Search

Future smart factories are expected to deploy an emerging dynamic Virtual Reality (VR) applications with high bandwidth wireless connections in the THz communication bands, where a factory worker can follow activities through 360{\deg}video streams with high quality resolution. THz communications, while promising as a high bandwidth wireless communication technology, are however known for low fault tolerance, and are sensible to external factors. Since THz channel states are in general hard to estimate, what is needed is a system that can adaptively react to transceiver configurations in terms of coding and modulation. To this end, we propose an adaptive control plane that can help us configure the THz communication system. The control plane implements a workflow algorithm designed to adaptively choose between various coding and modulation schemes depending on THz channel states. The results show that an adaptive control plane can improve throughput and signal resolution quality, with theoretically zeroed bit error probability and a maximum achievable throughput in the scenarios analayzed., Comment: This paper is uploaded here for research community, thus it is for non-commercial purposes

Intent-based networking (IBN) solutions to managing complex ICT systems have become one of the key enablers of intelligent and autonomous network management. As the number of machine learning (ML) techniques deployed in IBN increases, it becomes increasingly important to understand their expected performance. Whereas IBN concepts are generally specific to the use case envisioned, the underlying platforms are generally heterogenous, comprised of complex processing units, including CPU/GPU, CPU/FPGA and CPU/TPU combinations, which needs to be considered when running the ML techniques chosen. We focus on a case study of IBNs in the so-called ICT supply chain systems, where multiple ICT artifacts are integrated in one system based on heterogeneous hardware platforms. Here, we are interested in the problem of benchmarking the computational performance of ML technique defined by the intents. Our benchmarking method is based on collaborative filtering techniques, relying on ML-based methods like Singular Value Decomposition and Stochastic Gradient Descent, assuming initial lack of explicit knowledge about the expected number of operations, framework, or the device processing characteristics. We show that it is possible to engineer a practical IBN system with various ML techniques with an accurate estimated performance based on data from a few benchmarks only.

A few potential IoT communication protocols at the application layer have been proposed, including MQTT, CoAP and REST HTTP, with the latter being the protocol of choice for software developers due to its compatibility with the existing systems. We present a theoretical model of the expected buffer size on the REST HTTP client buffer in IoT devices under lossy wireless conditions, and validate the study experimentally. The results show that increasing the buffer size in IoT devices does not always improve performance in lossy environments, hence demonstrating the importance of benchmarking the buffer size in IoT systems deploying REST HTTP., Comment: This paper is uploaded here for research community, thus it is for non-commercial purposes

With the rapid advances in IoT, edge and cloud computing solutions, it is critical to educate and train students in computer science and engineering in various aspects of IoT-edge-cloud (IoT-E-C) system architecture implementations. We outline the design and development of an undergraduate laboratory course that sets the goal of implementing various interfaces and communication protocols to connect IoT, edge and cloud computing systems and evaluating their performance. The lab setup is modular and based on open source tools. In the IoT context, it consists of low-cost processing platforms with various sensors and actuators. In the edge and cloud computing context, we implement and deploy single board computers and Firebase cloud solutions, respectively. The modular lab setup allows students to engineer and integrate various communication protocol solutions, including MQTT, COAP and HTTP. In addition to the system implementation, students can evaluate and benchmark the performance of the entire system.

The most recent development in the underlying QUIC transport protocol has revived the interest in HTTP over QUIC (HTTP/3) as a communication protocol solution in cloud and edge computing. This development is notable especially given the rise in implementations of IoT edge/cloud continuum, a new computing paradigm that extends the cloud computing IoT via edge computing systems. In IoT cloud/edge continuum, benchmarking performance of the communication protocol of choice is critical. We focus for the first time on experimental benchmarking of HTTP/3 performance in an IoT cloud/edge continuum system. To benchmark the performance in terms of scalability and latency, we first implement HTTP/3 from a set of non-standardized open-source libraries, - which is challenge, and use the implementation in two main IoT scenarios: edge computing only (IoT devices running HTTP/3 clients and edge devices running an HTTP/3 server), and cloud/edge continuum (based on Google Firebase). Experimental results show that latency and scalability remain a challenge for HTTP/3, but its appeal also remains, in its wide adoption, embedded security and compatibility with existing networked systems.

