,

Accepted tutorial: Serverless Orchestration on the Edge-Cloud Continuum: From Small Functions to Large Language Models

We are happy to announce that our tutorial “Serverless Orchestration on the Edge-Cloud Continuum: From Small Functions to Large Language Models” (by Reza Farahani and Radu Prodan) has been accepted for IEEE ICDCS 2025, which will take place in Glasgow, Scotland, UK, in July 2025.

Venue: 45th IEEE International Conference on Distributed Computing Systems (ICDCS) (https://icdcs2025.icdcs.org/)

Abstract: Serverless computing simplifies application development by abstracting infrastructure management, allowing developers to focus on functionality while cloud providers handle resource provisioning and scaling. However, orchestrating serverless workloads across the edge-cloud continuum presents challenges, from managing heterogeneous resources to ensuring low-latency execution and maintaining fault tolerance and scalability. These challenges intensify when scaling from lightweight functions to compute-intensive tasks such as large language model (LLM) inferences in distributed environments. This tutorial explores serverless computing’s evolution from small functions to large-scale AI workloads. It introduces foundational concepts like Function-as-a-Service (FaaS) and Backend-as-a-Service (BaaS) before covering advanced edge-cloud orchestration strategies. Topics include dynamic workload distribution, multi-objective scheduling, energy-efficient orchestration, and deploying functions with diverse computational requirments. Hands-on demonstrations with Kubernetes, GCP Functions, AWS Lambda, OpenFaaS, OpenWhisk, and monitoring tools provide participants with practical insights into optimizing performance and energy efficiency in serverless orchestration across distributed infrastructures.

Hadi
,

Tutorial accepted: Perceptually Inspired Visual Quality Assessment in Multimedia Communication

ACM MM’25 Tutorial: Perceptually Inspired Visual Quality Assessment in Multimedia Communication

ACM MM 2025, October 27, 2025, Dublin, Ireland

https://acmmm2025.org/tutorial/

Tutorial speakers:

  • Wei Zhou (Cardiff University)
  • Hadi Amirpour (University of Klagenfurt)

Tutorial description:

As multimedia services like video streaming, video conferencing, virtual reality (VR), and online gaming continue to expand, ensuring high perceptual quality becomes a priority for maintaining user satisfaction and competitiveness. However, during acquisition, compression, transmission, and storage, multimedia content undergoes various distortions, causing degradation in experienced quality. Thus, perceptual quality assessment, which focuses on evaluating the quality of multimedia content based on human perception, is essential for optimizing user experiences in advanced communication systems. Several challenges are involved in the quality assessment process, including diverse characteristics of multimedia content such as image, video, VR, point cloud, mesh, multimodality, etc., and complex distortion scenarios as well as viewing conditions. The tutorial first presents a detailed overview of principles and methods for perceptually inspired visual quality assessment. This includes both subjective methods, where users directly rate their experience, and objective methods, where algorithms predict human perception based on measurable factors such as bitrate, frame rate, and compression levels. Based on the basics of perceptually inspired visual quality assessment, metrics for different multimedia data are then introduced. Apart from the traditional image and video, immersive multimedia and AI-generated content will also be involved.

Hadi
,

Journal paper accepted: ACM TOMM: Convex Hull Prediction Methods for Bitrate Ladder Construction: Design, Evaluation, and Comparison

URL: https://dl.acm.org/journal/tomm

Authors: Ahmed Telili (INSA, Rennes, France),  Wassim Hamidouce (INSA, Rennes, France), Hadi Amirpour (Alpen-Adria-Universität Klagenfurt), Sid Ahmed Fezza (INPTIC, Algeira), Christian Timmerer (Alpen-Adria-Universität Klagenfurt), and Luce Morin (INSA, Rennes, France)

