Network Kernel Architectures and Implementation ( ) Network Architecture Chaiporn Jaikaeo Department of Computer Engineering Kasetsart University Materials taken from lecture slides by Karl and Willig

2 Outline Network scenarios Network scenarios Optimization goals Optimization goals Design principles Design principles Gateway concepts Gateway concepts

3 Typical Views of WSN Self-organizing mobile ad hoc networks (MANETs) Self-organizing mobile ad hoc networks (MANETs) Peer-to-peer networks Peer-to-peer networks Multi/mobile agent systems and swarm intellegence Multi/mobile agent systems and swarm intellegence

4 Sensor Network Scenarios Sources: Any entity that provides data/measurements Sources: Any entity that provides data/measurements Sinks: Nodes where information is required Sinks: Nodes where information is required Source Sink Internet Sink Source Sink Source

5 Single-Hop vs. Multi-hop Multi-hop networks Multi-hop networks  Send packets to an intermediate node  Intermediate node forwards packet to its destination  Store-and-forward multi-hop network Store & forward multi-hopping NOT the only possible solution Store & forward multi-hopping NOT the only possible solution  E.g., collaborative networking, network coding Source Sink Obstacle

6 Multi-hopping Always Efficient? Obvious idea: Multi-hopping is more energy- efficient than direct communication Obvious idea: Multi-hopping is more energy- efficient than direct communication  Suppose we put a relay at distance d /2  Energy for distance d is reduced from cd  to 2 c ( d /2)   c - some constant   - path loss coefficient (  2) Usually wrong, or over-simplified Usually wrong, or over-simplified  Need to take constant offsets for powering transmitter, receiver into account

7 Multiple Sinks, Multiple sources

8 Outline Network scenarios Network scenarios Optimization goals Optimization goals Design principles Design principles Gateway concepts Gateway concepts

9 Goal: Quality of Service QoS in WSN is more complicated (compared to MANET) QoS in WSN is more complicated (compared to MANET)  Event detection/reporting probability  Event classification error, detection delay  Probability of missing a periodic report  Approximation accuracy (e.g, when WSN constructs a temperature map)  Tracking accuracy (e.g., difference between true and conjectured position of the pink elephant) Related goal: robustness Related goal: robustness  Network should withstand failure of some nodes

10 Goal: Energy efficiency Many definitions Many definitions  Energy per correctly received bit  Energy per reported (unique) event  Delay/energy tradeoffs  Network lifetime  Time to first node failure  Network half-life (how long until 50% of the nodes died?)  Time to partition  Time to loss of coverage  Time to failure of first event notification

11 Sharpening the Drop Sacrifice long lifetimes in return for an improvement in short lifetimes Sacrifice long lifetimes in return for an improvement in short lifetimes

12 Outline Network scenarios Network scenarios Optimization goals Optimization goals Design principles Design principles Gateway concepts Gateway concepts

13 Distributed Organization WSN participants should cooperate in organizing the network WSN participants should cooperate in organizing the network  Centralized approach usually not feasible Potential shortcomings Potential shortcomings  Not clear whether distributed or centralized organization achieves better energy efficiency Option: “limited centralized” solution Option: “limited centralized” solution  Elect nodes for local coordination/control  Perhaps rotate this function over time

14 In-Network Processing WSNs are expected to provide information WSNs are expected to provide information  Gives additional options  E.g., manipulate or process the data in the network Main example: aggregation Main example: aggregation  Apply aggregation functions to a collection tree in a network  Typical functions: minimum, maximum, average, sum, …  Not amenable functions: median

15 Aggregation Example

16 Signal Processing Another form of in-network processing Another form of in-network processing E.g., E.g.,  Edge detection  Tracking/angle detection of signal source Exploit temporal and spatial correlation Exploit temporal and spatial correlation  Observed signals might vary only slowly in time  Signals of neighboring nodes are often quite similar Compressive sensing Compressive sensing

17 Adaptive Fidelity Adapt data processing effort based on required accuracy/fidelity Adapt data processing effort based on required accuracy/fidelity E.g., event detection E.g., event detection  When event occurs, increase rate of message exchanges E.g., temperature E.g., temperature  When temperature is in acceptable range, only send temperature values at low resolution  When temperature becomes high, increase resolution and thus message length

18 Data Centric Networking Interactions in typical networks are addressed to the identities of nodes Interactions in typical networks are addressed to the identities of nodes  Known as node-centric or address-centric networking paradigm In WSN, specific source of events might not be important In WSN, specific source of events might not be important  Several nodes can observe the same area Focus on data/results instead Focus on data/results instead  Data-centric networking  Principal design change

19 Implementation Options Publish/subscribe (NDN – Named Data Networking) Publish/subscribe (NDN – Named Data Networking)  Nodes can publish data, can subscribe to any particular kind of data  Once data of a certain type has been published, it is delivered to all subscribers Databases Databases  SQL-based request

20 Outline Network scenarios Network scenarios Optimization goals Optimization goals Design principles Design principles Gateway concepts Gateway concepts

21 Gateways in WSN/MANET Allow remote access to/from the WSN Allow remote access to/from the WSN Bridge the gap between different interaction semantics Bridge the gap between different interaction semantics  E.g., data vs. address-centric networking Need support for different radios/protocols Need support for different radios/protocols

22 Gateway nodes Internet Gateway WSN tunneling Use the Internet to “tunnel” WSN packets between two remote WSNs Use the Internet to “tunnel” WSN packets between two remote WSNs

6LoWPAN IPv6 over Low-power Wireless Personal Area Networks IPv6 over Low-power Wireless Personal Area Networks Nodes communicate using IPv6 packets Nodes communicate using IPv6 packets An IPv6 packet is carried in the payload of IEEE data frames An IPv6 packet is carried in the payload of IEEE data frames 23

Example 6LoWPAN Systems 24

25 Summary Network architectures for WSNs look quite different from typical networks in many aspects Network architectures for WSNs look quite different from typical networks in many aspects Data-centric paradigm opens new possibilities for protocol design Data-centric paradigm opens new possibilities for protocol design