Energy-Efficient Multicast Protocols in Wireless Ad Hoc Networks

1
Energy-Efficient Multicast Protocols in Wireless
Ad Hoc Networks

• Sandeep K. S. Gupta
• Computer Science and Engineering Department
• Arizona State University
• Tempe, AZ, USA
• Sandeep.Gupta_at_asu.edu

2
Outline

• Multicasting
• Techniques for Conserving Energy Wireless Network
• Multicasting in Wireless Network
• Node Metric and Cost Models
• Protocols for Constructing Energy-Efficient
  Multicast Trees
• A Framework for Energy-Efficient Multicasting
• Conclusions

3
Multicasting

• Allow one entity to send messages to multiple
  entities residing in a subset of the nodes in the
  network
• Why multi-destination delivery in a single
  message?
• Transparency Efficiency Concurrency
• Applications
• distributed database, distributed games,
  teleconferencing

4
Techniques for Conserving Energy in Wireless
Network

• Turn-off non-used transceivers
• Scheduling transmission among nodes
• Reduce communication overhead, such as defer
  transmission when channel conditions are poor
• Transmission Power Control

5
Why Multicasting is different in Wireless
Networks?

• Wireless medium is broadcast medium (Wireless
  multicast Advantage)
• One time local transmission can possibly reach
  all the neighbors

6
Why Multicasting is different in Wireless
Network?

• Power control allows a node to determine who are
  its neighbors.
• More power used ?
• more interference
• Reduces simultaneous transmissions (thrput)
• Consumes energy at a faster rate ? node can die
  faster leading to disconnections.

7
Why Not Single-Hop Multicast?

• Single source multicast reach a subset of nodes
  from a given source s
• s increases its transmission range to such an
  extent that it can reach all the group members
• Increased interference and power wastage
• source may have limited transmission range

8
Multi-hop Approach

• Multi-hop Solution ? Problem of constructing
  multicast tree
• What is a link?
• Depends on power level
• Using maximum transmission power results in too
  many links
• link weight?
• 1. 2. ? Link-based view not appropriate!
• Node-based view construct tree with
  minimum/maximum summation of node cost

9
Energy Metric

• Two Criteria of Energy Optimization
• Total Energy Consumption (TEC)
• System Lifetime (SL)
• Node Cost
• Node Energy Cost
• Lifetime of a Node
• Type of Multicast Trees
• Source-based (this talk is restricted to
  source-based trees)
• Group-shared

10
Energy Metric
Initial battery energy at nodes 1, 2, and 3 are
200 EU
Minimum Energy Multicast Tree
Maximum Lifetime Multicast Tree
11
Nodes Energy Cost

• Energy consumed (per bit) at node i in
  source-based multicast tree T

where and are energy cost (per
bit) of transmission processing and reception
processing, is maximum energy cost (per
bit) of the link between node i and is children.
12
Nodes Energy Cost

• Energy cost (per bit) of node i for reliable
  multicast in source-based multicast tree T

where is the error rate for node i to
forward the multicast packet to all of node is
children reliably, and is the error
rate of node is parent to forward the packet to
all of its own children.
13
Nodes Energy Cost

• Nodes energy cost in group-shared Tree
• Tree Links attach to the node
• Direction of Message coming from
• Incorporate message generation rates of all the
  multicast sources in the tree.

Assume message generation rates of nodes 1 and 3
are 7pck/second and 13 packets/second. Average
energy cost of node 2
14
Nodes Lifetime

• Node is lifetime maximum number of multicast
  packets that may be forwarded by the node
  iwhere Ri(t) is remaining battery energy of
  node i at time t.

15
Cost of Multicast Tree

• The Total Energy Cost (TEC) of a multicast tree T
  
• The minimum TEC multicast tree T iswhere
  TG is the set of all possible multicast trees for
  the multicast group G in a given graph o.
• NP-Complete Problem
• Minimizing TEC of multicast tree ? Minimizing sum
  energy cost of all the tree nodes.

