[aosd-announce] PhD position in SOA and GreenIT

[Thomas Ledoux]

A PhD position is open with Ascola (www.emn.fr/x-info/ascola), a joint 
project-team between INRIA and LINA, Nantes, France.
Director: Pierre Cointe ; Advisor: Thomas Ledoux

"QoS Self-Optimization in Service-oriented Architectures for Energy 
Consumption Management"

Context
======
Web services adoption to build software architectures for grid 
computing, mobile networks, mashup environments, etc. require “agile” 
solutions able to dynamically evolve to manage resources instability 
(bandwidth, response time, storage capability, workload, etc.). The goal 
is to propose some guarantees to end-users on QoS but also to optimize 
the QoS for context-aware architectures.
Today, energy consumption is a critical issue for IT organizations and 
energy efficiency is becoming one of the main QoS criteria in hardware 
and software administration. Several companies develop assessment tools 
for designing energy-efficient data centers but it could be interesting 
to reify energy consumption in software development. Indeed, 
Service-oriented Architectures (SOA) build applications out of software 
services dealing with computation, data and finally energy consumption. 
An attractive idea could be to take energy efficiency into account in 
Service Level Agreement (SLA) and to dynamically adapt the architecture 
for optimizing energy.
However, current work in SOA architectures do not propose reliable 
dynamic compositions of different QoS concerns and there has been little 
research on energy consumption management.

Goal
====
We propose to tackle this challenge with a language-based approach. 
Recent work advocates aspect-oriented programming to add specific 
concerns (e.g. reliable messaging) on web services composition or a 
declarative approach to supervise the QoS of a web services 
orchestration with self-healing capabilities. However, the different 
solutions are limited (complex, not enough expressive, not reliable, 
syntax verification only with XML). We plan to use a Domain-Specific 
Languages (DSL) approach to solve these different issues: QoS concepts 
will be expressed concisely and their properties will be formally analyzed.
Then, to build an “agile” SOA architecture, able to automatically adapt 
to changes in the execution context, we propose to develop an autonomic 
infrastructure following the Autonomic Computing Initiative. The 
resulting infrastructure will be able to self-optimize the QoS and more 
specifically the energy criteria to ensure the most efficient energy 
consumption management. We will experiment Green IT use cases in the 
galaxy project (http://galaxy.gforge.inria.fr) aiming at providing an 
open SOA platform.

Candidate Profile
============
The PhD candidate should have good skills in the design of programming 
languages; be knowledgeable in the domains of software engineering, web 
services, autonomic computing.

Application
========
Applicants should send a curriculum vitae, a cover letter, 2-3 reference 
letters to Thomas Ledoux (ledoux at emn.fr)
All applications must be received as soon as possible.