Research Archive

Journal Article Open Access Green Computing

Energy Proportionality in Large-Scale Data Centers: Power Usage Effectiveness Optimization Through Workload Consolidation, Cooling System Intelligence, and Renewable Energy Integration

Data centers globally consume approximately 1-2 percent of world electricity supply, with power usage effectiveness (PUE) -- the ratio of total facility power to IT equipment power -- serving as the primary industry efficiency benchmark. Despite significant progress in server-level energy proportionality, facility-level PUE optimization through the joint management of workload consolidation, cooling system intelligence, and renewable energy procurement presents substantial unresolved engineering challenges. This paper presents DataGreen, an integrated energy management framework for large-scale data centers that coordinates workload scheduling, cooling control, and renewable energy utilization through a model predictive control architecture. DataGreen is evaluated through a combination of simulation using a validated data center energy model and a 12-month deployment in a 15 MW hyperscale data center. DataGreen achieves a mean annual PUE of 1.12 compared to the baseline PUE of 1.34, representing a 16.4 percent reduction in total facility power. Workload consolidation using thermal-aware VM placement contributes a 7.1 percent power reduction, DeepMind-inspired ML-based cooling optimization contributes 6.8 percent, and renewable energy time-shifting (scheduling deferrable workloads to periods of high renewable availability) contributes an effective 9.2 percent reduction in carbon intensity. We introduce the Data Center Carbon Efficiency Index (DCCEI) that combines PUE with renewable energy fraction and workload carbon intensity, and demonstrate that DCCEI provides a more complete sustainability picture than PUE alone.

Seun Adesanya, Hanna Bergstrom, Ryo Kawamoto, Nadia Khalil· Oct 2019· 312 citations

Cite View PDF

Journal Article Subscription Software Engineering

Database DevOps: Version-Controlled Schema Migration, Automated Database Testing, and Continuous Delivery of Persistent Data Layer Changes

Despite significant maturation of application-layer continuous delivery practices, the database tier remains one of the most common manual intervention points in otherwise automated DevOps pipelines. Schema migrations, data migrations, and stored procedure deployments are frequently managed through ad-hoc scripts and manual DBA approval gates that fragment delivery pipelines and introduce deployment risk. This paper presents Database DevOps as a coherent engineering discipline addressing the full lifecycle of database change management within CI/CD workflows. We evaluate three schema migration frameworks — Flyway, Liquibase, and Alembic — against six operational criteria: migration idempotency, rollback capability, branching support, CI integration maturity, drift detection, and cross-engine portability. We further characterize a taxonomy of database testing strategies — schema contract tests, referential integrity tests, performance regression tests, and data quality invariant tests — and provide empirical evidence of their fault detection effectiveness from a controlled experiment involving 1,200 intentionally injected database defects. Organizations implementing full Database DevOps practices in our case studies reduced database-related deployment failures by 71% and eliminated manual DBA gates in 83% of deployment pathways. We provide a Database DevOps Implementation Roadmap as a practitioner artifact.

Adunola Fashola, Erik Johansson, Daisuke Mori, Ana Sofia Ferreira· Oct 2019· 256 citations

Cite View PDF

Journal Article Open Access Cloud Computing

Federated Learning in Distributed Cloud Environments: Communication Efficiency, Differential Privacy Guarantees, and Model Convergence Under Non-IID Data Distributions

Federated learning enables collaborative model training across distributed clients without centralizing raw data, making it particularly valuable in regulated industries where data sharing is legally constrained. However, the practical deployment of federated learning in production cloud environments exposes significant engineering challenges: communication overhead from model update aggregation, degraded convergence under non-IID (heterogeneous) data distributions, and the tension between differential privacy noise injection and model utility. This paper presents FedCloud, a production-oriented federated learning framework designed for multi-organization deployment across cloud environments, and reports its empirical evaluation across three deployment scenarios: cross-hospital clinical NLP, cross-bank fraud detection, and cross-retailer demand forecasting. FedCloud implements adaptive FedAvg with gradient compression (achieving 7.3x communication reduction), client selection strategies optimized for stragglers in cloud environments, and Renyi differential privacy with per-round epsilon tracking. Under non-IID distributions simulating real organizational data heterogeneity, FedCloud achieves within 2.8% of centralized training accuracy for the clinical NLP task and within 4.1% for fraud detection, while satisfying epsilon less than 3 per training round. We characterize the accuracy-privacy-communication three-way trade-off surface empirically and provide a configuration selection guide for practitioners deploying federated learning under specific regulatory and performance constraints.

