Blog

In recent years, the reliance on digital services in critical sectors such as e-commerce, financial services, e-government, and healthcare applications has made availability a requirement as important as performance and security. Recent outages of major cloud platforms and widely used services such as Amazon Web Services (AWS) affected approximately 500 companies due to the disruption—some of which in Brazil included iFood, Mercado Livre, Itaú, and others. Even though they were short-lived, these outages caused global downtime and resulted in millions in losses, highlighting how just a few minutes of disruption can compromise entire business chains and user trust. In this scenario, availability and recovery time metrics have evolved from being merely operational indicators to becoming strategic inputs in capacity planning and the definition of service level agreements (SLAs).

This paper, authored by Jonas César Gomes Nunes (UFPI), Iure de Sousa Fé (UFPI), Lucas Lopes (UFPI), José Miqueias (UFPI), Elias P. Duarte Jr. (UFPR), and Francisco Airton Silva (UFPI), introduces Chaos-K8s, an innovative tool developed to systematically assess availability in Kubernetes clusters.

The solution focuses on automating infrastructure discovery and the empirical calibration of recovery times, allowing administrators to test failures in the control plane, nodes, and pods (a pod is a small part of an application containing everything it needs to function). By performing controlled simulations, the system helps quantify how different configurations impact the end-user experience and service resilience.

The study validates the tool’s effectiveness through mathematical modeling, demonstrating a high degree of statistical accuracy between practical tests and theoretical results. Ultimately, the work offers a structured methodology for companies to predict outages and optimize the robustness of their cloud operations.

Automatic Discovery and Calibration

The tool automatically maps the cluster components (control plane components such as etcd and kube-apiserver, worker nodes, and deployed applications) to generate an initial configuration without manual effort. Unlike other tools that use generic values, Chaos-K8s can perform a calibration phase where it injects controlled failures to measure the actual Mean Time to Recovery (MTTR) of each component in the specific environment. This ensures that availability calculations reflect the reality of the evaluated infrastructure.

The system acts as an external orchestrator that interacts with the cluster to apply failures at different layers: to Pods—terminating processes to force a restart by Kubernetes; Worker Nodes—restarting critical services such as kubelet and containerd via SSH; and the Control Plane—interrupting core processes such as etcd and kube-apiserver.

This week, we were thrilled to receive the news that ICoNIoT’s proposal for the São Paulo School of Advanced Science (ESPCA), titled “São Paulo School of Advanced Science in Artificial Intelligence for Communication Networks,” has been approved. The project is coordinated by researcher Carlos Kamienski (UFABC), coordinator of Education and Outreach at INCT ICoNIoT. The courses are scheduled to take place between September 13 and 22, 2027.

The review report approving the School highlights the quality of the proposed content, which is multidisciplinary and spans the fields of computer networks, telecommunications, and AI. The reviewers stated that the project has great potential to contribute to increasing the visibility of research and to encourage participation in doctoral programs and postdoctoral fellowships in the State of São Paulo, especially for students from other countries and Brazilian states. They also emphasized that the proposal offers a framework capable of providing significant benefits to participants, students, and researchers in the state of São Paulo. The topics to be addressed by the speakers were deemed current and relevant.

The review highlighted that the scientists responsible for the courses (speakers/guest lecturers) are internationally renowned in their fields, and that the proposed group of scientists (speakers/guest lecturers) reflects a good balance between foreign and Brazilian scholars.

Student Participation

Each ESPCA is expected to host up to approximately 100 students, and priority will be given to proposals that demonstrate significant participation by students from outside Brazil. At least 50% of participants are expected to come from other countries.

Participating students will have the opportunity to present the results of their ongoing research in poster sessions included in the program.

Dissemination

It was emphasized that the proposal offers new ways to disseminate information and ideas in a manner that could not be achieved through traditional communication channels such as scientific publications and presentations at scientific events. In this regard, the focus is on the group activities and discussions, the Student Lightning Talks, the field trips, and the roundtable discussions with industry representatives, as outlined in the project. These include guided field trips to major centers, research institutions, academic institutions, and companies relevant to the topics covered. The following field trips are planned: 1) Telecom Operator; 2) Samsung / Motorola R&D Center; 3) Smart City / Urban Operations Center (COP-SP).

