Title: Routing In Bluetooth
1Department of Information Engineering University
of Padova, Italy
A note on the use of these ppt slidesWere
making these slides freely available to all,
hoping they might be of use for researchers
and/or students. Theyre in PowerPoint form so
you can add, modify, and delete slides
(including this one) and slide content to suit
your needs. In return for use, we only ask the
followingIf you use these slides (e.g., in a
class, presentations, talks and so on) in
substantially unaltered form, that you mention
their source.If you post any slides in
substantially unaltered form on a www site, that
you note that they are adapted from (or perhaps
identical to) our slides, and put a link to the
authors webpage www.dei.unipd.it/zanella Thank
s and enjoy!
COST273 May 30-31, 2002 Helsinki
TD (02)-062
2On the performance of AODV and FSR routing
algorithms on Bluetooth scatternets preliminary
results
Department of Information Engineering University
of Padova, Italy
COST273 May 30-31, 2002 Helsinki
TD (02)-062
3Outline of the contents
- Bluetooth basic
- Ad-hoc routing algorithms
- Ad-hoc On demand Distance Vector (AODV)?
- Fisheye State Routing (FSR)?
- Simulation model
- Experimental results
- Conclusions and future work
4Bluetooth Technology
- What is Bluetooth?
- A wireless technology
- Proposed as cable replacement for leakage
portable electronic devices, BT provides
short-range low-power point-to-(multi)point
wireless connectivity - A global industry standard in the making
- Initially developed by Ericsson, now BT is
promoted by an industry alliance called Special
Interest Group (SIG)?
5Bluetooth piconet
- Two up to eight Bluetooth units sharing the same
channel form a piconet - In each piconet, a unit acts as master, the
others act as slaves - Channel access is based on a centralized polling
scheme
6FH TDD
- Each piconet is associated to frequency hopping
(FH) channel - The pseudo-random FH sequence is imposed by the
master - Time is divided into consecutive time-slots of
625 ?s - Each slot corresponds to a different hop
frequency - Full-duplex is supported by Time-division-duplex
(TDD) - Master-to-slave (downlink) transmissions start on
odd slots - Slave-to-Master (uplink) transmissions start on
even slots
7Bluetooth scatternets
- Piconets can be interconnected by Inter-piconet
Units (IPUs)? - IPUs may act as gateways, forwarding traffic
among adjacent piconets - IPUs must time-division their presence among the
piconets - Time division can be realized by using SNIFF mode
8Next in the line
- Bluetooth basic
- Ad-hoc routing algorithms
- Ad-hoc On demand Distance Vector (AODV)?
- Fisheye State Routing (FSR)?
- Simulation model
- Experimental results
- Conclusions and future work
9Motivations of the work
- Bluetooth gets out typical MANET scenario
- Physical proximity does not imply connection
- Connection set-up may take infinite time
- Broadcast is supported only within piconets
- Hence, MANET algorithms have to be tested in
Bluetooth environment - Table-driven algorithms LSR, DSDV,WRP,FSR
- On demand algorithms DSR,TORA,AODV?
10IP-layer routing
- Routing is performed at IP layer, making use of
IP addresses - Pros
- No address-mapping
- Independence of the network details
- Cons
- Each node must support IP functionalities
- IP datagrams must be reassembled before forwarding
11AODV Algorithm (1)?
- Route discovery
- Broadcast Route Request packet (RREQ), containing
source and destination IP addresses - Intermediate nodes that receive the RREq for the
first time - Increment by one the hop count field in the
packet - Add an entry containing source IP, destination
IP, predecessor IP - Broadcast the RREQ packet
IP5
IP3
Source node IP1
Destination node IP7
IP4
IP6
IP2
12AODV Algorithm (2)?
- Destination responds to the RREQ by unicasting a
Route Reply (RREP) packet to the source - The RREP flows backward along the path traced by
the RREQ - Intermediate nodes that process the RREP update
their entry - Entries that are not updated expire after a given
timeout
IP5
IP3
Source node IP1
IP4
IP7 Destination node
IP6
IP2
13AODV Algorithm (3)?
- In case of link failure, AODV propagates a Route
Error (RERR) message to the upstream nodes - Receiving an RERR, nodes set to infinity the
distance to the destination - If the path is still needed, nodes start a new
path discovery procedure
IP5
X
IP3
Source node IP1
X
IP4
IP7 Destination node
IP2
14FSR Algorithm (1)?
- Each node maintains link state information for
every other node - FSR generates route update on a periodic basis
- Routing information is propagated to neighbours
only - Updates occur on the basis of the Fisheye
algorithm - Nodes are divided in scopes, on the basis of
their distance to the source - Routing information for a given destination is
updated with a frequency that is inversely
proportional to the scope of the destination
15FSR Algorithm (2)?
- Fisheye scope set of nodes within a given number
of hops - Closer nodes are update more frequently than
farther ones
16FSR Algorithm (3)?
- Getting close to the destination, the routing
information becomes progressively more accurate
17Next in the line
- Bluetooth basic
- Ad-hoc routing algorithms
- Ad-hoc On demand Distance Vector (AODV)?
- Fisheye State Routing (FSR)?
