INSTITUT D’INFORMATIQUE 1 R&D on Pervasive/Ubiquitous Computing at the Institute of Informatics, University of Namur, Belgium Presentation to ERCIM Emobility.

INSTITUT D’INFORMATIQUE 1

R&D on Pervasive/Ubiquitous Computingat the Institute of Informatics,University of Namur, Belgium

Presentation to ERCIM Emobility working groupKickoff meeting,

Basel, Switzerland, October 27 2006

INSTITUT D’INFORMATIQUE 2

General introduction

The Institute of Informatics Established in 1970 Bachelor, Master, PhD programs in Information

Systems, Software Engineering and Computer Science

Staff• 15 professors• 50 research assistants• 10 PhD students• 7 administrative and support staff members

Students• 300+ currently enrolled• 1,500+ graduated since 1970

INSTITUT D’INFORMATIQUE 3

LScIn

fras

truc

ture

s

CoordinationOptimisation

Pervasivecomputing

JMJWVa

Rad

io R

esou

rce

Man

agem

ent

Java Optim

isation

Competing systems optimisation3 fieldsof expertise Infrastructures Coordination models Code analysis

and optimisation3 research focuses RRM Random access networks Java bytecode optimisation3 professors Jean-Marie JACQUET (JMJ) Laurent SCHUMACHER (LSc) Wim VANHOOF (WVa)

Pervasive/Ubiquitous Computing

INSTITUT D’INFORMATIQUE 4

Field expertise #1 – Infrastructures

NodeB #1

NodeB #2

NodeB #3

NodeB #4

RNC

UEs

INSTITUT D’INFORMATIQUE 5

Field expertise #2 – Coordination models

Interaction model for asynchronous systems (e.g. web request/response)

Common blackboard for concurrent data sharing between processes

Paradigm extended with• virtual blackboards• real-time primitives• mobile processes

Blackboard

P1 P2 P3

tellgetasknask

Bb #1

P1

Bb #2

Primitives

Processes

INSTITUT D’INFORMATIQUE 6

Field expertise #3 – Code optimisation

Embedded systems have limited capabilities w.r.t. power, memory, CPU. Nevertheless, new applications demand ever more resources.

Code running on such systems should be optimised accordingly

What kind of optimisations?• Improving execution speed• Lowering memory usage• Lowering power usage

In-house expertise on analysis, transformation and optimisation of (declarative) programming languages.

INSTITUT D’INFORMATIQUE 7

Research focus #1 – Radio Resource Management

Spreading code allocation in CDMA-based networks dealt with as a constraint-based programming problem

To be generalised to shared media/resource networks (e.g. FDMA, TDMA, OFDMA)

Orthogonal Variable Spreading Factor (OVSF) tree

INSTITUT D’INFORMATIQUE 8

Research focus #2- Random access networks

Process communication modelled with temporary channels instead of shared blackboards

Birth/death process of channels based on neighbour discovery

Suited for modelling multiple access in ad hoc and Disruption Tolerant (DTN) networks

Model of communicationbetween processes

Temporarycommunication

channel

INSTITUT D’INFORMATIQUE 9

Research focus #3 – Java bytecode optimisationfor embedded systems

Java regarded useful for portability Java run-time environment however not optimised for

embedded systems Some examples

• Java vulnerable to memory leak or stack overflow• Undeterministic behaviour of Java VM due to the garbage

collector which is triggered unpredictably/irregularly→ Memory consumption is hard to predict→ Execution time is hard to predict

Research topics• A Java VM which behaves (more) deterministically• Compile-time garbage collection for Java (as developed for

declarative languages like Haskell, Mercury)

INSTITUT D’INFORMATIQUE 10

Additional information

IPv6 Linux-based UTRAN Testbedhttp://www.info.fundp.ac.be/~hvp/rech/doc/index_en.html

Personal web pageshttp://www.info.fundp.ac.be/~jmj/http://www.info.fundp.ac.be/~lsc/http://www.info.fundp.ac.be/~wva/