Service Discovery in Highly Heterogeneous Mobile Environments

1
Service Discovery in Highly Heterogeneous Mobile
Environments

• Carlos Flores, Gordon Blair, Paul Grace
• c.florescortes, gordon, gracep_at_comp.lancs.ac.uk
• Monday Reading Group
• 26th Jun 2005

COMPUTING DEPARTMENT
2
Agenda

• Motivation
• Existing middleware solutions
• Research Goal
• Ad-hoc Service discovery protocols, common
  functionality
• Service discovery framework
• Future work
• Conclusions

3
Motivation Ad-hoc Environments

• Service discovery protocols make it possible for
  participating nodes in a network to locate and
  advertise services.

SLPB
S3
ALLIA

• Heterogeneity of discovery platforms

S2
SLPB
SSD
ALLIA
S1

• Dynamic discovery mechanisms are required by
  mobile nodes to cope with the heterogeneity of
  discovery platforms

SSD
GSD
?
S1
S2
GSD
S3
S4
S4
A SERVICE is an entity that can be used by a
person, a software program, or another service.
It may be a computation, storage, a communication
channel to another user, a software filter, a
hardware device, or another user (Liang Cheng).
1.- Motivation 2.- Existing middleware solutions
3.- Goal 4.- Ad-hoc SDP 5.- SD Framework 6.-
Future work 7.- Conclusions
4
Existing Middleware SolutionsReMMoC INDISS
UPnP
SLP
1.- Motivation 2.- Existing middleware solutions
3.- Goal 4.- Ad-hoc SDP 5.- SD Framework 6.-
Future work 7.- Conclusions
5
Research GoalService Discovery Framework

• A component framework approach for the
  development of a configurable and dynamically
  reconfigurable multi-personality discovery
  middleware for operation in mobile ad-hoc
  networks
• Common functionality
• Benefits
• Common component architecture
• Deployment
• Minimization of resource usage by component
  re-use
• Simplification of the configuration and
  reconfiguration process for multiple concurrent
  protocols.
• Development
• Code re-use to simplify the development of
  existing or new discovery protocols.
• Suite of protocols which can easily be plugged
  into a reconfigurable architecture.

1.- Motivation 2.- Existing middleware solutions
3.- Goal 4.- Ad-hoc SDP 5.- SD Framework 6.-
Future work 7.- Conclusions
6
Ad-hoc Service Discovery ProtocolsDiscovery
Protocols Suite

• Group Service Discovery (GSD)
• A novel group-based service discovery protocol
  for manets
• Dipanjan Chakraborty, Anupam Joshi, Yelena Yesha,
  Tim Finin
• University of Maryland, Baltimore County
• Scalable Service Discovery (SSD)
• Scalable service discovery for MANET
• Francoise Sailhan, Valérie Issarny
• INRIA-Rocquencourt, France
• Alliance Based Service Discovery (ALLIA)
• Allia Alliance-based service discovery for
  ad-hoc environments
• Olga Ratsimor, Dipanjan Chakraborty, Anupam
  Joshi, Timothy Finin
• Univeristy of Maryland, Baltimore County
• SLP-Based Service Discovery (SLP-B)
• SLP-based service management for dynamic ad-hoc
  networks
• Stefan Penz
• Aachen University, Germany

1.- Motivation 2.- Existing middleware solutions
3.- Goal 4.- Ad-hoc SDP 5.- SD Framework 6.-
Future work 7.- Conclusions
7
Ad-hoc Service Discovery ProtocolsGSD, SSD,
ALLIA and SLP-B - Common Core Functionality
Pull-based mode
Advertisement
Push-based mode
Centralized
Cache
Decentralized
Request
Reply
Policy
Network

• Cache refreshment rate
• Cache size
• Cache replacement policy
• Advertisement rate
• Vicinity diameter
• Forwarding policy
• Directory deployment policy

