Sonia Fahmy’s research while affiliated with Purdue University West Lafayette and other places

... Challenges of near-far LoD schemes include lack of visual continuity between the near and far regions, lack of support for dynamic far regions, and lack of support for motion parallax in the far region, i.e., far region objects at different depths do not move with different speeds in the frame as the user FoV moves. These challenges are being addressed by current research [14], as seen in Figure 2. , yielding frames comparable to ground truth frames rendered from the geometry of the entire virtual environment (right) [14]. ...

Reference:

Ten Ways in which Virtual Reality Differs from Video Streaming

... They propose mechanisms leveraging state-sharing among NFs and exploit parallel execution to reduce the cost of transactional statelessness. A piggybacking mechanism in proposed in [13], similar to the one in this work. However, their approach considers only AMF, SMF and UPF, unlike our Piggyback-based approach that spans all the involved NFs in the execution of control procedures. ...

Reference:

Procedure-Aware Stateless Systems for 5G & Beyond Core Networks

... Broadening our scope, the works we see the closest to ours come from the neighbouring area of microservices-the state-of-the-art style for cloud architectures. Proposals in this direction are by Baarzi and Kesidis [6], who present a framework for the deployment of microservices that infers and assigns affinity and anti-affinity traits to microservices to orient the distribution of resources and microservices replicas on the available machines; Sampaio et al. [33], who introduce an adaptation mechanism for microservice deployment based on microservice affinities (e.g., the more messages microservices exchange the more affine they are) and resource usage; Sheoran et al. [34], who propose an approach that computes procedural affinity of communication among microservices to make placement decisions. Looking at the industry, Azure Service Fabric [25] provides a notion of service affinity that ensures that the replicas of a service are placed on the same nodes as those of another, affine service. ...

Reference:

Affinity-aware Serverless Function Scheduling

... The cloud-native paradigm often emphasizes stateless services to simplify management and improve scalability. However, this approach has trade-offs, such as increased latency due to frequent interactions with external databases for state management [3]. Unlike stateless services, which can be easily restarted elsewhere, stateful microservices must preserve in-memory state during migration. ...

Reference:

Optimizing Stateful Microservice Migration in Kubernetes with MS2M and Forensic Checkpointing

... Depending on the campus layout, computing nodes may be distributed at the locations of the 5G NR base stations or throughout the network infrastructure that connects the 5G NR base stations with the 5GC. Importantly, the 5GC processing is based on cloud-native microservices that can be flexibly processed in cloud computing units [113,114]. ...

Reference:

Ubi-Flex-Cloud: ubiquitous flexible cloud computing: status quo and research imperatives

... Impressive research have been accomplished for coding [3,15,35,36], streaming [1,4,7,11,13,16,18,21,24,[29][30][31][32]37], FoV prediction [2,23,26], and edge-assisted delivery [6,19,27] of 360 video recently. However, a 360 video only covers the whole viewing sphere captured from a center position, and is typically projected onto a 2D plane and processed as a planar video. ...

Reference:

Progressive Frame Patching for FoV-based Point Cloud Video Streaming

... Secondly, they consider RNN, LSTM and Linear Regression (LR) models to predict load traffic in 5G cells and thirdly, a closed-loop automation model is implemented to predict SMF resource usage and automatically instantiate SMF instances. The authors in [46] propose the adoption of AI techniques in order to optimize placement and scaling aspects in the 5G CN. They explore how AI-based scaling algorithms combined with functionality-aware placement can enable the design of network slices. ...

Reference:

An AI-Assisted Framework for Lifecycle Management of Beyond 5G Services

... In the last decade, Machine Learning (ML) and Artificial Intelligence (AI) algorithms achieved tremendous performance in several real-life applications (e.g., Banking [1], Marketing [12], Energy [7]). Modern companies are using these cutting-edge solutions to achieve market differentiation and drive value growth through customer experience. ...

Reference:

Interpretation of net promoter score attributes using explainable AI

... For instance, a logjam attack allows an attacker to downgrade vulnerable transport layer security connections to 512-bit export-level encryption (Schinianakis, 2017). A downgrade attack can also force a UE to use a legacy network, vulnerable to many threats addressed by the newest generation (Angelogianni et al., 2020;Ghosh et al., 2019;Sheoran et al., 2019;Peltonen et al., 2021). ...

Reference:

A systematic literature mapping on the security of 5G and vertical applications

... n j=1 * PW cj + z 3j * L ) + (11) The parameters in equation (11) are similar to those described in equations (3), (4), (5), (6), (7), (8), and (9). ...

Reference:

Derivation and Comparative Analysis of End-To-End Power Models for 5G RAN Architectures