The report also highlighted that in-person discussions and hands-on activities allow for interaction and the exchange of information and ideas in ways that are not possible through usual communication channels, such as scientific publications and presentations at scientific events.

How ESPCA Works

FAPESP’s São Paulo School of Advanced Science (ESPCA) Program aims to support the organization of short-term courses that contribute to the advanced training of students and young researchers, promoting interaction with researchers of recognized excellence and expanding the international reach of research conducted in the state of São Paulo.

The ESPCA program has already funded dozens of São Paulo Schools of Advanced Science, which have brought together thousands of students and researchers from various countries. ESPCA provides funding for the organization of short-term courses in advanced research across different fields of knowledge in the state of São Paulo.

This talk will present the evolution of horizontal scaling strategies in Kubernetes environments, addressing intelligent mechanisms based on artificial intelligence techniques—including fuzzy logic, machine learning, and deep learning—to optimize resource allocation, predict computational workloads, and dynamically adapt to the demands of distributed systems. Next, we will discuss the specific challenges of running distributed control systems on edge infrastructures, highlighting the limitations of traditional scaling strategies based exclusively on CPU utilization. The presentation will delve into the concept of control-aware sizing, in which control quality (Quality of Control, QoC) metrics, temporal constraints, and computational costs are considered jointly to determine the appropriate number of shared controllers.

Experimental results will be presented, and prospects for the development of resource management mechanisms capable of integrating scalability, computational efficiency, and control requirements in IIoT systems will be discussed.

The speaker

The speaker, Marcelo Augusto Costa Fernandes, is an associate professor in the Department of Computer Engineering and Automation at the Federal University of Rio Grande do Norte (UFRN). He holds a Ph.D. in Electrical Engineering from the State University of Campinas (UNICAMP) and has served as a visiting researcher at King’s College London (United Kingdom), Harvard University (United States), and the University of Granada (Spain). He is currently the director of the Center for Artificial Intelligence and Data Science (nIACD) at the Digital Metropolis Institute (IMD/UFRN) and coordinator of the InovAI Lab. His research focuses on artificial intelligence, deep learning, digital signal processing, embedded systems, reconfigurable computing, and bioinformatics. He is an associate researcher at INCT ICoNIoT and participates in national and international projects focused on the development of intelligent systems and AI technologies.

The paper by Lucas Airam C. de Souza (UFRJ), Nadjib Achir (INRIA Saclay, France), Miguel Elias Mitre Campista (UFRJ), and Luis Henrique Maciel Kosmalski Costa (UFRJ) was presented at SBRC 2026. The research was also featured in a mini-course at the event, which included a hands-on activity using a simulation of the mobility studied by the researchers.

The research addresses one of the challenges of Federated Vehicle Learning: the high mobility of users. As vehicles move, communication conditions can vary significantly, affecting the time required for local training and for transmitting updates to the server. When an update exceeds the time limit set for a training round, it may end up being discarded.

To address this problem, the researchers propose CAIROS, a strategy that uses a Long Short-Term Memory (LSTM) model to dynamically estimate the network and computing conditions of each client.
The proposal is to use information available during training—such as the maximum time limit set for the round, the time already elapsed, the time remaining, the network conditions experienced by the client, and the amount of available data—to estimate the time needed to complete processing and transmit the update. In this way, the system aims to prevent the client’s work from being lost due to exceeding the time limit.

In CAIROS, the LSTM model acts as a dynamic delay estimator. Its function is to predict the communication conditions for each vehicle client and estimate whether the time required to process and transmit the local update might exceed the limit set for the round.
Based on this prediction, the vehicle can decide whether to continue local training or send the parameters to the server early. The strategy thus aims to reduce the discarding of updates and make better use of the work performed by the clients.
In the experiments presented by the researchers, CAIROS increased training efficiency and achieved gains of up to 25% in the accuracy of the global model in scenarios with severe time constraints, compared to the traditional FedAvg algorithm.

Prof. Nelson Fonseca presented ICoNIoT in Brasília at the event organized by CNPq

The INCT 2026 Seminar was held in Brasília from September 22 to 24. The event brought together coordinators and representatives from the currently active National Institutes of Science and Technology, as well as officials from CNPq, the Ministry of Science, Technology, and Innovation (MCTI), Capes, the National Council of State Research Support Foundations (Confap), and other partner institutions.