Abstract:
HTTP adaptive streaming (HAS ) has emerged as a prevalent approach for over-the-top (OTT ) video streaming services due to its ability to deliver a seamless user experience. A fundamental component of HAS is the bitrate ladder, which comprises a set of encoding parameters (e.g., bitrate-resolution pairs) used to encode the source video into multiple representations. This adaptive bitrate ladder enables the client’s video player to dynamically adjust the quality of the video stream in real-time based on fluctuations in network conditions, ensuring uninterrupted playback by selecting the most suitable representation for the available bandwidth. The most straightforward approach involves using a fixed bitrate ladder for all videos, consisting of pre-determined bitrate-resolution pairs known as one-size-fits-all. Conversely, the most reliable technique relies on intensively encoding all resolutions over a wide range of bitrates to build the convex hull, thereby optimizing the bitrate ladder by selecting the representations from the convex hull for each specific video. Several techniques have been proposed to predict content-based ladders without performing a costly, exhaustive search encoding. This paper provides a comprehensive review of various convex hull prediction methods, including both conventional and learning-based approaches. Furthermore, we conduct a benchmark study of several handcrafted- and deep learning ( DL )-based approaches for predicting content-optimized convex hulls across multiple codec settings. The considered methods are evaluated on our proposed large-scale dataset, which includes 300 UHD video shots encoded with software and hardware encoders using three state-of-the-art video standards, including AVC /H.264, HEVC /H.265, and VVC /H.266, at various bitrate points. Our analysis provides valuable insights and establishes baseline performance for future research in this field.
Dataset URL: https://nasext-vaader.insa-rennes.fr/ietr-vaader/datasets/br_ladder

,

Paper accepted: End-to-End Learning-based Video Streaming Enhancement Pipeline: A Generative AI Approach

Authors: Emanuele Artioli (Alpen-Adria Universität Klagenfurt, Austria), Farzad Tashtarian (Alpen-Adria Universität Klagenfurt, Austria), Christian Timmerer (Alpen-Adria Universität Klagenfurt, Austria)

Venue: ACM 35th Workshop on Network and Operating System Support for Digital Audio and Video (NOSSDAV’25)

Abstract: The primary challenge of video streaming is to balance high video quality with smooth playback. Traditional codecs are well tuned for this trade-off, yet their inability to use context means they must encode the entire video data and transmit it to the client.
This paper introduces ELVIS (\textbf{E}nd-to-end \textbf{L}earning-based \textbf{VI}deo \textbf{S}treaming Enhancement Pipeline), an end-to-end architecture that combines server-side encoding optimizations with client-side generative in-painting to remove and reconstruct redundant video data. Its modular design allows ELVIS to integrate different codecs, in-painting models, and quality metrics, making it adaptable to future innovations.
Our results show that current technologies achieve improvements of up to 11 VMAF points over baseline benchmarks, though challenges remain for real-time applications due to computational demands. ELVIS represents a foundational step toward incorporating generative AI into video streaming pipelines, enabling higher quality experiences without increased bandwidth requirements.
By leveraging generative AI, we aim to develop a client-side tool, to incorporate in a dedicated video streaming player, that combines the accessibility of multilingual dubbing with the authenticity of the original speaker’s performance, effectively allowing a single actor to deliver their voice in any language. To the best of our knowledge, no current streaming system can capture the speaker’s unique voice or emotional tone.

 

,

Journal article accepted: Energy-Time Modeling of Distributed Multi-Population Genetic Algorithms with Dynamic Workload in HPC Clusters

We are glad that the paper was accepted for publication in Future Generation Computer Systems. This journal publishes cutting-edge research on high-performance computing, distributed systems, and advanced computing technologies for future computing environments.

Authors: Juan José Escobar, Pablo Sánchez-Cuevas, Beatriz Prieto, Rukiye Savran Kızıltepe, Fernando Díaz-del-Río, Dragi Kimovski

Abstract: Time and energy efficiency is a highly relevant objective in high-performance computing systems, with high costs for executing the tasks. Among these tasks, evolutionary algorithms are of consideration due to their inherent parallel scalability and usually costly fitness evaluation functions. In this respect, several scheduling strategies for workload balancing in heterogeneous systems have been proposed in the literature, with runtime and energy consumption reduction as their goals. Our hypothesis is that a dynamic workload distribution can be fitted with greater precision using metaheuristics, such as genetic algorithms, instead of linear regression. Therefore, this paper proposes a new mathematical model to predict the energy-time behaviour of applications based on multi-population genetic algorithms, which dynamically distributes the evaluation of individuals among the CPU-GPU devices of heterogeneous clusters. An accurate predictor would save time and energy by selecting the best resource set before running such applications. The estimation of the workload distributed to each device has been carried out by simulation, while the model parameters have been fitted in a two-phase run using another genetic algorithm and the experimental energy-time values of the target application as input. When the new model is analysed and compared with another based on linear regression, the one proposed in this work significantly improves the baseline approach, showing normalised prediction errors of 0.081 for runtime and 0.091 for energy consumption, compared to 0.213 and 0.256 shown in the baseline approach.

