• HOME
  • NEWS
  • EXPLORE
    • CAREER
      • Companies
      • Jobs
    • EVENTS
    • iGEM
      • News
      • Team
    • PHOTOS
    • VIDEO
    • WIKI
  • BLOG
  • COMMUNITY
    • FACEBOOK
    • INSTAGRAM
    • TWITTER
Thursday, November 19, 2026
BIOENGINEER.ORG
No Result
View All Result
  • Login
  • HOME
  • NEWS
  • EXPLORE
    • CAREER
      • Companies
      • Jobs
        • Lecturer
        • PhD Studentship
        • Postdoc
        • Research Assistant
    • EVENTS
    • iGEM
      • News
      • Team
    • PHOTOS
    • VIDEO
    • WIKI
  • BLOG
  • COMMUNITY
    • FACEBOOK
    • INSTAGRAM
    • TWITTER
  • HOME
  • NEWS
  • EXPLORE
    • CAREER
      • Companies
      • Jobs
        • Lecturer
        • PhD Studentship
        • Postdoc
        • Research Assistant
    • EVENTS
    • iGEM
      • News
      • Team
    • PHOTOS
    • VIDEO
    • WIKI
  • BLOG
  • COMMUNITY
    • FACEBOOK
    • INSTAGRAM
    • TWITTER
No Result
View All Result
Bioengineer.org
No Result
View All Result
Home NEWS Science News Technology

Open Microservices Benchmark Publicitaki Delivers High Cloud Scalability at Minimal Cost

by
November 19, 2026
in Technology
Reading Time: 5 mins read
0
Open Microservices Benchmark Publicitaki Delivers High Cloud Scalability at Minimal Cost

Open Microservices Benchmark Publicitaki Delivers High Cloud Scalability at Minimal Cost

Share on FacebookShare on TwitterShare on LinkedinShare on RedditShare on Telegram

Computer science researchers have developed and validated an open-access microservices application named Publicitaki, designed to provide academic and research laboratories with a realistic, technologically diverse, and inexpensive experimental testbed. Distributed software architectures that decompose large enterprise platforms into autonomous microservices have become standard practice across modern industry. However, investigating their performance bottlenecks, debugging distributed workflows, and diagnosing operational anomalies has remained persistently difficult for academic groups. Commercial microservices architectures frequently present sprawling interdependencies known informally as death stars, yet private enterprises rarely release proprietary source code to researchers. Meanwhile, existing demonstration platforms developed for academia frequently suffer from single-language codebases, discontinued maintenance, or mandatory deployment on complex Kubernetes clusters that incur substantial financial overheads.

To overcome these systemic constraints, investigators Tiago Simões and Filipe Araújo engineered an architecture that deliberately reflects the heterogeneity, communication patterns, and operational demands of enterprise software while prioritizing budget accessibility. Described in the International Journal of Networked and Distributed Computing, the Publicitaki benchmark models an online price-tracking and e-commerce ecosystem consisting of three separate systems: Publicitaki itself, a connected Store platform, and an asynchronous Bank system. Instead of constraining the application to a single runtime or requiring expensive container orchestration frameworks, the authors implemented the benchmark using multiple programming languages, independent database engines, and automated Infrastructure as Code scripts. When deployed on Amazon Web Services using AWS Elastic Beanstalk, the entire test suite executed with cloud resource charges totaling under one United States dollar.

The structural motivation behind the project addresses specific technical trade-offs inherent in microservices software engineering. Transitioning from monolithic architectures to decoupled services enables independent team development, rapid release cycles, and isolated failure domains, but it fundamentally shifts architectural complexity from internal software logic to the underlying distributed networking environment. Services must communicate across physical and virtual machine boundaries using remote application programming interfaces, introducing network latencies, complex serialization steps, and concurrency issues. In evaluating previous research benchmarks, the authors noted that systems such as Sock Shop, Robot Shop, DeathStarBench, and Tea Store frequently relied on Kubernetes, which demands either costly managed control planes or labor-intensive manual cluster configuration. Furthermore, older suites often lacked representative asynchronous messaging, polyglot programming environments, or native mechanisms for automated cloud infrastructure provisioning.

Publicitaki addresses these technological gaps by dividing its e-commerce functionality into three modular software systems interacting across distinct protocols. The primary Publicitaki platform operates as an aggregator search engine that allows users to look up consumer goods, observe historical price fluctuations across multiple merchants, and select preferred vendors. The Store platform functions as an independent e-commerce retail portal where referred consumers can browse product catalogs, inspect descriptions and images, read customer reviews, and complete purchases. The third component, the Bank system, acts as a dedicated financial service that manages user balances, verifies available liquidity, and executes monetary transfers between buyers and merchant accounts. Together, these systems establish a realistic transaction pipeline requiring continuous cross-service collaboration.

