Division of Computer and Network Systems

We propose a unified “shift-left” security validation framework that integrates static vulnerability scanning, SBOM generation, image signature verification, policy-as-code enforcement, and best-practice scoring into a single Kubernetes admission-control webhook invoked within the CI/CD pipeline. By atomically intercepting container admission requests, the system produces a Software Bill of Materials, cross-references CVE feeds, validates digital signatures, applies dynamically loaded JSON/YAML policies, and computes a weighted KubeScore for Pod specifications prior to deployment. Evaluation on realistic workloads demonstrates end-to-end processing latency below 200 ms and detection rates exceeding 95 % for critical vulnerabilities and misconfigurations. This consolidated approach eliminates post-deployment scans, accelerates feedback loops, strengthens compliance auditability with immutable logs, and lays the foundation for future AI-driven remediation and multi-cluster policy synchronization. Index Terms – Admission-Control Webhook, SBOM (Software Bill of Materials), Static Vulnerability Analysis, CVE (Common Vulnerabilities and Exposures), Image Signature Verification, Cosign, Notary v2, Policy-as-Code, Open Policy Agent, Kyverno, KubeScore, kube-bench, Dynamic Policy Loading, Hot-Reload, Automated Remediation, Self-Healing, Immutable Audit Logs, CI/CD Pipeline Integration, Kubernetes Best Practices, SBOM Formats: CycloneDX, SPDX, Threat Modeling, Compliance Automation.

Federated learning (FL) enables collaborative model training without sharing raw data, which is ideal for telecom applications spanning on-premises datacenters, Azure, and third-party (3PC) clouds. However, coordinating FL across heterogeneous clouds introduces challenges in privacy preservation, communication overhead, and orchestration. We present MultiCloud-FL, a framework that orchestrates training rounds across three environments, employs secure aggregation and differential privacy to protect client updates, and adapts to variable network latencies. In experiments training an image-classification model on a synthetic telecom dataset partitioned across environments, MultiCloud-FL achieved: • 94.2 % accuracy (vs. 95.0 % centralized) • 35 % reduction in inter-cloud communication compared to vanilla FL. • Differential-privacy (1.0) with 1.5 % accuracy loss We detail system architecture, secure protocols, mermaid diagrams, quantitative evaluation, and discuss limitations and future directions.

The rapid expansion of multi-cloud infrastructures amplifies energy consumption and carbon footprint, challenging telecom operators to balance performance, cost, and sustainability. This paper introduces EcoSched, a green-cloud scheduler that dynamically places and migrates workloads across on-premises clusters and public clouds (Azure, AWS) to minimize total energy use while satisfying latency and availability SLAs. EcoSched combines workload forecasting, carbon-intensity APIs, and a multi-objective optimization engine. In a 30-day experiment with video-processing and signaling services under bursty loads, EcoSched reduced energy consumption by 24 % and carbon emissions by 18 % compared to baseline round-robin scheduling, with SLA violation rate under 1 %. We detail system design, prediction models, scheduler algorithm, architecture, quantitative evaluation, and discuss deployment considerations

ABSTRACT: Telecom applications often experience sudden traffic spikes (e.g., call-setup surges, signaling bursts) that challenge infrastructure elasticity and cost controls. This paper compares serverless functions (AWS Lambda) and containerized microservices (Kubernetes on EC2) under bursty load, measuring end-to-end latency, request-processing throughput, and cost per 1 million requests. In a simulated telecom signaling workload with 100 to 1,000 requests/sec spikes, here are my observations: Serverless cold-start latency up to 300 ms vs. container steady latency 50 ms • Throughput saturation at 800 req/sec for Kubernetes vs. 1,200 req/sec for Lambda (with concurrent executions) • Cost per 1 M requests: $12.5 (Lambda) vs. $8.7 (container) under sustained bursts We present architecture and sequence diagrams, detailed methodology, results analysis, and discuss trade-offs for telecom operators.

Abstract The paper describes a scheme for moving objects in image sequence. The scheme adopts a distribution-based model approach to moving object detection. The methodology involves pattern classification using higher-order statistics (HOS) in a clustering framework. The proposed method effectively combines statistical knowledge about the class of moving objects with motion information. The unknown distribution of the image patterns of objects is approximately modeled using HOS information derived from sample images. Our proposed algorithm uses an HOS-based decision measure which is derived from a series expansion of the multivariate probability density function in terms of the multivariate Gaussian and the Hermite polynomial. A data clustering that uses HOS is proposed to facilitate the decision process. The higher-order closeness measure is used to classify test clusters as foreground or background.

