CS 268: Lecture 24
Sensor Network
Architecture (SNA)
Ion Stoica
Computer Science Division
Department of Electrical Engineering and Computer Sciences
University of California, Berkeley
Berkeley, CA 94720-1776
1
Sensor Network Protocols Today
Obligatory David Culler Slide…
Appln
EnviroTrack
Hood
FTSP
Transport
Routing
Scheduling
SPIN
TTDD
TORA
CGSR
AODVDSR
ARA
GSR
DBF
DSDV
TBRPF
Resynch
Phy
Ascent
GPSR
SPAN
ReORg
PC
Drip
Arrive
MintRoute
GRAD
GAF
FPS
Yao
SMAC
PAMAS
Link
Trickle
Deluge
MMRP
Topology
TinyDB
Diffusion
Regions
WooMac
TMAC
Pico
WiseMAC
Bluetooth
RadioMetrix
RFM
CC1000
eyes
BMAC
802.15.4
nordic
What if I want to use any two protocols together??
2
Network Model

Patch
Network
Sensor Node
-
Sensor Patch
Gateway

Many resource constrained
Non-homogeneous
Modalities, roles, HW, SW
Power, BW
Transit tier
- Often specialized wireless
- Provides gateways
Transit Network

Client Data Browsing
and Processing
Dense patches of sensing nodes
Internet Tier
- Multiple connections to infra
- Deep storage, proc. Viz
Basestation

Base-Remote Link

Internet

SNA should not require
unconstrained nodes
Should utilize unconstrained nodes
to reduce burden on constrained
ones
Mobility within physically
embedded context
Data Service
3
What is an Architecture?

Architecture is how to “organize” implementations
- What interfaces are supported
- Where functionality is implemented

Architecture is the modular design of the network

Architecture is not the implementation itself
4
Internet vs Sensor Nets
Internet goals









Interconnect separate networks
Resilience to loss and failure
Support many comm. services
Accommodate variety
Distributed management
Cost effective
Low effort attachment
Resource accountability
Network Architecture
Sensor Nets







Resource efficiency
Data centric design
Deal with intermittent
connectivity
Self-managed
Observation, monitoring of
various environments
Cost effective
Scalability
5
Internet vs Sensor Nets
Internet goals









Interconnect separate networks
Resilience to loss and failure
Support many comm. services
Accommodate variety
Distributed management
Cost effective
Low effort attachment
Resource accountability
Network Architecture
Sensor Nets








Dense real world monitoring
Resilience to loss, failure and
noise
Support many applications
Scale to large, small, long
Cost effective
Evolvable in resources
Composable
Security
6
Why not IP?

One or very few applications running on a sensornet vs huge number
running in the Internet

Large variety of traffic patterns (most not point-to-point):
- Any-to-any, many-to-one, many-to-few, one-to-many
- Inneficient to impl. these patterns over point-to-point

IP does not address (well):
-
Resource and energy constraints
Unattended operation
Intermittent connectivity
Space embeded nodes
...
7
A Sensor Network Architecture (SNA)

Narrow waist: Sensornets Protocol (SP)
-
Goals: generality and efficiency
Position: between data-link and network layers
Service: best-effort, single hop
Common to both single- vs multiple-hop deployments
8
Properties of SP

SP provides mechanisms for network protocols to operate
- Network protocols may introduce policy

Three key elements of SP:
- Data Reception
- Data Transmission
- Neighbor Management
9
Collaborative Interface

Control
- Reliability Best effort to transmit the msg
- Urgency Priority mechanism

Feedback
- Congestion
Was the channel busy?
• Should I slow down?
- Phase
Was there a better time to send?
• Decouple appl sampling from communication
10
Message Reception
Receive
SP



Message arrives from link
SP dispatches
Network protocols establish
- naming/addressing
- filtering
11
Message Transmission
Send
Receive
Msg Pool
SP

Messages placed in shared message pool
- All entries are a promise to send a
packet in the future

Messages include
- Pointer to first packet and # of packets
- Control information: reliability and urgency
- Feedback information: congestion and phase
12
Neighbor Management
Neighbors
Neighbor Table
Send
Receive
Msg Pool
SP