- Simulation model
- Experimental results
- Conclusions and future work
18Simulation platform
- Simulator Tool OPNET Modeler Ver. 8.0
- The simulator does support
- Baseband protocols
- Frequency Hopping, Paging, Inquiry, Scan
- Link manager (LM) protocol
- Link layer control and adaptation protocol
(L2CAP) - Connection setup/release, Sniff Mode
- The simulator does not support
- Handover for Bluetooth units
- Multi-slot data packets
19Model assumptions
- Pre-formed Scatternet
- Roles of master/slave/gateway are preassigned
- Pure Round Robin polling strategy
- Nodes in a piconet have the same priority and get
polled in cyclic order - 2 piconets per IPU
- IPU divides it time equally between the piconets
by means of the sniff mechanism - IPUs are not coordinated
- Network layer routing algorithms AODV FSR
20Scatternet topology
21Next in the line
- Bluetooth basic
- Ad-hoc routing algorithms
- Ad-hoc On demand Distance Vector (AODV)?
- Fisheye State Routing (FSR)?
- Simulation model
- Experimental results
- Conclusions and future work
22Average end-to-end delay (1)?
- Simulation parameters
- CBR traffic 20 kbit/sec
- 5 hops connection
- 2 gateway units (IPUs)?
- Results
- End-to-end delay grows almost linearly with the
Sniff period - Short IP datagrams better exploit the pipeline
effect
23Average end-to-end delay (2)?
- Simulation parameters
- Sniff period 100 slots
- 5 hops connection
- Results
- Short IP dtgs achieve lower end-to-end delay but
saturate earlier - Long IP dtgs incur in higher end-to-end delay but
increase capacity utilization
24AODV route discovery delay (1)?
- Simulation parameters
- Sniff period 100 slots
- Route length increasing
- Results
- Using Link Layer (AODV-LL) messages to refresh
table entries the discovery time is shorter - Path discovery query ends one hop earlier
- Route discovery delay grows almost linearly with
the distance of the destination
25AODV route discovery delay (2)?
- Simulation parameters
- Sniff period ranges from 50 to 200 slots
- 5 hops connection
- Results
- Route discovery delay grows almost linearly with
the Sniff period - AODV-LL shows much better performance than
AODV-std - Impact of Sniff period is higher on longer path
26Fisheye control traffic
- Simulation parameters
- Sniff time 50 slots
- Different Refresh periods
- Different number of scopes
- Remark Scope0 is the Global State Routing
- Results
- As expected, the more the number of scopes the
less the control traffic - Refresh Times less than 0.1s absorb more than 10
of the system capacity - Refresh time must be longer than 0.1s, but this
increases the route updating delay
27Fisheye update delay (1)?
- Simulation parameters
- 5 hops connection
- 2 scopes
- Sniff period 50 slots
- Refresh period ranges from 0.1 to 0.25 s
- Results
- A refresh period of 0.1 s updates nodes 6-hop
faraway from the source within 1 s
28Fisheye update delay (2)?
- Simulation parameters
- 2 scopes
- Sniff perriod ranges from 50 to 300 slots
- Refresh period0.25 s
- Path length increasing
- Results
- Update delay mainly due to the path length
- Sniff period has smaller impact
29Next in the line
- Bluetooth basic
- Ad-hoc routing algorithms
- Ad-hoc On demand Distance Vector (AODV)?
- Fisheye State Routing (FSR)?
- Simulation model
- Experimental results
- Conclusions and future work
30Final Remarks
- Datagram size and Sniff period have a
considerable impact on IP end-to-end delay and
AODV route discovery time - FSR refresh time must be carefully chosen
- AODV appears suitable in case of
- Sparse connections
- Relaxed latency constraints
- Semi-static topology
- FSR appears more convenient for
- Dense connections
- Dynamic and wide topology
31Future work
- Coming Soon (maybe)?
- Mathematical analysis of the scatternet
efficiency - Simulator enhancements
- Multi-slot packets
- Handover
- Comparison with Link Layer Routing algorithms
- Implementation of dynamic scatternet formation
algorithms
32Spare slides
33Table-Driven algorithms
- Each node maintains one or more tables with
routing information for every other node - Nodes periodically exchange tables information
- Algorithms differ for the number of
routing-related tables and updating strategy - Examples
- Fisheye State Routing (FSR)
- Hierarchical State Routing (HSR)?
- Wireless Routing Protocol (WRP)?
34On Demand algorithms
- Nodes maintain route information only to the
nodes for which there is an actual need - Routes are built on a demand basis, by means of a
route discovery mechanism - Changes on the network topology are propagated
only to the interested nodes - Examples
- Ad hoc On demand Distance Vector (AODV)?
- Cluster Based Routing (CBR)?
- Dynamic Source Routing (DSR)
- Associativity Based Routing (ABR)?
35Routing on Bluetooth scatternet
- Bluetooth baseband packets contain AMA_ADDR
- Temporary and local meaning only
- Routing must be performed above the baseband
layer! - Possible approaches
- Data Link Layer Routing performed between L2CAP
and IP - Routing Vector Method (RVM)?
- Bluetooth Network Encapsulation Protocol (BNEP)?
- Network Layer routing performed at IP layer
- MANET algorithms
36Data Link layer routing
- RVM lies above L2CAP and beneath IP
- It uses 3 bit local IDs to identify the piconets
- Routing is performed by means of the routing
vector mechanism - Pros simplicity, bandwidth conservation, low
resources requirement - Cons Topological changes determine address re-map