Cloud-Edge-IoT Research area

The Distributed and Parallel Systems (DPS) Group conducts advanced research in the area of Cloud-Edge-IoT computing, addressing the challenges of distributed, heterogeneous, and large-scale environments. Our work focuses on creating novel programming abstractions, efficient runtime systems, and adaptive infrastructures to enable scalable, reliable, and performance-aware computing across the continuum from cloud to edge to IoT devices.

Research Topics

Our research agenda in Cloud-Edge-IoT computing spans the following areas:

New Programming Models and Paradigms

• Development of novel abstractions, such as Collaborative State Machines, to simplify the design and execution of distributed applications in dynamic, heterogeneous environments.

Workflow Applications

• Efficient execution models and scheduling techniques for large-scale, data- and computation-intensive workflows across cloud and edge platforms.

Stream Processing

• Real-time processing frameworks for continuous data streams, supporting low-latency decision-making in IoT and edge scenarios.

Containerization Management and Orchestration 

• Lightweight deployment and orchestration mechanisms for applications spanning cloud-to-edge infrastructures, leveraging container technologies for portability and scalability.

Distributed Resource Management, Autoscaling, and Scheduling 

• Adaptive strategies for resource allocation, autoscaling, and application placement to optimize performance, cost, and energy efficiency.

Replication Protocols and Distributed Storage 

• Reliable data management solutions, including replication mechanisms and scalable storage systems, to ensure consistency, fault-tolerance, and availability in distributed settings.

Distributed Machine Learning and AI Systems

• Large Language Models (LLMs). Exploration of distributed training, fine-tuning, and inference at scale.
• Frameworks for training and deployment across heterogeneous computing resources.
• Federated Learning. Privacy-preserving approaches for collaborative learning over decentralized IoT and edge data sources.
• Multi-Agent and Collaborative Systems Architectures and algorithms enabling intelligent, cooperative, and adaptive distributed systems across diverse devices and infrastructures.

IoT Computing and Networks 

• Scalable frameworks for processing, communication, and management of data generated by large numbers of IoT devices, ensuring reliability, security, and efficiency.

Our goal

Our objective is to design and implement next-generation computing infrastructures that seamlessly integrate cloud, edge, and IoT systems. These infrastructures will enable data- and computation-intensive applications in life sciences, natural sciences, engineering, business analytics, and social domains, providing the foundations for collaborative, performance-oriented, and globally distributed science and applications.

By addressing programmability, scalability, resource efficiency, and reliability across the Cloud-Edge-IoT continuum, DPS research aims to empower scientists and practitioners to conduct research faster, better, and in fundamentally new ways.