Digital Auction Platform

To enable the sale of diverse art from a local art gallery I worked together with others at VizworX to develop a state-of-the-art (pun intended) web application to support the display and auction of art pieces digitally. This application investigated existing platforms for a quick solution but it was determined that the client's needs did not justify existing third-party options. A new application was developed to allow the client to store, track, and display artwork with auctions handled directly in the same application.

Development

The application was developed using Next.JS being used for the primary platform, this enabled us to quickly prototype and develop alongside designers and produce a modern and minimalist interface. For all secure pages and endpoints we leveraged a third-party OAuth 2.0 identity provider, this provider is used for both administrators and public users of the platform.

The auction interfaces are connected to an Express service developed specifically to handle the complex logic of bids, auction states, and rapid data synchronization between bidders. This backend service provided REST and WebSocket endpoints to balance the needs of singular requests and subscription-based updates to users. This combination allowed us to implement auctions in a more modern form than existing competitors.

Deployment

For this project, our client had minimal existing applications and also had limited plans for future software or expansion. For these reasons, we did not have or plan to deploy a dedicated Kubernetes cluster and instead went for a service that could run our two specific containers with scaling. For this purpose, the Elastic Container Service from AWS was selected in addition to an Aurora PostgreSQL database cluster.

The deployment is set up so that cost is minimal when auctions are not occurring and has automatic scaling configured to account for traffic spikes during auctions. This was achieved through the separation of concerns relating to serving content and managing bids, with a separate micro-service handling bid details from users.

The project was fully deployed in early 2025 with a first auction planned for later in the year. Through our testing, we have reduced areas of concern by leveraging features of PostgreSQL and Socket.IO to improve throughput and reduce wasted traffic.

Skills: Next.JS, React, Express, Socket.io, Stripe, PostgreSQL, RBAC, Docker, AWS