Nkechi Eze, Lars Berggren, Akira Yamamoto, Nadia Khalil· Jul 2019· 441 citations

Cite View PDF

Journal Article Subscription Cloud Computing

Serverless Computing in DevOps Pipelines: Performance Trade-offs, Cold Start Mitigation, and Organizational Adoption Patterns

Serverless computing platforms have disrupted traditional DevOps deployment models by abstracting infrastructure management entirely, yet their integration into continuous delivery pipelines introduces novel performance, cost, and observability challenges. This paper presents a comprehensive empirical evaluation of serverless-native DevOps workflows, combining platform benchmarking experiments with a multi-organization case study involving six companies that migrated production workloads to serverless architectures. Our benchmarks quantify cold start latency distributions across AWS Lambda, Azure Functions, and Google Cloud Functions under varying memory configurations, runtime environments, and invocation patterns. We find that cold starts impose median latency penalties of 180–740 ms depending on runtime and configuration, with Java and .NET runtimes exhibiting the highest variance. We introduce four cold start mitigation strategies — Provisioned Concurrency, Keep-Warm Scheduling, Lightweight Runtime Selection, and Dependency Minimization — and evaluate their cost-performance trade-offs. Case study findings reveal that serverless adoption fundamentally reshapes team topologies, reducing infrastructure operations burden by 65% while requiring new competencies in cost observability and function granularity design. This work provides the most comprehensive empirical treatment of serverless DevOps integration to date.

Kenji Watanabe, Amara Diallo, Florian Huber, Priscilla Nwosu· Jul 2019· 334 citations

Cite View PDF

Journal Article Open Access Software Engineering

DevOps Adoption in Small and Medium Enterprises: Tailoring Practices, Toolchains, and Organizational Models for Resource-Constrained Engineering Teams

The DevOps literature is disproportionately dominated by case evidence from large technology companies — Google, Netflix, Amazon — whose organizational scale, engineering budgets, and talent pools are unrepresentative of the vast majority of software-producing organizations globally. This paper redresses this imbalance through a focused empirical study of DevOps adoption in small and medium enterprises (SMEs), defined as organizations with fewer than 250 employees. We conducted a longitudinal study across 19 SMEs over 24 months, combining quarterly interviews, pipeline telemetry analysis, and a dedicated SME DevOps survey instrument (n=488 respondents from 141 SMEs). Our findings reveal that SMEs face a distinct set of adoption challenges: role conflation (developers as operators as security engineers), toolchain cost sensitivity, absence of dedicated platform teams, and regulatory naivety. We develop the SME DevOps Adaptation Framework (SDAF), which tailors the DORA research model`s four key metrics and associated practices to SME constraints, proposing lightweight toolchain stacks, role-sharing governance models, and incremental adoption roadmaps. SMEs that adopted SDAF practices over 12 months achieved deployment frequency improvements of 340% and MTTR reductions of 52% from baseline, demonstrating that DevOps value is highly accessible to smaller organizations when appropriately adapted.

Obinna Ikechukwu, Helena Svensson, Ryota Hayashi, Lucia Barbosa· Apr 2019· 287 citations

Cite View PDF

Journal Article Open Access Augmented and Virtual Reality

Reducing Cybersickness in Virtual Reality Through Adaptive Field-of-View Restriction, Frame Rate Stabilization, and Predictive Head Motion Compensation