,

Journal article accepted: Overview of Research Trends and Challenges in 6G Mobile Networks and the Computing Continuum

We are glad that the paper was accepted for publication in SCSA Journal. The journal covers research in smart computing systems and applications, with a focus on next-generation networking, cloud, and edge computing solutions.

Authors: Stojan Kitanov, Dragi Kimovski, Fisnik Doko, Kurt Horvath, Shpresa Tuda, Blerta Idrizi

Abstract: The rapid proliferation of IoT devices, coupled with the generated exponential growth of data, has necessitated the development of advanced network architectures. As a result, 5G mobile networks have already begun to face challenges such as network congestion, latency, and scalability limitations. Therefore, the need for a robust and future-proof solution becomes increasingly evident. In this direction, many research initiatives and industrial communities started to work on the development of 6G mobile networks. On the other hand, the emerging concept of Computing Continuum encompasses the seamless integration of edge, fog, and cloud computing resources to provide a unified and distributed computing environment, and it aims to enable real-time data processing, low-latency response, and intelligent decision-making at the network edge. The primary objective of this research paper is to address the shortcomings of existing network infrastructures and to overcome these shortcomings by integrating advanced AI capabilities in 6G mobile networks with the Computing Continuum. Moreover, it would be proposed a Computing Continuum Middleware for Artificial Intelligence over 6G networks, which would offer high-level and well-defined (“standardized”) interfaces which would create an automated, sustainable loop for managing IoT applications utilizing AI approaches over 6G networks.

,

Paper accepted: ADApt: Edge Device Anomaly Detection and Microservice Replica Prediction

We are glad that the paper was accepted for publication at ICFEC 2025. ICFEC focuses on innovations in cloud and edge computing, bringing together researchers and practitioners to discuss emerging challenges and solutions.

Title: ADApt: Edge Device Anomaly Detection and Microservice Replica Prediction

Authors: Narges Mehran, Nikolay Nikolov, Radu Prodan, Dumitru Roman, Dragi Kimovski, Frank Pallas, Peter Dorfinger

Venue: 9th IEEE International Conference on Fog and Edge Computing 2025, in conjunction with CCGrid 2025, 19-22 May, 2025 – Tromso, Norway

Abstract: The recent shift towards increasing user microservices in the Edge computing infrastructure brings new orchestration challenges, such as detecting overutilized resources and scaling out overloaded microservices in response to augmenting requests. In this work, we present the ADApt using the monitoring data related to Edge computing resources, detecting the utilization-based anomalies of resources (e.g., CORE or MEM), investigating the scalability in microservices, and adapting the application executions. To reduce the bottleneck while using computing resources, we first explore monitored devices executing microservices with various requirements, detecting overutilization-based processing events, and scoring them. Thereafter, based on the memory requirements, ADApt predicts the processing requirements of the microservices and estimates the number of replicas running on the overutilized devices. The prediction results show that the gradient boosting regression-based replica prediction reduces the MAE, MAPE, and RMSE compared to other models. Moreover, ADApt is able to estimate the number of replicas for each microservice close to the actual data without any prediction and to reduce the utilization of the device.

