Abstract

TBA

About the speaker

TBA

Abstract

The commercial success of edge computing hinges on the ability to allocate and price computational resources efficiently while meeting stringent service latency requirements. Many edge applications are expected to contain machine learning (ML) components, yet our understanding of the interference between co-located ML workloads, often referred to as the noisy neighbors problem, and its impact on how to orchestrate applications under latency constraints is rather limited. In this talk we address this problem from two complementary perspectives, resource management and pricing. We first discuss the problem of profiling the inference time of ML applications for energy aware service orchestration. We introduce a lightweight, data-driven framework, and show how it can be used for scalable placement and resource management decisions, balancing latency guarantees with energy efficiency. We then turn to the problem of pricing in multi-tenant edge environments, formulated as a game among network operators, service providers, and autonomous, price sensitive users, where contention among workloads potentially affects performance. We show how a model-assisted learning approach can be used by the operator for learning a near-optimal pricing policy under resource and latency constraints. We posit that combining data driven performance modeling, optimization, and learning, as discussed in this talk, could enable more sustainable, economically viable edge systems that adapt intelligently to contention, latency, and energy constraints.

About the speaker

György Dán is professor of teletraffic systems at KTH Royal Institute of Technology, Stockholm, Sweden. He received the M.Sc. in computer engineering from the Budapest University of Technology and Economics, Hungary in 1999, the M.Sc. in business administration from the Corvinus University of Budapest, Hungary in 2003, and the PhD in Telecommunications from KTH in 2006. He worked as a consultant in the field of access networks, streaming media and videoconferencing 1999-2001. He was a visiting researcher at the Swedish Institute of Computer Science in 2008, a Fulbright research scholar at University of Illinois at Urbana-Champaign in 2012-2013, and an invited professor at EPFL in 2014-2015. He served as area editor of Computer Communications 2014-2021, as editor of IEEE Transactions on Mobile Computing 2019-2023, serves on the TPC of conferences like IEEE Infocom, ACM e-Energy, IJCAI, AAAI, and is vice-chair of the steering committee of IEEE SmartGridComm. He has received several best paper awards from IFIP and IEEE in recent years. His research interests include the design and analysis of mobile computing systems, game theoretical models of networked systems, and cyber-physical system security and resilience.