Cybersickness -- the constellation of nausea, disorientation, and discomfort experienced by a significant proportion of users during VR immersion -- remains the primary barrier to widespread consumer VR adoption, with prevalence estimates ranging from 20 to 80 percent depending on content type, session duration, and individual susceptibility. This paper presents a comprehensive engineering study of cybersickness mitigation through rendering pipeline interventions, evaluating three techniques -- adaptive field-of-view (FOV) restriction, frame rate stabilization through Asynchronous Spacewarp (ASW) and ATW mechanisms, and predictive head motion compensation using IMU-driven pose prediction -- individually and in combination. A controlled user study (n=168 participants) employs the Simulator Sickness Questionnaire (SSQ) and physiological measures (galvanic skin response, heart rate variability) under three VR content categories: locomotion-heavy, rotation-heavy, and stationary content. Combined FOV restriction and predictive motion compensation reduces SSQ total severity scores by 51 percent for locomotion-heavy content and 43 percent for rotation-heavy content, with no statistically significant reduction in presence scores. Frame rate stabilization contributes most significantly for users with high flicker sensitivity (SSQ reduction of 34 percent in the high-sensitivity subgroup). We introduce the Cybersickness Mitigation Effectiveness Index (CMEI) and provide an adaptive rendering pipeline reference architecture for VR headset manufacturers and game engine developers.

Chiamaka Eze, Erik Magnusson, Akira Nakamura, Sara Rodrigues· Apr 2019· 356 citations

Cite View PDF

Journal Article Open Access Human-Computer Interaction

Conversational Agents in Enterprise Software Workflows: Usability, Trust Calibration, and Productivity Impact of Chatbot Integration in Knowledge Work Environments

Enterprise chatbot deployments have proliferated across knowledge work environments, yet rigorous evaluation of their usability, trust calibration accuracy, and measurable productivity impact remains sparse relative to the volume of deployment activity. This paper presents a mixed-methods study of enterprise conversational agent integration across five organizations in legal, financial, and healthcare knowledge work domains, combining a 12-week longitudinal experiment (n=214 participants) with qualitative interviews and log analysis of 340,000 conversational interactions. We evaluate chatbot usability using the Conversational Agent Usability Scale (CAUS), which we develop and validate as part of this work across 11 usability dimensions including intent recognition accuracy, response coherence, context retention, and error recovery behavior. The longitudinal experiment finds that well-designed chatbot integration reduces time spent on information retrieval tasks by 31% and on routine document generation by 44%, but increases task completion time by 18% for complex multi-step reasoning tasks where chatbot error rates are highest. A central finding is trust miscalibration: 67% of users exhibit overtrust in chatbot outputs for factual queries within their domain of expertise, leading to unchecked propagation of erroneous information. We propose a Trust Calibration Interface Design framework comprising four evidence-presentation patterns that reduce overtrust incidence by 48% in a controlled follow-up study.

Adanna Obi, Marcus Eriksson, Yuko Tanaka, Sara Fonseca· Jan 2019· 367 citations

Cite View PDF

Journal Article Open Access Cloud Computing

GitOps: Declarative Infrastructure Management and Its Impact on Deployment Reliability and Audit Compliance in Cloud Environments

GitOps has emerged as a deployment methodology that uses Git repositories as the single source of truth for both application configuration and infrastructure state, enabling automated reconciliation between desired and actual system state. Despite growing practitioner adoption, rigorous empirical evaluation of GitOps impact on operational outcomes remains limited. This paper presents a mixed-methods study combining a controlled experiment with a practitioner survey (n=412) to evaluate GitOps adoption outcomes across reliability, compliance, and team productivity dimensions. In our controlled experiment, teams using GitOps-based workflows with Flux CD and ArgoCD achieved a 52% reduction in failed deployments and a 44% improvement in audit log completeness relative to script-based deployment teams. Survey findings indicate that Git-based change control satisfies regulatory audit requirements more completely in 71% of cases compared to ad-hoc deployment scripts. We also identify three anti-patterns — Secret Sprawl, Repo Monolithism, and Drift Blindness — that undermine GitOps implementations and propose mitigation strategies for each. This work provides both an empirical foundation for GitOps adoption decisions and actionable engineering guidance for practitioners.

Alexandre Dubois, Chinyere Uzoho, Vikas Sharma, Emma Thorvaldsen· Jan 2019· 467 citations

Cite View PDF