16
Cost of Multicast Tree

• Lifetime of Multicast tree T is
• The maximum lifetime multicast tree T
  iswhere TG is the set of all possible
  multicast trees for the multicast group G in a
  given graph o.
• NP-Complete Problem
• Maximizing multicast tree lifetime ? Maximizing
  the lifetime of trees bottleneck node

17
Protocols for Constructing Energy-Efficient
Multicast Trees

• Centralized Protocols
• Needs global knowledge (High Overhead) Not
  scalable!
• Adaptivity Expensive to adapt to dynamic
  changes, such as remaining battery at nodes
  Offline Approach.
• Distributed Protocols
• Local knowledge (Low Overhead) Scalable
• Adapt to dynamic changes Online approach

18
BIP/MIP Algorithm

• Constructing minimum TEC source-based broadcast
  tree T.
• Centralized ApproachU is the set of all nodes in
  the networkEi,j is the minimum energy cost of i
  to cover node j as a child.

• MIP Algorithm Pruning all of the non-group nodes
  which are leaf nodes in BIP tree.

19
BIP Algorithm

• Limitations of BIP
• Performance depends on the order of adding nodes
  in the tree.
• View is limited by adding one node at a time.

Minimum TEC Tree
BIP Tree
3 EU/pck
2 EU/pck
2 EU/pck
2 EU/pck
TEC 4 EU/pck
TEC 3 EU/pck
20
Distributed BIP

• Distributed Version of BIP
• Every node constructs BIP tree locally
• Dist-BIP-A Connect all the locally generated BIP
  trees (one-hop neighbor information)
• Dist-BIP-G Connect the locally generated BIP
  tree by the gateway nodes (two-hop neighbor
  information)

21
BIP/MIP Algorithm

• Combine energy cost and lifetime of a node as
  node cost in BIP/MIP

• Limitations
• Minimizing
• Ci is not the lifetime of node i, even when ?1
• Node Cost is a function of time, so the tree
  should be periodically refined
• ??, BIP/MIP chooses higher remaining battery
  nodes ?minimum TEC(T) or maximum LT(T)

22
EWMA Algorithm

• EWMA Algorithm refine MST to minimum TEC
  source-based broadcast tree (Centralized
  Approach)
• New Transmission Energy Selection Node i selects
  a downstream node j. The incremental energy of
  node i to cover js children isEnergy Gain is
  
• Selects the node j with highest positive Gain.
  Increase node is transmission energy to cover
  all of node js children and eliminate the
  redundant transmissions which are already covered
  by node i.

23
EWMA Algorithm

• Limitations
• Greedy nature not suitable for multicast tree.

EWMA Multicast Tree
Minimum TEC Multicast Tree
3
4
4
3
8 EU/pck
4 EU/pck
7 EU/pck
1
1
2
2
2 EU/pck
2 EU/pck
TEC 8 EU/pck
TEC 6 EU/pck
24
Distributed EWMA

• EWMA-Dist
• Two-hop neighbor information
• Using breadth first search, Parent tries to
  reduce TEC by excluding childrens transmission ?
  Shorter and Boarder tree

25
REMiT Approach

• Refinement-based?- (Take an initial solution and
  make it better)
• Needed anyways because of dynamic changes in the
  network
• Interference
• S-REMiT Minimize TEC of source-based tree
• L-REMiT Maximize Lifetime of source-based tree
• G-REMiT Minimize TEC of group-shared tree
• How to distribute the computation?

26
Refinement Operation Change

• Reduce TEC of the source-based tree by moving
  node xs farthest child (say node i) to another
  node (say node j)

27
Refinement Criterion
28
Oscillation Disconnection Avoidance

• Lemma 1 Nodes j and x are the only nodes in the
  source-based multicast tree whose node cost may
  be affected by .
• Lemma 2 If j is not a descendant of node i in
  tree T, then the tree remains connected after
  .