The deployment of deep neural network (DNN) models in software applications is increasing rapidly with the exponential growth of artificial intelligence. Currently, such models are deployed manually by developers in the cloud considering several user requirements, while the decision of model selection and user assignment is difficult to take. With the rise of edge computing paradigm, companies tend to deploy applications as close as possible to the user. Considering this system, the problem of DNN model selection and the inference serving becomes harder due to the introduction of communication latency between nodes. We present an automatic method for DNN placement and inference in edge computing; a mathematical formulation to the DNN Model Variant Selection and Placement (MVSP) problem is presented, it considers the inference latency of different model-variants, communication latency between nodes, and utilization cost of edge computing nodes. Furthermore, we propose a general heuristic algorithm to solve the MVSP problem. We provide an analysis of the effects of hardware sharing on inference latency, on an example of GPU edge computing nodes shared between different DNN model-variants. We evaluate our model numerically, and show the potentials of GPU sharing, with decreased average latency by 33% of millisecond-scale per request for low load, and by 21% for high load. We study the tradeoff between latency and cost and show the pareto optimal curves. Finally, we compare the optimal solution with the proposed heuristic and showed that the average latency per request increased by more than 60%. This can be improved using more efficient placement algorithms.

In Network Function Virtualization (NFV), the re-liability of Service Function Chain (SFC) is one of the most critical performance parameters. Since each SFC has different composition, placement, configuration and deployment requirements, the exact analysis of SFC reliability is a challenge. In this paper, we address this challenge by analyzing the end-to-end service reliability as a combined function of SFC placement strategy over Inter- and Intra-DCNs and failures of hardware and software components. The analysis introduces a novel reliability benchmark, the so-called Affordable Component Failure which can take into account the interdependency of different component failures, i.e., failure of data centers, racks, servers and virtual network function as well as their impact on availability of other SFC components and the overall SFC reliability.

While edge computing is envisioned to superbly serve latency sensitive applications, the implementation-based studies benchmarking its performance are few and far between. To address this gap, we engineer a modular edge cloud computing system architecture that is built on latest advances in containerization techniques, including Kafka, for data streaming, Docker, as application platform, and Firebase Cloud, as realtime database system. We benchmark the performance of the system in terms of scalability, resource utilization and latency by comparing three scenarios: cloud-only, edge-only and combined edge-cloud. The measurements show that edge-only solution outperforms other scenarios only when deployed with data located at one edge only, i.e., without edge computing wide data synchronization. In case of applications requiring data synchronization through the cloud, edge-cloud scales around a factor 10 times better than cloud-only, until certain number of concurrent users in the system, and above this point, cloud-only scales better. In terms of resource utilization, we observe that whereas the mean utilization increases linearly with the number of user requests, the maximum values for the memory and the network I/O heavily increase when with an increasing amount of data.

Edge computing is a promising technology that can fulfill the requirements of latency-critical and computation-intensive applications. To further enhance the performance, coded edge computing has emerged because it can optimally utilize edge devices to speed up the computation. In this paper, we tackle a major security issue in coded edge computing: how to verify the correctness of results and identify attackers. Specifically, we propose an efficient verification scheme, namely Decode-and-Compare (DC), by leveraging both the coding redundancy of edge devices and the properties of linear coding itself. To design the DC scheme, we conduct a solid theoretical analysis to show the required coding redundancy, the expected number of decoding operations, and the tradeoff between them. To evaluate the performance of DC, we conduct extensive simulation experiments and the results confirm that the DC scheme can outperform existing solutions, such as homomorphic encryption and computing locally at the user device.

