Computing is increasingly present in our lives, and we rely on digital services and applications for most daily activities, from work to entertainment, and from education to society services and governance. It is predicted that, by year 2030, ICT (Information and Communication Technology) might consume as much as 30% of the worldwide energy. As such, ICT raises significant sustainability concerns and research is needed to increase the energy efficiency of applications and digital services.
One major contribution to increasing the sustainability of ICT is to reduce waste in computing. Informally, we define waste in computing as the inefficient use of computing resources. In this project, we aim to (1) define metrics and methods to measure computing waste reliably and systematically in different applications and systems across the computing continuum, (2) propose and prototype tools to reduce such waste (depending on its causes and target system), and (3) demonstrate our approach (and its impact on ICT sustainability) on several case-studies.
Do you want to join our efforts to increase ICT sustainability? The Computer Architecture for Embedded Systems (CAES) group at the University of Twente invites applications for a 4-year PhD position on the topic of "
Zero-waste computing". We are looking for a talented and focused researcher with knowledge in using/programming distributed systems, cloud/edge computing, experience with empirical research in deploying and analysing real applications running on real (large) heterogeneous systems, and experience with performance and energy measurement and prediction models and tools.
The successful applicant will join University of Twente as PhD student, supervised by Dr. Ir. Ana-Lucia Varbanescu, and will work as part of the CAES group, also in collaboration with a second PhD student working, in the context of the Graph-Massivizer EU project, on high-performance, energy-efficient graph processing applications deployed in the computing continuum. The student will also collaborate closely with VU Amsterdam, the EnergyLabels project at University of Amsterdam, and with SURF; we expect to address the sustainability of applications ranging from machine learning (from UvA, Radboud University) and big science (UvA, Nikhef).
The PhD student will have the opportunity to broaden their knowledge by joining international exchange programs, participating in national and international conferences and workshops, and visiting other research institutes and universities in the project and beyond.
Concretely, the project is foreseen to include the following activities:
- Define metrics (and evaluation processes for these metrics) for quantifying waste in the computing continuum
- Analyse 5-7 case-study applications, including dense and sparse matrix computation, machine-learning workloads, scientific applications, and graph processing workloads. Quantify their energy waste.
- Propose methods to reduce waste across the computing continuum and evaluate them empirically (in simulation, emulation, and/or in the lab).
- Assess trade-offs between functional and non-functional properties, and their impact on zero-waste computing.
- Propose methods to reduce compute waste and assess their efficacy on the selected case-studies.
- Assess the impact of reduced-waste computing on sustainability of ICT.
These steps are further expected to be disseminated in open publications and accompanied with proof-of-concept tools as demonstrators.