,

Workshop paper accepted: Serverless Large Language Models: Edge vs. Cloud Deployment Trade-offs

Authors: Reza Farahani, Radu Prodan

Venue: Workshop on AI for Sustainable Energy Systems and Green AI, Brdo Estate, Slovenia, 11-12 March 2025

 

 

Hadi
,

Paper accepted at ACM TOMM: Counterfeiting Attacks on a RDH-EI scheme based on block-permutation and Co-XOR

Authors: Fan Chen (Southwest Jiaotong University, China),  Lingfeng Qu (Guangzhou University, China), Hadi Amirpour (Alpen-Adria-Universität Klagenfurt), Christian Timmerer (Alpen-Adria-Universität Klagenfurt), and Hongjie He (Southwest Jiaotong University, China)

Journal: ACM Transactions on Multimedia Computing Communications and Applications (ACM TOMM)

Abstract: Reversible data hiding in encrypted images (RDH-EI) has gained widespread attention due to its potential applications in secure cloud storage. However, the security challenges of RDH-EI in cloud storage scenarios remain largely unexplored.} In this paper, we present a counterfeiting attack on RDH-EI schemes that utilize block-permutation and Co-XOR (BPCX) encryption. We demonstrate that ciphertext images generated by BPCX-based RDH-EI are easily tampered with to produce a counterfeit decrypted image with different contents imperceptible to the human eye. This vulnerability is mainly because the block permutation key information of BPCX is susceptible to known-plaintext attacks (KPAs). Taking ciphertext images in telemedicine scenarios as an example, we describe two potential counterfeiting attacks, namely fixed-area and optimal-area attacks. We show that the quality of forged decrypted images depends on the accuracy of the estimated block-permutation key under KPA conditions. To improve the invisibility of counterfeit decrypted images, we analyze the limitations of existing KPA methods against BPCX encryption for 2×2 block sizes and propose a novel diagonal inversion rule specifically designed for image blocks. This rule further enhances the accuracy of the estimated block-permutation key. The experiments show that, compared to existing KPA methods, the accuracy of the estimated block-permutation key in the UCID dataset increases by an average of 11.5%. In the counterfeiting attack experiments on Camera’s encrypted image, we successfully tampered with over 80% of the pixels in the target area under the fixed-region attack. Additionally, we achieved a tampering success rate exceeding 90% in the optimal-region attack.

 

Hadi
,

Accepted tutorial: Video Coding Advancements in HTTP Adaptive Streaming

Tutorial title: Video Coding Advancements in HTTP Adaptive Streaming

Venue: IEEE International Conference on Multimedia & Expo (ICME) 2025 (https://2025.ieeeicme.org/tutorials/)

We are happy to announce that our tutorial “Video Coding Advancements in HTTP Adaptive Streaming” (by Hadi Amirpourazarian and Christian Timmerer) has been accepted for IEEE ICME 2025, which will take place in Nantes, France, June 30 – July 4, 2025.

Description: This tutorial provides a comprehensive exploration of the HTTP Adaptive Streaming (HAS) pipeline, covering advancements from content provisioning to content consumption. We begin by tracing the history of video streaming and the evolution of video coding technologies. Attendees will gain insights into the timeline of significant developments, from early proprietary solutions to modern adaptive streaming standards like HAS. A comparative analysis of video codecs is presented, highlighting milestones such as H.264, HEVC, and the latest standard, Versatile Video Coding (VVC), emphasizing their efficiency, adoption, and impact on streaming technologies. Additionally, new trends in video coding, including AI-based coding solutions, will be covered, showcasing their potential to transform video compression and streaming workflows.

Building on this foundation, we explore per-title encoding techniques, which dynamically tailor bitrate ladders to the specific characteristics of video content. These methods account for factors such as spatial resolution, frame rate, device compatibility, and energy efficiency, optimizing both Quality of Experience (QoE) and environmental sustainability. Next, we highlight cutting-edge advancements in live streaming, including novel approaches to optimizing bitrate ladders without introducing latency. Fast multi-rate encoding methods are also presented, showcasing how they significantly reduce encoding times and computational costs, effectively addressing scalability challenges for streaming providers.

The tutorial further delves into edge computing capabilities for video transcoding, emphasizing how edge-based architectures can streamline the processing and delivery of streaming content. These approaches reduce latency and enable efficient resource utilization, particularly in live and interactive streaming scenarios.

Finally, we discuss the QoE parameters that influence both streaming and coding pipelines, providing a holistic view of how QoE considerations guide decisions in codec selection, bitrate optimization, and delivery strategies. By combining historical context, theoretical foundations, and practical insights, this tutorial equips attendees with the knowledge to navigate and address the evolving challenges in video streaming applications.