O coordenador geral do ICoNIoT, prof. Nelson Fonseca, esteve presente realizando apresentação sobre o INCT ICoNIoT. Ele falou para o público do seminário sobre a estrutura do ICoNIoT; as produções realizadas até o momento; os pesquisadores formados; a interação do INCT com o setor produtivo – cooperações e parcerias com empresas privadas, órgãos públicos, secretarias; as iniciativas de internacionalização do INCT ICoNIoT – que incluem universidades presentes em quase 20 países, e as contribuições do INCT à inovação em políticas públicas. 

Segundo o prof. Nelson, “o workshop organizado pelo CNPq teve como participantes os coordenadores dos INCT vigentes, os gestores e os funcionários do programa. O INCT ICoNIoT foi selecionado para apresentação e recebeu um retorno muito positivo pela sua atuação e excelência na sua área de atuação”. 

Impacts and Contributions of ICoNIoT

The scientific and technological impacts of INCT ICoNIoT were also highlighted; the institute has been pushing the frontiers of knowledge in areas such as federated learning, 5G/6G and OpenRAN networks, cybersecurity, edge, fog, and cloud computing, IoT and vehicular systems, optical networks, and distributed AI. The Institute has contributed to the development of new methodologies, analytical models, experimental platforms, databases, and computational tools focused on network traffic analysis, anomaly detection, DDoS attack prediction, protection of IoT and IIoT devices, privacy preservation, and monitoring of distributed systems. These results have served as the foundation for new research projects, the training of highly qualified professionals, and the transfer of knowledge to strategic sectors of the digital economy.

The scientific outreach initiatives aimed at bridging the gap between science, technology, education, society, and the productive sector were highlighted. The strategy aims to help disseminate the work carried out at the INCT through communication in languages appropriate for both internal and external audiences; foster a culture of collaboration among researchers to ensure that the knowledge produced leads to new partnerships; support a strategy for training professionals in the INCT’s areas of focus; and help enhance other aspects of science and technology management at the INCT, such as technology transfer.

Check out some photos from the seminar:

This work was conducted by researchers João Victor Feio Gonçalves (UFPA), John Sousa (UFPA), Rafael Veiga (UFPA), Lucas de Lima Bastos (UNIFESSPA), Lucas de Sousa Pacheco (UFPA), Iago Medeiros (UFPA), Denis Rosário (UFPA), and Eduardo Coelho Cerqueira (UFPA).

They developed ASTRA, a new method designed to optimize Federated Learning in scenarios where data across devices is highly diverse and disconnected. The core of the innovation lies in Dual-Space Regularization, which combines physical constraints on model weights with semantic self-distillation to ensure that local predictions remain aligned with the global objective.

To enable use on resource-constrained devices, the authors implement Curriculum-Based Scaling, activating the “instructor” model intensively only at the beginning and subsequently on a periodic basis. The purpose of the study is to mitigate client drift and prevent catastrophic system divergence, achieving higher accuracy than traditional methods with only a small increase in computational cost.

The project developed by Amanda Lopes (UFPA), John Sousa (UFPA), Lucas de Lima Bastos (UNIFESSPA), Lucas de Sousa Pacheco (UFPA), Iago Medeiros (UFPA), Denis Rosário (UFPA), and Eduardo Coelho Cerqueira (UFPA) was presented at SBRC 2026.

Vehicles are increasingly capable of collecting large amounts of data through sensors and IoT devices. In the context of autonomous driving, one approach is to equip cars with multiple sensors and sufficient computing power to process this information and estimate, in real time, the best route to take.

With this processing, the vehicle itself can make decisions, such as braking, accelerating, or changing its course. However, the potential is even greater when different vehicles can participate in the same network and share information. Based on the data and predictions generated by these vehicles, it is possible, for example, to identify the best routes and anticipate traffic situations.

It is in this context that connected autonomous vehicles have emerged: an attempt to expand the capabilities of vehicle systems to solve problems that the hardware available in a single vehicle cannot solve on its own. The idea is to leverage the knowledge generated by academia and create solutions that can, in the future, be implemented by the industry.

