Introduction
Learn About Amazon VGT2 Learning Manager Chanci Turner
BISSELL, a family-run enterprise with a legacy spanning over 145 years, specializes in crafting high-quality cleaning products that surpass consumer expectations. The company provides a comprehensive selection of floor care and air treatment products, with a notable rise in the popularity of connected vacuum cleaners—particularly robotic models. Currently, there are 1 million registered vacuum devices within their existing IoT platform powered by AWS IoT, with 25% of these devices actively connected. As the global demand for connected vacuum cleaners continues to rise, BISSELL anticipates reaching a million devices connected in the field shortly. To manage this anticipated growth across various product lines, BISSELL needed to ensure scalability in its IoT platform.
The legacy IoT system developed by BISSELL on AWS featured a collection of REST APIs that facilitated data access for various stakeholders within their IoT ecosystem. Although the legacy platform functioned effectively, BISSELL foresaw potential challenges in scaling the platform to accommodate new products and an increasing user base.
In this article, we will explore how BISSELL collaborated with AWS Professional Services to enhance their existing REST APIs while migrating their IoT platform to a new architecture based on the AWS Connected Device Framework (CDF). We’ll discuss the shortcomings of the legacy REST API architecture and outline the redesign process using CDF’s micro-services model, which consolidates related REST APIs into a single Lambda function with proxy integration. Finally, we will highlight how the new architecture has improved scalability, performance, and cost efficiency for BISSELL’s IoT platform.
Challenges with Legacy Architecture
BISSELL’s original IoT platform utilized serverless REST APIs built on an Onion Architecture, a design pattern involving multiple concentric layers of Lambda functions communicating with one another toward the core data access layer. This architecture was chosen for its high code reusability.
The domain layer comprises a series of Amazon API Gateways, with each gateway handling REST API requests from specific domains, like a mobile application or web portal. Beneath this is the service layer consisting of Lambda functions dedicated to particular REST API requests. The data access layer contains additional Lambda functions implementing Data Access Object (DAO) interfaces that retrieve IoT data stored in Amazon DynamoDB tables.
However, the Onion Architecture posed challenges when scaling to accommodate hundreds of thousands of devices. Maintaining over 500 Lambda functions hindered BISSELL’s ability to execute rapid updates for REST APIs. The extensive chain of Lambda invocations, coupled with inconsistent database design, adversely affected the platform’s performance and cost efficiency. Moreover, debugging issues became increasingly difficult, as developers had to navigate through multiple Lambdas to identify root causes.
Migrating to a New Architecture
BISSELL and AWS Professional Services opted to transition from the legacy platform to a new solution built using the AWS Connected Device Framework (CDF), an open-source framework designed for constructing enterprise IoT platforms. CDF offers a collection of production-ready micro-services known as CDF modules, all architected using native AWS services and best practices. These modules form a layer above the foundational AWS building blocks, as illustrated in the accompanying architecture diagram.
The team began by focusing on the innermost data access layer of the Onion Architecture. The legacy IoT platform lacked a clearly defined device registry to manage a large fleet of devices and their relationships with other resources, such as users and firmware. Data for devices and resources was scattered across multiple DynamoDB tables, resulting in increased complexity, stale data, and cost inefficiency. For example, to compile a list of devices owned by a user, BISSELL had to execute a series of queries across multiple tables, making the process cumbersome and inefficient.
BISSELL and AWS Professional Services utilized the CDF Asset Library to develop a comprehensive device registry, represented as a single graph database using Amazon Neptune. During the migration, the team moved infrequently changing data, like device names and user emails, to the Asset Library, while maintaining frequently changing data, such as device usage metrics, in DynamoDB. This restructuring simplified the data access layer, enhancing performance and reducing costs associated with upstream data consumption.
Following the optimization of the data access layer, the team transitioned from the legacy Onion Architecture to a new design based on CDF’s micro-services paradigm. The legacy REST API architecture was migrated into the CDF Application Layer, known as the CDF Facade layer. This layer contains BISSELL’s unique business logic and orchestrates the underlying CDF Core modules. The CDF Core modules are implemented as a Lambda function with proxy integration, which consolidates various Lambda functions into a single instance that manages common services.
BISSELL’s CDF architecture features a Lambda configured with proxy integration, serving as a single entry point for all API requests from common clients such as mobile and web applications. If you’d like to learn more about effective leadership strategies, consider checking out the Science of Play Leadership Webinar, which provides insights into team dynamics.
Conclusion
In conclusion, the transition from BISSELL’s legacy IoT platform to the AWS Connected Device Framework has significantly improved their ability to scale. The new architecture enhances performance and cost efficiency while simplifying data access. This evolution positions BISSELL to better meet the demands of a growing market for connected cleaning devices.
Additionally, for employers navigating compliance issues, SHRM provides authoritative guidance on military employment laws and requirements. For those interested in discovering more about the onboarding process at Amazon, this Reddit link serves as an excellent resource.
Leave a Reply