To replicate real-world enterprise engineering practices, the developers implemented each system across different technology stacks and programming languages. The Publicitaki platform utilizes four microservices: a presentation frontend developed with React, Vite, and TypeScript, an account management service built on Quarkus with Java, and specialized product catalog and notification services constructed using Node.js and TypeScript. Communication within Publicitaki predominantly relies on Representational State Transfer APIs over Hypertext Transfer Protocol. To simulate real-world financial settlement latencies, asynchronous communication between the services and the banking backend is routed through a HiveMQ message broker utilizing the Message Queue Telemetry Transport protocol under a publish-subscribe paradigm.

The Store and Bank platforms further broaden this technological diversity by introducing additional frameworks and remote procedure call mechanisms. The Store system comprises four microservices: an ASP.NET MVC frontend written in C#, a product management backend developed in Django with Python, and separate user and notification services implemented in Spring Boot with Java. Within the Store, the frontend communicates with the user management microservice via Google Remote Procedure Calls, exemplifying high-performance synchronous inter-service communication. The Bank platform incorporates three microservices, combining an ASP.NET MVC C# presentation tier with Spring Boot Java services for customer account handling and notification delivery. Across the entire application, state persistence is handled independently by MySQL and PostgreSQL database instances, reinforcing database-per-service decoupling.

In total, the benchmark comprises 11 distinct microservices and 7 dedicated database management systems, encompassing over 116,000 lines of code across its three subsystems. Deployment flexibility is achieved through a two-tiered provisioning strategy. For local defect correction and component testing, each system includes dedicated Docker Compose configuration files, enabling complete execution on developer workstations without cloud expenses. For cloud deployment, the authors developed automated Terraform scripts tailored for AWS. Terraform was selected over proprietary alternatives because its provider-agnostic infrastructure definitions allow research groups to adapt deployment routines across alternative cloud providers with minimal re-engineering, eliminating the need to maintain manual cloud configurations.

Rather than deploying onto a managed Kubernetes control plane such as Amazon Elastic Kubernetes Service, which entails fixed baseline expenses regardless of usage, the researchers configured the cloud deployment using AWS Elastic Beanstalk. Elastic Beanstalk operates as a Platform as a Service that automatically orchestrates virtual machine provisioning, Classic Elastic Load Balancing, and autoscaling across Amazon Elastic Compute Cloud instances without base management charges. The evaluation environment was instantiated in the europe-north-1 AWS region using economical t3.micro virtual machine instances equipped with two virtual central processing units and one gibibyte of random-access memory, alongside db.t3.micro instances for the relational database tier.

The research team conducted extensive performance evaluations using the Locust load-testing framework to determine whether the architecture could satisfy target non-functional requirements under rising traffic volumes. The non-functional specification established that all individual microservices should respond within 500 milliseconds when overall system traffic reaches 100 requests per second. The investigators simulated realistic user sessions featuring a 90 percent read to 10 percent write distribution across five distinct scenarios, scaling client populations from initial baselines of 10 users up to high-stress levels of 1,000 concurrent clients generating sustained traffic against eleven individual service endpoints.

Under low-load conditions of up to 10 concurrent clients, Elastic Beanstalk maintained stable operations using a single virtual machine instance per service, with individual microservices responding in approximately 100 milliseconds. Under moderate loads ranging up to 100 clients, the automated scaling engine triggered dynamic scale-out routines, provisioning two to three virtual machine instances based on outbound network traffic thresholds. In sustained 60-minute stress tests with 250 concurrent clients, Locust generated 400,111 total requests, achieving an average throughput of 111.1 requests per second. The system registered an overall average response time of 171 milliseconds, with the 90th and 95th percentiles remaining below 170 and 340 milliseconds respectively, well within the target threshold.

High-load testing also exposed architectural choke points, demonstrating the benchmark’s utility for distributed systems research. When the store frontend microservice was restricted to a maximum threshold of four compute instances under 1,000 clients, latency increased dramatically, culminating in 504 server timeout errors. Performance metrics revealed that central processing unit utilization across instances remained modest, staying between 10 and 20 percent. Instead, performance degradation stemmed from outbound network saturation in the downstream store-products microservice, which transfers heavy image files to the frontend orchestrator. When the investigators expanded the scaling ceiling to ten instances, Elastic Beanstalk successfully provisioned five frontend and four product service instances, reducing average response times back below 100 milliseconds and stabilizing network traffic at approximately 5 mebibytes per second.

