Sensor Network Routing

1
Sensor NetworkRouting Data Dissemination

• Lynn Choi
• Korea University

2
Characteristics Like MANET

• No static infrastructure
• No routers, base stations, or access points
• Dynamic Topology
• The topology of the network changes very
  frequently due to
• Node power-down to save energy
• Node failures
• Sensors may be inaccessible, i.e. thrown into
  inhospitable terrain
• New nodes can be additionally deployed
• Node mobility
• The validity of path information in a router can
  change spontaneously without warning

3
Characteristics Unlike MANET

• Lack of global identification
• Sensor nodes may not have global identification
  (i.e. IP address) due to the large amount of
  overhead and large number of sensors
• Lead to data-centric (content-based) addressing
  and routing
• Limited in energy, computation, and memory
• Energy-aware routing is a must!
• The number of nodes can be several orders of
  magnitude higher
• Require scalable protocols
• Multiple sensors collaborate to achieve one goal
• Difficult to pay special attention to any
  individual node
• Collecting information within the specified
  region
• Collaboration/aggregation among neighbors within
  a geographical region

4
Why Different and Difficult?

• Data centric (content-based) addressing and
  routing
• Data centric addressing and routing
• Where are nodes whose temperatures exceed more
  than 30 degrees for the past 10 minutes?
• Tell me the location of the object every 100ms
  for 2 minutes
• High level tasks are needed
• At what speed and in what direction was that
  elephant traveling?
• Is it the time to order more inventory?
• Data aggregation and caching
• Intermediate nodes can perform data aggregation
  and caching in addition to routing to minimize
  communication and thus to save energy
• where, when, how?

5
Challenges

• Goal Minimize energy dissipation
• Energy-limited nodes
• Computation
• Aggregate data
• Suppress redundant routing information
• Communication
• Bandwidth-limited
• Energy-intensive
• Scalability ad-hoc deployment in large scale
• Fully distributed w/o global knowledge
• Large numbers of sources and sinks

6
Communication Procedures

• Terminology
• Source the node that generates data/event
• Sink the node that are interested in the
  data/event
• Interest a descriptor for a particular kind of
  data/event
• 1/2/3 step procedures
• Data advertisement
• Source broadcasts/advertises the availability and
  meta-data of data
• Query/Interest propagation
• Sink broadcasts its interests
• Data propagation
• Source sends the data to sink if an event is
  detected
• The event detection can be periodic or on-demand

7
Flooding

• Idea
• Each node broadcasts a packet if the maximum
  hop-count of the packet is not reached and the
  node itself is not the destination
• Does not require topology maintenance or route
  discovery
• Disadvantages
• Implosion
• A node receives copies of the same message
• Overlap
• The same event may be sensed by more than one
  node due to overlapping regions of coverage
• Resource blindness
• The protocol does not consider the available
  energy at the nodes, reducing the network lifetime

8
Gossiping

• A modified version of flooding
• Instead of broadcasting, send it to a randomly
  selected neighbor
• Plus
• Avoid the problem of implosion
• Lower overhead, less traffic
• Minus
• Longer delay
• Does not guarantee the delivery

9
Directed Diffusion

• Data centric
• Data is named by attribute-value pairs (rather
  than node address)
• A node requests data by sending interests for
  named data
• Detect vehicle location in 100,100 and send me
  events every 20ms.
• Data diffusion procedure
• Interest propagation
• Sinks broadcast interest to neighbors
  (request-driven)
• Gradients are set up pointing back to where
  interests came from
• Interests and data are cached by intermediate
  nodes
• Data propagation
• Once a source receives an interest, it routes
  measurements along gradients
• Reinforcement
• After sink starts receiving events, it enforces a
  particular (low-delay) path to receive
  high-quality data
• Gradients from Source (S) to Sink (N) are
  initially small but increase during reinforcement
• Supports multiple sources and multiple sinks

10
Directed Diffusion
11
TTDD (Two-Tier Data Dissemination Protocol)

• Mobile sinks bring new challenges
• The location of a mobile sink needs to be
  continuously propagated throughout the sensor
  field
• Frequent location updates from multiple mobile
  sink lead to
• Excessive energy drains of sensor nodes
• Increased collisions
• Two-tier forwarding model
• Assumption
• Sensor nodes are stationary and location-aware
  (GPS-enabled)
• Each data source proactively builds a grid
  structure
• Each source forwards its data to a set of sensor
  nodes called disseminating nodes (at grid points)
• This enables each mobile sink to receive data by
  flooding in each local cell only
• Localize impact of sink mobility on data
  forwarding

12
TTDD (Two-Tier Data Dissemination Protocol)
Dissemination Node
Data Announcement
Data
Query
Immediate Dissemination Node
13
TTDD (Two-Tier Data Dissemination Protocol)
Dissemination Node
Trajectory Forwarding
Data Announcement
Immediate Dissemination Node
Data
Immediate Dissemination Node