One of the challenges is enabling the vehicle to “read” its surrounding context and use that information to make more appropriate decisions. This is where PRINCE comes in—a federated learning (FL) architecture designed for autonomous vehicle networks.

The proposal aims to select, from among the available vehicles, those best suited to participate in a federated learning process. Instead of considering all vehicles in the network indiscriminately, the system evaluates different characteristics to determine which ones can contribute most significantly to the learning process.

Proactive customer selection, as exemplified by PRINCE, offers significant advantages over traditional reactive methods in Connected and Autonomous Vehicles (CAVs) environments, particularly in terms of resource efficiency, convergence speed, and model stability. While reactive approaches select vehicles and address any failures only after they occur, the proactive strategy seeks to anticipate potential problems and select, from the outset, the clients most likely to complete training.

One of the main benefits lies in the efficient use of resources. In reactive methods, bandwidth and processing capacity can be wasted when selected vehicles drop out or fail during a training round. The proactive approach seeks to pre-filter unstable nodes, achieving—according to the presented results—an improvement of up to 3.5 times in resource efficiency. This proactive approach also contributes to a significant reduction in connection failures—up to 2.2 to 2.5 times fewer compared to reactive approaches—thanks to the use of stochastic models to predict instabilities and pre-filtering based on signals such as RSSI.

These gains are also reflected in the model’s performance. PRINCE achieved a final accuracy of 83.90%, outperforming state-of-the-art reactive models by up to 9.22%. Furthermore, proactive selection accelerates convergence: in one of the evaluated scenarios, the system reached 60% accuracy three times faster than random selection methods, while also outperforming reactive approaches such as ACSP-FL and ECS-HDSR.

Another key advantage lies in reduced latency and computational costs. In reactive strategies, such as ECS-HDSR, when a client fails, the system must find a replacement and restart or reorganize part of the process, which adds steps, time, and processing load. The proactive approach seeks to avoid this situation by pre-selecting a set of clients with a high probability of remaining connected and completing the training.

This proactive approach also promotes training stability. While reactive methods may exhibit fluctuations in accuracy and spikes in the loss function due to the inclusion of models from failed clients, proactive selection tends to produce a more stable and consistent learning trajectory.

In summary, the fundamental difference lies in how each strategy handles uncertainty. Reactive selection waits for a problem to occur before attempting to correct it; proactive selection uses vehicle behavior history, network conditions, and mobility forecasts to anticipate potential failures and prevent them. In a dynamic environment such as that of connected and autonomous vehicles, this predictive capability can make each communication cycle more efficient, reduce waste, and contribute to global models that converge more quickly and stably.

Factors Considered in the Active Selection of Vehicles

The first factor considered is the location and connectivity of the vehicles, which are determined by the Received Signal Strength Indicator (RSSI). Vehicles that are closer or farther away—and whether or not they have internet coverage—are evaluated. Vehicles with weak RSSI values (i.e., values below -85 dbM) are excluded from further selection, as they do not have sufficient connectivity in that area to actively participate in the federated learning client selection process. Vehicles with good RSSI values (i.e., values above -85 dbM) are selected for the next stage.

To estimate the likelihood that a vehicle will remain connected during the process, in the second step the architecture uses the concept of a Markov chain—a stochastic process in which the system can be in a given state and either remain in that state or transition to another. Based on this model, the probability that a given vehicle will remain connected to the network is calculated.

This information is important because, in a federated learning process, it is not enough for a vehicle to have relevant data: it must also be able to remain connected for the time necessary to participate in the training.

The Importance of Data Heterogeneity

The third and final criterion used by PRINCE is the diversity of the data available in each vehicle. To this end, the architecture takes into account Shannon entropy, a measure related to the variability and diversity of information.

The reasoning is that vehicles with more diverse data can make a more meaningful contribution to the learning process, since they allow the model to encounter different situations and patterns.

Thus, the system seeks to prioritize vehicles with a higher entropy index—that is, greater diversity in their data. The architecture therefore combines different criteria to select the most suitable vehicles to participate in federated learning. Generally speaking, three dimensions are considered: proximity—prioritizing vehicles that are close to the region or situation of interest; connectivity—selecting vehicles most likely to remain connected during the process; and data heterogeneity—prioritizing vehicles whose data exhibit greater diversity.

