)
AODV(RERR)
97
3
14
1
6
11
10
2
5
12
4
8
13
discover new route (latency)
AODV(RERR)
14?
7
3
14
1
6
2
11
10
5
12
4
8
9
14
14 via 5 via 814
14
14 via 10
via 1313
what about a local repair?
initiate RREQ from point of failure
AODV(RERR)
14?
7
3
14
1
6
2
11
10
5
12
4
8
9
14
14 via 5 via 814
14
14 via 10
via 1313
what about a local repair?
initiate RREQ from point of failure
AODV(RERR)
via 10
7
3
14
1
6
2
11
10
5
12
4
8
9
14
14 via 5 via 814
14
14 via 10
via 13
14
13
lower latency
AODV(RERR)
via 10
7
3
14
1
6
2
11
10
5
12
4
8
9
14
14 via 5 via 814
14
14 via 10
via 13
14
13
lower latency but longer routes
AODV(DYMO)
I (over)-simplified description here
I further complexities and optimizations
I AODV-UU implementation is
well-known
CriteriaI effectiveness
I convergence/recoveryI scalability (number of nodes, density)
I performanceI data throughputI route latency (delay)I route optimality
(hops/stability/diversity)I overhead cost
(packets/bandwidth/energy)
OLSR
Optimized Link State Routing
Clausen et.al
OLSR
Optimized Link State Routing
Clausen et.al
conventional link-state routing
I beacon to determine neighbors
I for each node, disseminate its links to
all other nodes
I use SPF algorithm to generate routing
table
OLSR
Optimized Link State Routing
Clausen et.al
conventional link-state routing
I beacon to determine neighborsI for each node, disseminate its links to
all other nodesI use SPF algorithm to generate routing
table
high overhead, exchange information for links
that are never used
OLSR
variant of conventional link state routing
OLSR
variant of conventional link state routing
I for each node, disseminate only some of
its links
I for each node, only disseminate
information received via some links
I use SPF to generate routing table
OLSR
variant of conventional link state routing
I for each node, disseminate only some of
its links
I for each node, only disseminate
information received via some links
I use SPF to generate routing table
“some links” = multipoint selector set
OLSR
IETF RFC 3626 (experimental)
2-hop Neighborhood
broadcast periodic “hello” messages
I each message contains a list of
neighbors
I each node discovers its 2-hop
neighborhood
I discovers failed links
I discovers bi-directional links
Bi-directional Links
HELLO(3)={1,2,4,7}
HELLO(6)={2,3,7}
6
4
3
14
1
2
13
129
10
11
7
85
Bi-directional Links
NBR(3)={1,2,4,7}
NBR(3)={1,2,4,7}HELLO(6)={2,3,7}
5
3
14
1
2
13
129
10
11
7
8
6
4
Multipoint Relay
multipoint relay set (MPR): subset of a
node’s 1-hop neighbors, such that each of its
2-hop neighbors is a 1-hop neighbor of a
node in the MPR set
Multipoint Relay
multipoint relay set (MPR): subset of a
node’s 1-hop neighbors, such that each of its
2-hop neighbors is a 1-hop neighbor of a
node in the MPR set
minimum set of 1-hop neighbors that results
in the same set of 2-hop neighbors
Multipoint Relay
multipoint relay set (MPR): subset of a
node’s 1-hop neighbors, such that each of its
2-hop neighbors is a 1-hop neighbor of a
node in the MPR set
minimum set of 1-hop neighbors that results
in the same set of 2-hop neighbors
each node has an MPR set (not a “network”
MPR set)
Multipoint Relay(MPR set)
12
1
2
6
9
10
1185
3
7
4
14
13
1-hop and 2-hop neighbors of node 4 ode 5
is not needed in the multipoint relay set
Multipoint Relay(MPR set)
12
1
2
6
9
10
1185
3
7
4
14
13
node 5 is not needed in the multipoint relay
set
Dense Network
12
1
2
6
9
10
1185
3
7
4
14
13
with greater node density, the proportion of
relay nodes is smaller
Dense Network (MPR set)
{3,8}
{4,10}
{10}
{1,4,8}
{8,10}
{4,7}
{1,3}
{3,5}
{7,8,12} {9,10}
{10}{7,8}
{6,4,10}
{2,7}
9
13
14
1
5
11
10
127
32
4
8
6
nodes which are not in the MPR set are in
some sense redundant
Dense Network (MS set)
9
13
14
1
5
11
10
127
32
4
8
6
multipoint selector (MS) set is the inverse of
MPR set
i.e. nodes that have selected this node as an
MPR
Dense Network (OLSR)
Operation:
Dense Network (OLSR)
Operation:
I each node uses HELLO messages to find
and announce its MPR set
Dense Network (OLSR)
Operation:
I each node uses HELLO messages to find
and announce its MPR setI a node sends link state information only
for nodes in its MS set (for which it is
an MPR)
Dense Network (OLSR)
Operation:
I each node uses HELLO messages to find
and announce its MPR setI a node sends link state information only
for nodes in its MS set (for which it is
an MPR)I each node computes SPF routes using
all received messages
Dense Network (OLSR)
Operation:
I each node uses HELLO messages to find
and announce its MPR setI a node sends link state information only
for nodes in its MS set (for which it is
an MPR)I each node computes SPF routes using
all received messagesI a node only rebroadcasts link state
messages from nodes in its MS set
OLSR
only disseminate link data for green nodes
only rebroadcast data from green nodes1: 4 2 3 5 2: 1 3 6 3: 1 2 4 6 74: 1 3 5 7 8 5: 1 5 8 6: 2 3 77: 3 4 6 9 10 8: 4 5 9 10 11 9: 7 10 1210: 7 8 9 11 12 13 11: 8 10 13 12: 9 10 13 1413: 10 11 12 14 14: 10 12 13
CriteriaI effectiveness
I convergence/recoveryI scalability (number of nodes, density)
I performanceI data throughputI route latency (delay)I route optimality
(hops/stability/diversity)I overhead cost
(packets/bandwidth/energy)
top related