The financial assessment confirmed the platform’s viability for resource-constrained laboratories. In baseline idle mode, the entire deployment consisting of 11 compute instances and 7 database instances cost approximately 0.2474 United States dollars per hour. Under heavy load with default scaling caps of four replicas per service, operational costs reached approximately 0.5906 dollars per hour, while the worst-case stress scenario registered an expense of 0.5698 dollars per hour. Because cloud resources can be provisioned and destroyed rapidly via automated Terraform scripts, the team completed all experimental evaluation runs with less than one dollar in cumulative cloud service charges, demonstrating that realistic microservices research does not require substantial financial budgets.

Subject of Research: A low-cost, representative microservices research benchmark application deployed on AWS Elastic Beanstalk

Article Title: A Representative, Low-Cost, and Easy-to-Deploy Microservices Application for Research Purposes

Article References: Simões, T., & Araújo, F. (2026). A Representative, Low-Cost, and Easy-to-Deploy Microservices Application for Research Purposes. International Journal of Networked and Distributed Computing, 14(1), Article 16. https://doi.org/10.1007/s44227-025-00065-0

Image Credits: AI Generated

DOI: 10.1007/s44227-025-00065-0

Keywords: microservices, cloud computing, AWS Elastic Beanstalk, software engineering, distributed systems, horizontal scalability, Docker containerization, Terraform, Locust load testing, REST API, MQTT messaging, gRPC, performance benchmarking, DevOps, infrastructure as code, Publicitaki, system monitoring, cost optimization

News Source: Denise Maddox. (November 19, 2026). Open Microservices Benchmark Publicitaki Delivers High Cloud Scalability at Minimal Cost. Scienmag.

Tags: AWS Elastic Beanstalkcloud computingcost optimizationDevOpsDistributed systemsDocker containerizationgRPChorizontal scalabilityInfrastructure-as-CodeLocust load testingmicroservicesMQTT messagingperformance benchmarkingPublicitakiREST APISoftware Engineeringsystem monitoringTerraform
Share12Tweet7Share2ShareShareShare1

Related Posts

Silicomanganese Fume Emerges as Low-Carbon Supplementary Binder for Sustainable Concrete

Silicomanganese Fume Emerges as Low-Carbon Supplementary Binder for Sustainable Concrete

November 19, 2026
Mid-size Unmanned Ground Vehicles Face Severe Mobility and Scalability Constraints

Mid-size Unmanned Ground Vehicles Face Severe Mobility and Scalability Constraints

November 19, 2026

Industrial Monitoring Framework Couples Sensor Topology with Language Models for Operational Decisions

November 19, 2026

Ceramic Waste Additive Enhances Stereolithography 3D Printing Resins

November 19, 2026

POPULAR NEWS

  • Cyanoacrylate Adhesives Speed Extraoral Skin Closure but Certainty Remains Low

    Cyanoacrylate Adhesives Speed Extraoral Skin Closure but Certainty Remains Low

    29 shares
    Share 12 Tweet 7
  • Mid-size Unmanned Ground Vehicles Face Severe Mobility and Scalability Constraints

    29 shares
    Share 12 Tweet 7
  • Computational Framework Matches Biodegradable Implant Resorption Rates to Bone Healing Timelines

    29 shares
    Share 12 Tweet 7
  • Robotic Exoskeleton Gait Training Earns High Satisfaction Despite Wearability and Cost Challenges

    29 shares
    Share 12 Tweet 7

About

We bring you the latest biotechnology news from best research centers and universities around the world. Check our website.

Follow us

Recent News

Arabic KOOS-12 Demonstrates High Reliability and Validity for Knee Osteoarthritis Assessment

Silicomanganese Fume Emerges as Low-Carbon Supplementary Binder for Sustainable Concrete

Glaucoma Filtration Outcomes Rely on Perioperative Ocular Surface Preservation

Subscribe to Blog via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Join 89 other subscribers
  • Contact Us

Bioengineer.org © Copyright 2023 All Rights Reserved.

Welcome Back!

Login to your account below

Forgotten Password?

Retrieve your password

Please enter your username or email address to reset your password.

Log In
No Result
View All Result
  • Homepages
    • Home Page 1
    • Home Page 2
  • News
  • National
  • Business
  • Health
  • Lifestyle
  • Science

Bioengineer.org © Copyright 2023 All Rights Reserved.