SP provides a shared neighbor table
- Cooperatively managed
- SP mediates interaction using table
• No policy on admission/eviction by SP
• Scheduling information
13
SP Architecture
Network
Service
Manager
Network
Protocol 1
Neighbors
Network
Protocol 3
Receive
Msg Pool
SP Adaptor B
Data Link B
Link
Estimator
PHY A
Link
Estimator
Data Link A
Send
Neighbor Table
SP
SP Adaptor A
Network
Protocol 2
PHY B
14
Neighbor Table
Neighbor
Required
Link
Message Pool
Network
sp_message_t
control
destination
message
quantity
urgent
reliability
feedback
1
phase
D adjustment
congestion true or false
2
address
time on
time off
listen
quality
address_t
local time node wakes
local time node sleeps
true or false
estimated link quality
Neighbor Table
address_t
1st TOSMsg to send
# of pkts to send
on or off
on or off
Msg Pool
SP
15
SP Message Futures
1)
Network Protocol
SP Message
1st
packets
2)
3)
packet
(1)
(5)
Next Packet
Handler
4)
5)
6)
(6)
Send
Submit an SP Message for
Transmission
Message added to message pool
SP requests the link transmit the
1st packet
Link tells SP the transmission
completed
SP asks protocol for next packet
Protocol updates packet entry in
message pool
Msg Pool
SP
(2)
Message
Dispatch
(3)
msg*
(4)
Link Protocol
16
What SP Isn’t

SP does not dictate any header fields
- Messages are opaque to SP

Instead, rely on abstract data types
- Can query for address, length, etc

No explicit security mechanism
- Message content opaque to SP
- Link, Network, and App security can be built transparently to SP
17
Benchmarks

Minimal performance reduction in single hop
- Compare to B-MAC paper
- Compare to IEEE 802.15.4

Simpler multihop/network protocol code

Power consumption

Network protocol co-existence
18
Results: mica2 Throughput
16000
14000
0.9
12000
0.8
0.7
10000
0.6
8000
0.5
6000
0.4
4000
2000
0
0
0.3
B-MAC
SP
SP + CC
SP + LPL + CC
SP + LPL + CC + Phase
Channel Capacity
5
10
Nodes (n)
Percentage of Channel Capacity
Throughput (kbps)
1
0.2
0.1
15
0
20
19
Results: mica2 Throughput
16000
14000
0.9
12000
0.8
0.7
10000
0.6
8000
0.5
6000
0.4
4000
2000
0
0
0.3
B-MAC
SP
SP + CC
SP + LPL + CC
SP + LPL + CC + Phase
Channel Capacity
5
10
Nodes (n)
Percentage of Channel Capacity
Throughput (kbps)
1
0.2
0.1
15
0
20
20
Results: mica2 Throughput
16000
14000
0.9
12000
0.8
0.7
10000
0.6
8000
0.5
6000
0.4
4000
2000
0
0
0.3
B-MAC
SP
SP + CC
SP + LPL + CC
SP + LPL + CC + Phase
Channel Capacity
5
10
Nodes (n)
Percentage of Channel Capacity
Throughput (kbps)
1
0.2
0.1
15
0
20
21
Results: mica2 Throughput
16000
14000
0.9
12000
0.8
0.7
10000
0.6
8000
0.5
6000
0.4
4000
2000
0
0
0.3
B-MAC
SP
SP + CC
SP + LPL + CC
SP + LPL + CC + Phase
Channel Capacity
5
10
Nodes (n)
Percentage of Channel Capacity
Throughput (kbps)
1
0.2
0.1
15
0
20
22
Results: mica2 Throughput
16000
14000
0.9
12000
0.8
0.7
10000
0.6
8000
0.5
6000
0.4
4000
2000
0
0
0.3
B-MAC
SP
SP + CC
SP + LPL + CC
SP + LPL + CC + Phase
Channel Capacity
5
10
Nodes (n)
Percentage of Channel Capacity
Throughput (kbps)
1
0.2
0.1
15
0
20
23
Results:
Single Hop Benchmarks (802.15.4)
24
Conclusion

SNA: provide context for sharing our community work and
accelerate the development and deployment of sensornet
applications

Effective link abstraction, SP, allows network protocols to
run efficiently on varying power management schemes
25