29
S-REMiT Algorithm

• Minimizing TEC of source-based multicast tree
• Two phases
• First Phase Build an initial tree
• Second Phase
• Every node starts local refinement
• Once node i hears its neighbor just made
  refinement, it locks all of its neighbors.
• Node i selects the new parent for itself with the
  highest positive energy Gain, say node j.
• Node i changes its parent from x to j. (Node x
  may be pruned if it is leaf node and not in the
  group.)
• Node i unlocks its neighbors

30
L-REMiT Algorithm

• Maximizing LT of source-based multicast tree
• Two phases
• First Phase Build an initial tree
• Second Phase
• Find bottleneck node x in the tree, node i is the
  costliest node of node x.
• Node i selects the new parent for itself with the
  highest positive Lifetime LTGain, say node j. If
  no such node j exists, terminate L-REMiT.
• Node i changes its parent from x to j (Node x may
  be pruned if it is leaf node and is not in the
  group).
• Recompute the bottleneck node, go to step 1.

31
Performance Results
32
Performance Results
33
Performance Results
34
A Framework for Energy-Efficient Multicast
Policies QoS Requirement
Protocol of Constructing/Maintenance
Energy-Efficient Multicasting Tree
Tree Cost Computation
Node Cost Computation
Energy Cost Model
Link layer parameters feedback (mobility, link
error rate, etc)
35
A Framework for Energy-Efficient Multicast

• Energy Cost Model Nature of wireless
  transceivers
• long range
• short range radios
• Node Cost Computation QoS constraints (delay),
  optimization goals (TEC, LT), type of multicast
  trees (source-based, group-shared)
• Cross layer design combine network layer and
  link layer

36
Conclusions

• Wireless Multicasting is different from Wired
  Multicasting Wireless Multicast Advantage
• Energy-efficient multicast protocols
• Power control
• Different optimization goals Lifetime, Energy
• Type of trees source-based, group-shared
• Adaptive protocols
• Framework for energy-efficient multicasting
• Evaluation on actual wireless (sensor) ad hoc
  e.g. Berkeley Mica Motes

37
Reference

• 1 B. Wang and S. K. S. Gupta. S-REMiT
  S-REMiT A Distributed Algorithm for
  Source-based Energy Efficient Multicasting in
  Wireless Ad Hoc Networks , In Proceedings of
  IEEE GlobleCOM, San Francisco, CA, Dec. 2003, pp.
  3519-3524
• 2 B. Wang and S. K. S. Gupta,"On Maximizing
  Lifetime of Multicast Trees in Wireless Ad hoc
  Networks", In Proceedings of Thirty-Second
  International Conference on Parallel Processing
  (ICPP), Kaohsiung, Taiwan, China, October 2003,
  pp. 333-340.
• 3 B. Wang and S. K. S. Gupta, "G-REMiT An
  Algorithm for Building Energy Efficient of
  Multicast Trees in Wireless Ad Hoc Networks", In
  Proceedings of IEEE International Symposium on
  Network Computing and Applications (NCA),
  Cambridge, MA, April 2003, pp. 265-272.
• 4 J. E. Wieselthier, G. D. Nguyen, and A.
  Ephremides, Resource management in
• energy-limited, bandwidth-limited,
  transceiver-limited wireless networks for
  session-based multicasting. Computer Networks,
  39(2)113131, 2002.
• 5 J. E. Wieselthier, G. D. Nguyen, and A.
  Ephremides, Distributed algorithms for
  energy-efficient broadcasting in ad hoc networks,
  Proceedings of MilCom, Anaheim, CA, Oct. 2002.
• 6 M. Cagalj, J.P. Hubaux, and C. Enz.
  Minimum-energy broadcast in All-wireless
  networks NP-completeness and distribution
  issues. In Proceedings of ACM MobiCom 2002, pages
  172 182, Atlanta, Georgia, September 2002.
• 7 F. Li and I. Nikolaidis. On minimum-energy
  broadcasting in all-wireless networks. In
  Proceedings of the 26th Annual IEEE Conference on
  Local Computer Networks (LCN 2001), pages
  193202, Tampa, Florida, November 2001.