söndag 25 januari 2015

Sustainability Effects of Cloud Computing

I was part of a group that handed in a EU application some 10 days ago. It wasn't one of those truly huge applications with a dozen partners, but rather a "Marie Curie Innovative Training Networks" application. We applied for money to hire 10 ph.d. students at five nodes in four different countries.

The application is called "Energy-aware cloud-based software engineering" and the application is a marriage between technical and sustainability aspects of (future) cloud computing. Here's a summary of the central challenge in two short sentences: Cloud computing is slated to grow explosively. How can sustainability become part of the criteria by which (future) cloud computing infrastructure is planned and realised? Or more formally (from the application): "[there is a] lack of professional competencies - and hence of proper training - for engineering and deploying sustainable and energy-efficient cloud-based software applications in terms of both software development and system operation." A report from mid-2013 estimates the possibility of major (> 85%) economic and energy savings in migrating business software to the cloud - but how do we make that promise feasible? The proposed answer in our application is through developing the right competencies in a tight collaboration between industry and academia. Each of the five nodes in the application consists of two partners; one academic and one industry partner. The five nodes are:

- Academic institution Vrije Universitet Amsterdam (VUA) and industry/partner organisation GreenIT Amsterdam
- Academic institution University of Amsterdam (UVA) and industry/partner organisation SURFsara
- Academic institution Politecnico di Milano (PDM) and industry/partner organisation CEFRIEL
- Academic institution KTH Royal Institute of Technology (KTH) and industry/partner organisation The Interactive Institute Swedish ICT (TII)
- Academic institution University of Zürich (UZH) and industry/partner organisation Swiss Federal Laboratories for Materials Science and Technology (EMPA)

It's fair to say that the first three academic partners primarily focus on the technical challenges involved, while the last two partners primarily focus on issues pertaining to sustainability. The application is a cooperation though and an important point is that the ph.d. students we hope to recruit will visit other institutions in the network. It is definitely fair to say that the brunt of the work with putting the application together was done by people at the two Dutch universities. Despite this, I will here quickly mention what UZH plans to do and a tiny bit more about the Swedish node.

University of Zürich is primarily represented by professor Lorenz Hilty and "his" two planned ph.d. students will work on "life-cycle assessment" (e.g. investigating the relation between cloud software services and LCA) and "establishing sustainability as a non-functional requirement in software engineering".

The "Swedish node" consists of KTH (me) and The Interactive Institute (Cecilia Katzeff). It was Cecilia who inquired if I wanted to join forces with her and work on the application some time before Christmas. She then went away on vacation for two weeks and when she missed a before-Christmas telephone conference I kind of slipped from being her sidekick to being responsible for the Swedish node.

KTH is responsible for a work package called "Social Sustainability Effects of Cloud Computing Systems". The two proposed KTH ph.d. students will more specifically work on:

"User Behavours in Cloud Computing Systems". The main question here is:
- How do current user behaviour in relation to cloud services affect hardware and software architectures/requirements, efficiency and sustainability aspects of cloud computing?

"Social Sustainability Effects of Cloud Computing Systems". The main two questions here are:
- Which are the stakeholders in the development and deployment of current and future cloud computing architectures and services and which are the economic incentives of primary and secondary stakeholders?
- What are the social aspects (convenience, personalisation, perceived control, influence, surveillance, privacy) of moving from decentralised (p2p) architectures to more centralised cloud architectures and terminals (laptops, tablets, smartphones), i.e. with relatively sparse storage facilities but high bandwidth?

A larger number of technical questions and tasks have also been formulated in the application, for example concerning:
- New paradigms for sustainable cloud eco-systems
- Virtualization strategies
- Energy efficiency metrics and practices
- Applications for hotspot identification
- Multi-party services transparencies
- Hardware-software integration

The application is quite voluminous (as these things tend to be) but this is the only summary I will post on the blog for now. I will of course get back with more information if the application is granted!

Inga kommentarer:

Skicka en kommentar