Eduardo Almeida and Luís Vilaça (CTM), Carlos Pereira, Carlos Silva, João Pacheco, Ricardo Andrade, Ricardo Silva and Vasco Campos (CPES)

Eduardo Almeida and Luís Vilaça (CTM)

“The CTM coordinators would like to nominate the researcher Luís Miguel Vilaça and Eduardo Nuno Almeida. Luís and Eduardo showed tremendous dedication and team spirit during the implementation and configuration of the SLURM platform on the Centre’s computing resources, while working as grants holders within the scope of the project and carrying out their PhD work. Their contribution to the design of this solution, and further configuration and problem solving, were crucial to address the complexity of this process. We’d like to highlight the organisation of the training action ‘SLURM@CTM – New version and how to use’, promoted by both during the CTMeet UP – whose success can me measured by the significant participation of CTM researchers and the technical and pedagogical skills of Eduardo and Luís. This nomination stems from their professionalism, quality and responsiveness, and the results of their work, with direct impact on the way the Centre explores the computing resources (with a clear increase in efficiency and parity in the distribution of resources, namely the access to GPUs)”.

– CTM coordinators

The CTM coordinators highlighted your tremendous dedication and team spirit during the implementation and configuration of the SLURM platform. Can you explain the main goal of the platform, and how it works?

SLURM is the new shared computational infrastructure of CTM. This infrastructure is composed of a cluster of GPUs and CPUs that can be used to run complex tasks, such as multimedia processing and Machine Learning. Using an automatic centralised resource management mechanism, SLURM distributes and schedules the execution of tasks according to the resources available at a given time. To ensure an efficient and fair usage of resources by all CTM researchers, SLURM uses priority management mechanisms based on multiple metrics, such as the number of tasks scheduled by users. In addition to the installation and configuration of the platform, we wrote user manuals and an issue/ticket management system, so that researchers can report problems or provide suggestions to improve the platform.

Besides the implementation and configuration of the SLURM platform, you also promoted a training action about its features and usage. What are the main benefits to researchers?

The introduction of SLURM represents a paradigm shift in the way CTM researchers access and use the Centre’s computational resources. Given the high cost of some components, such as GPUs, it is essential to optimise their use, without limiting fairness and equity in their access.

In the previous paradigm, each researcher depended on a limited set of computational resources, whose access was managed manually, and which may not be always utilised. The centralisation of resources on the SLURM platform enables a more efficient and equitable use by all researchers, automatically managing the distribution of scheduled tasks by the available resources and adjusted to their requirements.

As a complement to the documentation and user manuals, this training action allowed us to explain the most important concepts about the new platform, through practical examples, so that CTM researchers could learn how to use the system in their research activities. The interactive and practical aspect of the session also allowed for the clarification of frequently asked questions and the collection of feedback to improve the platform.

Which aspects of your job do you enjoy the most?

As PhD students and researchers in telecommunications and multimedia communications technologies, we feel particularly motivated by the opportunity to install and manage a new shared computing platform to support CTM’s research activities. This opportunity allowed us to acquire and consolidate important skills in systems administration and DevOps, which are highly demanded by the technology industry. Moreover, the possibility of using SLURM for our own research and PhD, makes this project even more attractive and challenging.

How do you comment on this nomination?

We would like to thank the CTM coordinators for acknowledging our work and effort. The implementation of this project entailed multiple technical challenges, considering the complexity of the platform, the objectives outlined by the coordinators and the requirements of the researchers. Most of these challenges derived from the need to create an effective, scalable, and easy-to-maintain solution, whose goals were achieved through an excellent team spirit and knowledge sharing. We feel especially fulfilled with the work developed, translating into a growing use of SLURM by the researchers. Finally, we would like to thank the CTM coordination, in particular Filipe Ribeiro, for the support provided during the implementation of this project.

Carlos Pereira, Carlos Silva, João Pacheco, Ricardo Andrade, Ricardo Silva and Vasco Campos (CPES)

“Carlos Pereira, Carlos Silva, João Paulo Pacheco, Ricardo Andrade, Ricardo Silva and Vasco Campos contributed in a committed and decisive way to the demonstration of several use cases within the scope of the H2020 InterConnect project, addressing different target-audiences, e.g., the European Commission. They were responsible for the high quality and success of the demonstration of semantic interoperability concepts (based on SAREF) on digital platforms that will soon be available to partners like Sonae, Schneider, Sensinov and E-Redes. These will exploit energy flexibility in services and domestic buildings and improve the efficiency of grid operation”.

– CPES coordinators

You’ve been involved in InterConnect, the largest European research project to date, led by INESC TEC. What have been the main challenges to work you’ve been carrying out?

“Challenge” is one of the words that best characterise this project – which, curiously, despite its dimension and complexity, aims to simplify and improve interoperability between digital systems in the electricity sector. However, as you can imagine, in major projects like this one (50 partners from seven EU countries), expectations are always high; the main challenges are keeping calm and fostering hard work and commitment among all partners, in order to address the proposed challenges, with proven quality.

You’ve demonstrated the concepts of semantic interoperability on digital platforms that will soon be available to different partners. What platforms are you talking about and what will be the main benefits to these entities.

In this last demonstration, we presented multiple prototypes that already resort to semantic interoperability to communicate in an intelligent, safe, and interoperable way. Regarding the monitoring of distribution networks, we are developing, together with E-REDES, a prototype for real-time monitoring of faults and their origin in the power grid. In addition, and in order to prevent possible day-ahead problems on the power grid, we have explored the flexibility needs of E-REDES, communicating back, through an aggregator, the respective flexibility of end users. As this aggregator can host flexibility from several “Energy Managers” with different characteristics, these communications are fully interoperable and validated by the SAREF ontology, to ensure the abstraction of the exchanged data. We also have a commercial pilot with two use cases that exploit the energy flexibility on the commercial areas of SONAE MC.

The first is the demonstration of an interoperable IoT system for monitoring, forecasting and controlling the flexibility of consumption of supermarket refrigeration systems, and these systems will play an increasingly active role in responding to the needs of the distribution system operators during power grid management tasks. In addition to the innovative nature of the systems involved, i.e., the consumption flexibility forecasting system developed by INESC TEC, this is an example of an ecosystem based on multi-partner platforms (SONAE MC, Schneider Electric, Sensinov, E-REDES and INESC TEC) that communicate with each other in a simple and uniform way, using the concepts of semantic interoperability explored in this project.

As to the second use case, we are developing a prototype (created by using a combination of machine learning and optimisation techniques) that uses real data from supermarket electric vehicle charging stations to generate incentives capable of promoting vehicle charging during periods with low electricity prices (e.g., surplus photovoltaic generation).

Which aspects of your job do you enjoy the most?

It is a privilege to participate in this type of project, where we implement a set of ambitious ideas, originally conceived at INESC TEC meeting rooms (online or in person), which later turn into real impact and recognition. We all value the enormous challenge of feeling that we are contributing to something great, with immediate applicability, not only concerning the academia and scientific dimensions, but also in terms of commercial and industrial enterprises, to the great benefit of all

How do you comment on this nomination?

We are grateful for this nomination by the CPES coordinators; it acknowledges the quality of this specific work, not only carried out by us, but also by those who have participated in the project, actively contributing to its success and the prominence of INESC TEC at national and international level. We’d also like to thank Fábio Coelho (HASLab) and Rúben Queirós (CTM) for their tireless work, and all those involved in this demonstration phase.