With the edge computing becoming an increasingly adopted concept in system architectures, it is expected its utilization will be additionally heightened when combined with deep learning (DL) techniques. The idea behind integrating demanding processing algorithms in Internet of Things (IoT) and edge devices, such as Deep Neural Network (DNN), has in large measure benefited from the development of edge computing hardware, as well as from adapting the algorithms for use in resource constrained IoT devices. Surprisingly, there are no models yet to optimally place and use machine learning in edge computing. In this paper, we propose the first model of optimal placement of Deep Neural Network (DNN) Placement and Inference in edge computing. We present a mathematical formulation to the DNN Model Variant Selection and Placement (MVSP) problem considering the inference latency of different model-variants, communication latency between nodes, and utilization cost of edge computing nodes. We evaluate our model numerically, and show that for low load increasing model co-location decreases the average latency by 33% of millisecond-scale per request, and for high load, by 21%.

Optical networks are prone to physical layer attacks, in particular the insertion of high jamming power. In this paper, we present a study of jamming attacks in elastic optical networks (EON) by embedding the jamming into the physical layer model, and we analyze its impact on the blocking probability and slots utilization. We evaluate our proposed model using a single link and a network topology and we show that for in-band- jamming, the slots utilization decreases with the increase of jamming power, and becomes null when the jamming power is higher than 3 dB, while for out-of-band jamming, the impact is maximal for a specific jamming power, 1.75 dB in our simulation. Considering multiple positions of attackers, we attained the highest blocking probability 32% for a specific jamming power 2 dB. We conclude that the impact of jamming depends on attacker positions as well as the jamming power.

Anonymity networks are an increasingly important part of Internet evolution to providing practical mechanisms to enhance anonymity and security of network users. The best known anonymity network, The Onion Routing (Tor), is not only subject to user deanonymization attacks but also subject to censorship attacks by blocking the public and secret Tor entry routers, e.g., bridges, or by jamming Tor traffic to prevent users to access Tor network. To help further evolution of anonymity networks, we focus in this work on censorship attacks, and propose to enhance the well-known multipath Tor technique with linear network coding (LNC). We analyze anonymity, privacy and censorship success considering different types of adversaries and the amount of honest and compromised components in the Tor and in the underlying network. The results show that LNC can significantly improve the robustness of Tor against censorship and maintain sufficient anonymity and privacy levels.

In recent years, edge computing has attracted increasing attention for its capability of facilitating delay-sensitive applications. In the implementation of edge computing, however, data confidentiality has been raised as a major concern because edge devices may be untrustable. In this paper, we propose a design of secure and efficient edge computing by linear coding. In general, linear coding can achieve data confidentiality by adding random information to the original data before they are distributed to edge devices. To this end, it is important to carefully design code such that the user can successfully decode the final result while achieving security requirements. Meanwhile, task allocation, which selects a set of edge devices to participate in a computation task, affects not only the total resource consumption, including computation, storage, and communication, but also coding design. In this paper, we study task allocation and coding design, two highly-coupled problems in secure coded edge computing, in a unified framework. In particular, we take matrix multiplication, a fundamental building block of many distributed machine learning algorithms, as the representative computation task, and study optimal task allocation and coding design to minimize resource consumption while achieving information-theoretic security.

Spectrum defragmentation (DF) as a connection reconfiguration method is essential in elastic optical networks (EONs) in order to minimize connection blocking probability. This paper proposes the first exact Markov model to computing exact blocking probabilities in EONs with DF by taking into account the occupancy status of spectrum slices of all network links, and waiting and serving connections during a DF process. Since the complexity of the exact Markov model increases exponentially with the network capacity and size, we propose a reduced state model where a link occupancy state is defined by the total occupied slices on a fiber link. Furthermore, using a spectrum fragmentation factor in each occupancy state we calculate state-and class-dependent connection setup rates, which is used to compute approximate blocking in EONs with DF. Notably, we show individually the distinct blocking values, one due to resource unavailability and the other due to fragmentation, both under a random-fit and a first-fit spectrum allocation policies. Our numerical results show that the DF process is very useful in reducing overall connection blocking in EONs. We also observe that blocking due to spectrum fragmentation can be reduced, but not eliminated in a mesh network topology even when an optimal DF scheme is deployed.

