TECoSA Seminar – Computation offloading in Edge and cloud environment: Survey, taxonomy, applications and open challenges

We aim to bring you a TECoSA Seminar on the first Thursday of each month during term-time. For Spring 2023, the talks will be on-line or hybrid. All are welcome to attend and we look forward to some lively discussions. Members can accept the invitations, non-members can email tecosa-admin@kth.se to register.

Our June seminar is with Ali Balador, senior researcher and project manager at Ericsson Research.

The Zoom link is https://kth-se.zoom.us/j/66857695267.

ABSTRACT:  With the rapid development in the area of Internet of Things (IoT), the number of delay-sensitive and computationally intensive IoT applications has drastically increased recently. The resource-constrained nature of the IoT devices, different characteristics of IoT tasks such as delay sensitivity, and heterogeneity of the edge cloud servers make the task offloading problem an important challenge in the edge cloud continuum. Offloading can be defined as a mechanism to move the processing or computation from one device to another with more suitable capabilities. The main goal is to move resource-intensive tasks from a device with limited resources to an edge server or cloud server to achieve performance objectives, such as reducing overall computation time, minimizing network resources usage, maximizing battery life.

Offloading is a flexible mechanism whose implementation depends on use cases or situations. For this reason, it is very important to study different application domains and use cases to find out which use cases are the most promising and can get benefit of offloading. In this presentation, I will explore different application domains and use cases and explain why offloading is beneficial. I will also show an offloading taxonomy that helps to understand that flexibility and identify which are the most relevant offloading configurations for different application domains and use cases.

One example could be offloading to support time-critical industrial use cases. However, to ensure deterministic performance and support timing predictability of real-time applications that utilize the edge-cloud continuum, more enhancements are required in the computation domain and communication domain. In this case, I will present our initial plans for our new VINNOVA project, which is called AORTA. We will develop technologies that allow offloading of real-time services and functionality to the edge and cloud. We will build upon recent advances in 5G and cloud and networking technologies, as well as cutting-edge research results. The AORTA framework will support a fluid compute model where functionality will be dynamically deployed locally, at the edge, or in the cloud and support integration and real-time performance irrespective of where they execute.

BIO:  Dr. Ali Balador graduated in computer science at the Polytechnic University of Valencia, where he also achieved his Phd on the topic of wireless communication for vehicular environments in 2016. He held the role of Assistant Professor at Mälardalen University for 5 years. He was a visiting researcher at University of Bologna in Italy, Halmstad University in Sweden, and the National Institute of Informatics (NII) in Japan.

Ali has a background in a number of industries, such as automotive, railway, healthcare, and manufacturing but has spent most of his time focusing on the automotive sector, working in research and development projects where he acted as main technical coordinator or WP leader. He spent 6 years at RISE Research Institute of Sweden where, he focused on topics such as edge and cloud computing and the combination with artificial intelligence.

In April 2022, he joined Ericsson Research in Sweden in the role of senior researcher and project manager. His own personal research interests include wireless networks, distributed systems, cloud and edge computing, and security & privacy.