Event
{Event}  Bring a friend (or 2!) and save 50% on passes for MongoDB World! Register now and save >>

Scaling NFT trading to new heights with MongoDB Atlas and AWS

CNFT.io reduces costs by 4X with MongoDB’s Flex Consulting

INDUSTRY

Technology

PRODUCT

Atlas Database
Atlas Search

USE CASE

NFT ecommerce Marketplace

CUSTOMER SINCE

2021
INTRODUCTION

From $100M TO $22BN: The explosion of NFT trading

While the crypto market never fails to attract masses of attention, it was Non-Fungible Tokens – or NFTs as we more commonly know them – that exploded into the mainstream in 2021.

From a trading volume of around $100m in 2020, the arrival of celebrities, sportstars, and global brands – creating everything from collectibles and digital art to gaming and assets for the metaverse – drove NFT trading to an estimated $22bn in 2021. [1]

Launched in July 2021, CNFT.IO is a marketplace for NFTs, and the first ever built on the Cardano blockchain. Within just six months of its debut, close to $200m of NFTs had been traded on CNFT.IO, representing 80%+ of all NFTs sold on Cardano, and growing at around $1m [2] per day. As of early 2022, the CNFT.IO marketplace serves 150,000 users, and lists over 1.1 million assets in its catalog.

CNFT.IO’s phenomenal growth and rapid delivery of new, market defining features wouldn’t have been possible without the right technology stack. MongoDB Atlas the multi-cloud application data platform, is at the core of the company’s data infrastructure with Atlas Search powering the user experience.

“I’ve been involved with Bitcoin since 2011, but it was the arrival of Cardano in 2015 that ignited my passion for the true potential of blockchain platforms. Its early work with underserved communities in developing African countries, providing citizens with a digital ID that can verify things like academic credentials and land ownership, showcases the transformative power of applications built on the blockchain.”

CNFT.IO Co-Founder

THE CHALLENGE

Debut to downtime: Ditching cloud proprietary databases

CNFT.IO emerged from stealth with a soft-launch of its NFT marketplace on July 26, 2021. The company wanted to test with a few hundred users before announcing its platform to the world. However, news of the launch quickly spread through the Cardano community and within hours had attracted over 6,000 registered users eager to start trading.

While the opening day massively exceeded the company’s expectations, the proprietary cloud document database backing the marketplace did not.

Constant throttling imposed by the database service prevented users from signing up and listing their NFT assets. CNFT engineers frantically searched across product documentation and user forums over the next 24-hours before being able to identify and remediate the problem. Over the next few weeks, the CNFT.IO team encountered ongoing database issues which impacted the marketplace’s stability and the company’s growth.

Despite these poor experiences, the development team remained committed to building the marketplace on a document database. Having used relational databases in the past, they knew the rigidity and fragility of the tabular data model would slow down their ability to continuously ship new platform features. At the same time, trying to take a relational database to the scale needed in the future would add unsustainable cost and complexity. Document databases pose none of these limitations, triggering the CNFT.IO team to begin evaluating MongoDB Atlas as a replacement.

Beyond the database, the development team had built the first version of its marketplace search on Typesense. CNFT.IO’s developers enjoyed working with the search engine, but soon hit limitations in the scalability needed to support growing user search volumes. They also had to take on the complexity of standing up their own syncing mechanism to replicate data from the database to the search engine, adding further friction to the team’s development velocity.

THE SOLUTION

Migrating in weeks, scaling for years

Within a few days of evaluating MongoDB in the Atlas free-tier and moving it to AWS, the engineering team took the decision to migrate all of their data out of the existing proprietary database. By September 2021, the CNFT.IO marketplace had been totally re-platformed onto MongoDB Atlas and AWS.
“As soon as we started evaluating Atlas, MongoDB consulting engineers jumped in to help us optimize our data model and construct the search queries we needed to run. What we have loved about the engagement we’ve had since those earliest days on Atlas is that everyone at MongoDB wants to see us succeed. They are much more of a partner to us than simply a vendor.”