We study how erasure coding can improve service reliability in Data Center Networks (DCN). To this end, we find that coding can be best deployed in systems, where i) traffic is split into multiple parallel sub-flows, ii) each sub-flow is encoded; iii) SFC along with their corresponding Virtual Network Functions (VNF) concatenated are replicated into at least as many VNF instances as there are sub-flows, resulting in parallel sub- SFCs; and iv) all coded sub-flows are distributed over parallel paths and processed in parallel. We study service reliability as function of the level of parallelization within DCN and the resulting amount of redundancy. Based on the probability theory and by considering failures of path segments, VNF and server failures, we analytically derive the probability that parallel subflows are successfully processed by the parallelized SFC and that the original serial traffic can be successfully recovered without service interruptions.We compare the proposed failure protection with coding and the standard backup protection and evaluate the related overhead of both methods, including decoding, traffic redirection and VNF migration. The results not only show the benefit of our scheme for reliability, but also a reduced overhead required in comparison to backup protection.

A specific source of harmonic currents in power grid are fluorescent lamps with electromagnetic ballast. A new model of fluorescent lamps with electromagnetic ballast is presented in this paper. Also, applicability to simulations of fluorescent lamp performance for steady state source voltages and for its RMS time variations with nominal frequency is also demonstrated. The presented model is based on an extended curve fitting procedure against to the existing standard fluorescent lamp models. For the approximation of the lamp's i-u characteristic, a pair of exponential functions representing the fluorescent lamp branches is chosen. The obtained simulation results were verified by comparison against the measurements and an excellent agreement was found. It is also demonstrated that the suggested fluorescent lamp model delivers accurate results for different steady state source voltages including its variations. The presented and suggested lamp model can be efficiently used in harmonic studies and its implementation in the existing EMTP/ATP or MATLAB/SPS simulation tools is straightforward.

In this paper, we study end-to-end service reliability in Data Center Networks (DCN) with flow and Service Function Chains (SFCs) parallelism. In our approach, we consider large flows to i) be split into multiple parallel smaller sub-flows; ii) SFC along with their VNFs are replicated into at least as many VNF instances as there are sub-flows, resulting in parallel sub-SFCs; and iii) all sub-flows are distributed over multiple shortest paths and processed in parallel by parallel sub-SFCs. We study service reliability as a function of flow and SFC parallelism and placement of parallel active and backup sub-SFCs within DCN. Based on the probability theory and by considering both server and VNF failures, we analytically derive for each studied VNF placement method the probability that all sub-flows can be successfully processed by the parallelized SFC without service interruption. We evaluate the amount of backup VNFs required to protect the parallelized SFC with a certain level of service reliability. The results show that the proposed flow and SFC parallelism in DCN can significantly increase end-to-end service reliability, while reducing the amount of backup VNFs required, as compared to traditional SFCs with serial traffic flows.

Smart sensing and computing have become important concepts in the last few years, creating opportunities in the new sector of smart agriculture. A few commercial smart agriculture systems have been introduced to this end, and albeit closed for experimentation, are paving the way for high-tech innovations for crop and livestock agriculture. In this paper, we focus on open and low-cost concepts for smart fog (edge) computing systems to create a smart farm animal welfare monitoring system. We develop an open source system that enables networking and computing of edge devices but also processing of data in a server - all being a connected system that we refer to as smart farm computing system. We propose to use Raspberry Pis as edge devices to monitor the animals and the farm environment, and we let the edge devices communicate with a local farm controller. The proposed farm computing system conceptually creates a fog computing layer and is further connected with cloud computing systems and a mobile application. We demonstrate that a low-cost and open computing and sensing system can effectively monitor multiple parameters related to animal welfare.