Application Integration and Serverless System Design Patterns on AWS


The tutorial will cover how to integrate distributed, Serverless applications with less code on Amazon Web Services (AWS). Decoupling applications at any scale can reduce the impact of changes, making it easier to update and faster to release new features. We will start with an overview of AWS cloud services that enable communication between decoupled components within microservices, distributed systems, and Serverless applications. We will then show how these services can be orchestrated using AWS Step Functions, which is a low-code, visual workflow service that developers can use to build distributed applications and automate IT and business processes. Workflows manage failures, retries, parallelization, service integrations, and observability so developers can focus on higher-value business logic. Finally, we show how to use cloud-native services to externalize system integration design patterns.

  • Overview of cloud-native application integration best practices
  • Introduction to AWS Application integration services (Amazon API Gateway, Amazon Event Bridge, Amazon SNS, Amazon SQS, AWS Step Functions).
  • Serverless computing and AWS Lambda (function-as-a-service).
  • Workflow example using AWS Step Functions.
  • Externalizing system design patterns from your application code.
  • Speakers Bios

    Isael Pelletier

    Mohamed Sabri

    Occupying the space in the middle of 100 Billion+ radio-identifiable things and as many contextual applications


    Context-aware physical spaces make sense of who/what is where/how, providing a real-time data stream, via middleware, that enables countless applications. But as no two physical spaces nor applications are the same, this demands run-anywhere middleware that is as modular as it is extensible. In this tutorial we’ll dissect a variety of real-world examples based on the open source Pareto Anywhere middleware, with critical analysis of its architecture and design, providing firsthand experience with a tangible and relevant middleware challenge.

    Ambient data is all around us, throughout the physical spaces in which we live, work and play. Over 100 billion standard radio-identifiable “things”, in the form of Bluetooth devices and RAIN RFID tags, are discoverable and automatically-identifiable at a human scale of about 10m. Middleware can transform this endless source of data into a real-time stream of what is where/how, which in turn can feed countless complementary contextual applications, from occupancy to experiential to environmental sensing.

    Such middleware is not uncommon, however most examples are designed for specific, siloed applications, rather than general use cases, and have narrow rather than broad compatibility with radio-identifiable devices and infrastructure. First, we’ll examine what it entails for a middleware to enshrine broad compatibility with billions of devices from hundreds if not thousands of vendors, and the compromises that need to be made. Next, we’ll identify what is common across technologies and vendors, which will shape the structure and format of the data stream. Then we’ll present the Pareto Anywhere open source middleware in light of these constraints and compromises to understand the design decisions behind its architecture. With those foundations in place, we’ll examine as many real-world use cases and applications as we can, based on actual deployments (the good and the bad!) for an interactive analysis of the middleware, its strengths and weaknesses. The discussion should foster critical thinking and reflection on the challenges of developing a generic middleware for the real-world.

    The tutorial will include live demos, and participants are invited to experiment with the middleware on a Raspberry Pi, or equivalent device, before, during and/or following the tutorial.

  • Speakers Bios

    Jeffrey Dungen

    The tutorial will be presented by Jeffrey Dungen ( who is the lead developer of reelyActive’s Pareto Anywhere open source middleware. He has a B.A.Sc. in Computer Engineering from the University of Waterloo and a M.A.Sc. in Electrical Engineering (Artificial Intelligence) from Ecole Polytechnique Montreal. He has worked in a technical role at startups in real-time location & context, RFID and computer vision since 2004 (Purelink, Koozoo, reelyActive), and is the co-founder and CEO of reelyActive. He has several IEEE publications related to his work at reelyActive, including the middleware and the standardisation of its data structures. He regularly presents tutorials at the IEEE RFID conference (2017, 2019 & 2022) and taught an undergraduate course in digital logic four times. He is a regular conference speaker.


    All times are Anywhere on Earth (AoE)
    Events Dates
    Full Paper Submission May 15th May 25th, 2022
    Rebuttal July 29th-August 1st, 2022
    Author Notification August 9th, 2022
    Revised Submissions September 9th, 2022
    Notifications of Decisions of Revised Papers September 23rd, 2022
    Camera Ready October 3rd, 2022


    All times are Anywhere on Earth (AoE)
    Events Dates
    Workshop Proposal Submission May 21st, 2022
    Industry Track Full Paper Submission June 30th July 15th, 2022
    Doctoral Symposium Submission September 15 October 9, 2022
    Demo & Poster Submission September 9th September 16th , 2022
    Conference November 7th – 11th, 2022


    All times are Anywhere on Earth (AoE)
    Events Dates
    Full Paper Submission November 20th, 2021
    Rebuttal February 1st-3rd, 2022
    Author Notification February 14th, 2022
    Revised Submissions March 14th, 2022
    Notifications of Decisions of Revised Papers March 28th, 2022
    Camera Ready April 15th, 2022