The article presents the torchosr package - a Python package compatible with PyTorch library - offering tools and methods dedicated to Open Set Recognition in Deep Neural Networks. The package offers two state-of-the-art methods in the field, a set of functions for handling base sets and generation of derived sets for the Open Set Recognition task (where some classes are considered unknown and used only in the testing process) and additional tools to handle datasets and methods. The main goal of the package proposal is to simplify and promote the correct experimental evaluation, where experiments are carried out on a large number of derivative sets with various Openness and class-to-category assignments. The authors hope that state-of-the-art methods available in the package will become a source of a correct and open-source implementation of the relevant solutions in the domain.

In order to avoid well-known paradoxes associated with self-referential definitions, higher-order dependent type theories stratify the theory using a countably infinite hierarchy of universes (also known as sorts), Type_0 : Type_1 : *s. Such type systems are called cumulative if for any type A we have that A : Type_i implies A : Type_{i+1}. The Predicative Calculus of Inductive Constructions (pCIC) which forms the basis of the Coq proof assistant, is one such system. In this paper we present the Predicative Calculus of Cumulative Inductive Constructions (pCuIC) which extends the cumulativity relation to inductive types. We discuss cumulative inductive types as present in Coq 8.7 and their application to formalization and definitional translations.

SRAM-based weak physically unclonable functions (PUFs) have shown promise regarding tamper sensitive key storage and device ID generation. Weak PUFs rely on intrinsic process variations to produce repeatable and unique start-up behavior. However, noise in the system can affect the start-up behavior and introduce errors. A number of solutions, such as fuzzy extraction and error correcting codes have been proposed to alleviate the effect of noise and generate stable keys. Unfortunately, the overhead from these techniques grows superlinearly with increasing error rate. Alternatively, accelerating device aging or burn-in has been shown to reduce start-up error rate significantly, which leads to reduced overhead for error correction. Unfortunately, burn-in is a time-intensive process and accrues significant production cost. This paper proposes a method to reduce the cumulative burn-in time via quantification of the minimum burn-in requirements for chip. We propose a low-cost proxy to measure the degree of process variation of each device at birth and use previously proposed device aging models to determine the burn-in requirements. Our results show that this procedure reduces cumulative burn-in cost without compromising the resultant reliability of weak PUFs. Furthermore, we analyze the effect of having additional PUF cells, alternate weak PUF designs and transistor technologies on burn-in requirements.

In order to avoid well-know paradoxes associated with self-referential definitions, higher-order dependent type theories stratify the theory using a countably infinite hierarchy of universes (also known as sorts), Type$_0$ : Type$_1$ : $\cdots$ . Such type systems are called cumulative if for any type $A$ we have that $A$ : Type$_{i}$ implies $A$ : Type$_{i+1}$. The predicative calculus of inductive constructions (pCIC) which forms the basis of the Coq proof assistant, is one such system. In this paper we present and establish the soundness of the predicative calculus of cumulative inductive constructions (pCuIC) which extends the cumulativity relation to inductive types.

Purpose – The purpose of this paper is to study cyber-enabled manufacturing systems (CeMS) for additive manufacturing (AM). The technology of AM or solid free-form fabrication has received considerable attention in recent years. Several public and private interests are exploring AM to find solutions to manufacturing problems and to create new opportunities. For AM to be commercially accepted, it must make products reliably and predictably. AM processes must achieve consistency and be reproducible. Design/methodology/approach – An approach we have taken is to foster a basic research program in CeMS for AM. The long-range goal of the program is to achieve the level of control over AM processes for industrial acceptance and wide-use of the technology. This program will develop measurement, sensing, manipulation and process control models and algorithms for AM by harnessing principles underpinning cyber-physical systems (CPS) and fundamentals of physical processes. Findings – This paper describes the challenges facing AM and the goals of the CeMS program to meet them. It also presents preliminary results of studies in thermal modeling and process models. Originality/value – The development of CeMS concepts for AM should address issues such as part quality and process dependability, which are key for successful application of this disruptive rapid manufacturing technology.