The goal is to find a set of vehicles that is not only related to the problem we want to analyze but is also capable of effectively contributing to the model’s training.

Possible Applications

One possible application would be the use of PRINCE by a company that needs to transmit specific information to certain vehicles in a given region.

Imagine, for example, that a problem is occurring on a particular route. The company could use the architecture to identify vehicles that are in that area or heading toward it and share relevant information with those users.

Similarly, a driver traveling to a specific city could receive specific information related to conditions in that region, allowing the system to use the data generated by connected vehicles to support more informed decisions.

Read the full article in the SBRC 2026 proceedings 

The paper, authored by Rodrigo Almeida Bezerra (UFBA), Luiz Fernando Bittencourt (UNICAMP), Leandro Villas (UNICAMP), and Maycon Peixoto (UFBA), is set within the intersection of the fields of IoT and Brain-Computer Interfaces (BCI).

The study simulates the operation of a BCI. In this specific case, the BCI involves a game that connects via electroencephalography—a device that, when placed on the head, captures brain signals to establish a connection between the user’s brain and the game.

In this context, potential delays can negatively impact the user experience. Edge computing comes into play, then, to enable a fast, low-latency user experience. To achieve this, processing must take place at the edge; that is why the shift from the cloud to the edge is important for specific applications like this one.

In the case of this work, there is an additional dimension: user mobility. The user is moving from one place to another, and the application modules need to move along with the user. User mobility, therefore, complicates the interactions between the system’s elements; the edge could be, for example, a device in the user’s home, and when the user moves away, latency increases.

It is in this context that the MX algorithm solution comes into play, to track the user.

What is the MI-X algorithm, and what is its main objective?

This algorithm, developed by the researchers, is a strategy for migrating game applications from the fog to the cloud.

Mobility Impact-X (MI-X) is a dynamic allocation algorithm and a mobility-aware modular allocation strategy designed to operate in Hierarchical Fog Computing architectures. It was specifically designed to manage modular, latency-sensitive Brain-Computer Interface (BCI) applications, such as electroencephalogram-based games (EEG Games), within the cloud-fog-edge continuum.

Its primary goal is to minimize application latency and response times while simultaneously optimizing the use of computational resources. To achieve this, MI-X seeks to dynamically balance the following factors:

Computational proximity to the user: Prioritizing execution at the edge (cloudlets) to reduce perceived latency.

Communication cost: Assessing the impact of distributing or migrating interdependent modules across the network.

Operational stability: Avoiding abrupt migrations and long hierarchical hops that could cause system instability.

What sets MI-X apart is its ability to integrate the application’s modular structure with mobility management. This allows the algorithm to keep interdependent components physically close to one another within the infrastructure, preventing network costs from negating individual processing gains and ensuring consistent performance even amid continuous user mobility.

Challenges

Edge applications are close to the user but have limited resources; cloud applications, on the other hand, have more resources but are farther from the user, resulting in longer response times.

In urban mobility scenarios, edge resources can become saturated. The developed solution analyzes the impact of each option: it performs a search and attempts to balance latency, computational capacity, communication costs, etc.

An interesting fact about the development of this solution is that the history of this algorithm includes others, built in 2019, 2020, and 2023, which have evolved over time; MI-X is the first in the series to offer the possibility of reducing latency.

The researchers emphasize that it is not enough to simply place everything at the edge; processing must intelligently follow the user, tracking their movements.

Work continues to evolve through the research of master’s student Rodrigo Bezerra (UFBA)—now with a new strategy called HERMES (Heuristic Resource Management for Edge Systems). This approach builds on MI-X by considering not only where application modules should be executed as the user moves, but also when a migration is actually worthwhile. HERMES analyzes the impact of each possible reallocation, taking into account latency, resource availability, and the cost of transferring modules, thereby avoiding unnecessary migrations and the so-called “back-and-forth” effect between the edge, fog, and cloud. Thus, the research evolves from the idea of simply following the user to that of intelligently and stably orchestrating the entire path between the edge, the fog, and the cloud.

Read the paper