• Routing Schemes
• Multicast
• Unicast
• Bordercast
• Anycast
• Ad-hoc routing

Vicinity 2 hops
1.- Motivation 2.- Existing middleware solutions
3.- Goal 4.- Ad-hoc SDP 5.- SD Framework 6.-
Future work 7.- Conclusions
8
Service Discovery FrameworkComponents, Tasks and
Protocols Matching
ALLIA
1.- Motivation 2.- Existing middleware solutions
3.- Goal 4.- Ad-hoc SDP 5.- SD Framework 6.-
Future work 7.- Conclusions
9
Service Discovery FrameworkComponents, Tasks and
Protocols Matching
1.- Motivation 2.- Existing middleware solutions
3.- Goal 4.- Ad-hoc SDP 5.- SD Framework 6.-
Future work 7.- Conclusions
10
Service Discovery FrameworkEvaluation
GSD
SSD
SLP-B
ALLIA
1.- Motivation 2.- Existing middleware solutions
3.- Goal 4.- Ad-hoc SDP 5.- SD Framework 6.-
Future work 7.- Conclusions
11
Service Discovery FrameworkEvaluation
?
coordinator
1.- Waits for directory advertisement
2.- Send election advertisement
SSD
3.- Receive candidates
4.- Elects and notify directory
5.- Central directory notifies its presence
1.- Motivation 2.- Existing middleware solutions
3.- Goal 4.- Ad-hoc SDP 5.- SD Framework 6.-
Future work 7.- Conclusions
12
Service Discovery FrameworkEvaluation
GSD
Service Advertisement Message Structure
ID
TYPE
ADV. DIAMETER
HOPCOUNT
SSD
SOURCE ADD
SERVICE DESC.
SERVICE GROUPS
OTHER GROUPS
Ontology (DAMLOIL)
ID
TYPE
ADV. DIAMETER
SOURCE ADD
SERVICE DESC.
WSDL
1.- Motivation 2.- Existing middleware solutions
3.- Goal 4.- Ad-hoc SDP 5.- SD Framework 6.-
Future work 7.- Conclusions
13
Future Work

• Fine-grained architecture to increase code
  re-usability and configurability.
• Design and development of a monitor component to
  allow applications to automatically identify
  discovery protocols availabe in the environment.

REQUEST

Lookup
1.- Motivation 2.- Existing middleware solutions
3.- Goal 4.- Ad-hoc SDP 5.- SD Framework 6.-
Future work 7.- Conclusions
14
Future Work

• Evaluate our architecture with discovery
  protocols from different network styles
• Ad-hoc
• Bluethoot
• Fixed networks
• SLP, Salutation, Jini, UPnP, Twine
• Internet
• UDDI
• Overlay networks
• JXTA, Pastry
• Development of a middleware to provide service
  discovery interoperability across different
  network styles

1.- Motivation 2.- Existing middleware solutions
3.- Goal 4.- Ad-hoc SDP 5.- SD Framework 6.-
Future work 7.- Conclusions
15
Conclusions

• Dynamic discovery mechanisms are required by
  mobile nodes to cope with the heterogeneity of
  discovery platforms.
• Existing solutions like ReMMoC and Indiss
• Significant development effort is required for
  the development of each individual protocol.
• It is not possible to re-use components at
  development or run-time.
• Complex configuration and reconfiguration
  processes.

1.- Motivation 2.- Existing middleware solutions
3.- Goal 4.- Ad-hoc SDP 5.- SD Framework 6.-
Future work 7.- Conclusions
16
Conclusions

• Service discovery framework for the development
  of an adaptive multi-personality service
  discovery middleware.
• Common functionallity
• Common architecture
• Framework evaluation
• GSD, SSD, ALLIA, SLPB
• Re-used components Cache, Policy and Network
• More fine grained architecture is required
  Advertisement, Request, Reply.
• Future work
• Evaluate our architecture with discovery
  protocols from different network styles.
• Development of a middleware to provide service
  discovery interoperability across different
  network styles

1.- Motivation 2.- Existing middleware solutions
3.- Goal 4.- Ad-hoc SDP 5.- SD Framework 6.-
Future work 7.- Conclusions
17
Questions?
Carlos Flores, Gordon Blair, Paul Grace
c.florescortes, gordon, gracep_at_comp.lancs.ac.uk