CNFT.IO Co-Founder

Atlas Search enables users to quickly search and discover NFTs listed on the CNFT.IO marketplace

The CNFT.IO platform is built on Javascript, using React on the front-end and Node.js with Express on the backend. MongoDB Atlas running on AWS manages all of CNFT.IO’s user data, sales history, and NFT listings. Once the user selects which NFTs they want to sell from their Cardano wallets, the asset’s metadata is loaded into the listings catalog running in MongoDB Atlas, and then automatically indexed by Atlas Search.

Based on Apache Lucene, Atlas Search provides all of the capabilities users need to find exactly what they are looking for. Atlas Search features such as autocomplete for search-as-you-type and fuzzy search for typo-tolerance enable users to search across marketplace listings faster and more precisely. Lightning fast facets and counts in Atlas Search allows results to be instantly grouped and counted so users can quickly discover the NFTs they are most interested in buying. While most searches today are in English, Atlas Search’s support for more than 40 languages will help CNFT.IO continue to expand its global reach in the future.

As they were designing their technology stack the company’s founders had taken the decision to use managed services wherever possible. This would enable their engineers to focus on building great products and user experiences, rather than get bogged down on undifferentiated infrastructure management. Investing in MongoDB Atlas helps the company reduce costs and de-risk their platform. Atlas bakes in automated operational best practices while providing intelligent performance optimization to run mission-critical applications reliably and securely at scale.

“Since migrating to MongoDB Atlas, the performance on our site has significantly improved, and provided us the confidence to deploy enhancements every day.”

CNFT.IO Co-Founder

Using Atlas Search helps the company maximize the benefits of managed services. This is because as soon as data in the catalog database changes – for example a new listing is inserted, an existing listing is updated, or the schema changes – Atlas automatically synchronizes it with the search index. This avoids the engineering team having to build and maintain their own custom syncing mechanism and update index mappings as new platform functionality is released.

Beyond syncing, developers are also more productive, enabling them to ship new platform features faster. This is because they work with the MongoDB Atlas database and Atlas Search using a single driver and unified API, eliminating context switching between different query languages as they code, and unnecessary build dependencies. As a result, new platform features are delivered to users faster.

THE RESULTS

Continuous delivery at 4X lower cost

“For us, the NFT market is like the space race. To drive growth and stay competitive, we have to constantly innovate and deliver faster than the others. We ship new features and fixes daily, then deliver larger project epics every 2-4 weeks. We could never have delivered at this velocity without the agility of Atlas and AWS and the outstanding support we get from MongoDB.” CNFT.IO Co-Founder

Rather than go alone, CNFT made the strategic decision to turn to MongoDB's Flex Consulting and invest in expert advice early in their journey with MongoDB Atlas to optimize their Atlas cluster. The result was a 4x saving in monthly database costs. When the company’s engineering team migrated to MongoDB they selected the largest available Atlas instance sizes to avoid the service throttling they had encountered with the previous proprietary cloud database. Within just a couple of days, the MongoDB consulting engineer had optimized the MongoDB data model and query patterns, enabling the team to rightsize their Atlas clusters to match application load. The savings unlocked by Flex Consulting have been reinvested into growing the engineering team.

Today CNFT.IO is continuing its rapid headcount expansion. All new hires are put through MongoDB University to accelerate their ramp and time-to-productivity. If you are interested in working at the cutting edge of blockchain and cloud-native technologies, send your resume across to careers@cnft.io.

As CNFT.IO further expands into smart contract-based trading and auctions, it expects its platform to grow 5x-10x within the next year. Building on MongoDB Atlas as its application data platform and serving users with Atlas Search provides the company with the foundations it needs to meet that growth.

GET STARTED

Whether you are building a new application, extending an existing MongoDB workload, or looking to simplify your search estate, Atlas Search makes it easy to get started, and makes your user experiences more engaging and delightful.

Try Atlas Search on a free cluster today.

What will your story be?

MongoDB will help you find the best solution.