Skip to main content
Version: v3.x beta

Core Concepts

The sprawl of microservices and APIs which any given product needs to connect to in order to function has led to a new set of challenges for developers and architects.

In the modern era of building applications, the data layer is becoming increasingly complex; it's no longer just a database, but a collection of databases, services, and APIs.

Hasura DDN introduces the concept of a data supergraph to help you manage this complexity.

The state of the data layer​

Imagine the following sets of teams within a hypothetical company:

  • Product Management Team: Responsible for the relational database storing product information. This team focuses on cataloging products, managing inventory, and ensuring product details are accurate and up-to-date by use of a relational PostgreSQL database.

  • User Experience Team: Manages a MongoDB document database and related APIs for storing user information. Their focus is on maintaining user profiles, preferences, and ensuring privacy and security of user data.

  • Search Optimization Team: Handles the integration of an Algolia search service for products and partner stores. They work on optimizing search algorithms, ensuring relevant search results, and improving the overall search experience for users.

  • Finance and Transactions Team: Oversees the Stripe payment service and underlying relational database for processing payments. This includes managing transaction security, payment gateway integrations, and ensuring smooth financial transactions.

  • Logistics and Shipping Team: Responsible for the ShipStation shipping service, fulfilling orders, and the underlying relational database that supports this process. Their primary focus is on logistics, order tracking, and ensuring timely delivery of products.

  • Data Science and Recommendations Team: Manages the Weaviate vector database for storing user recommendations. They work on personalization algorithms, user behavior analysis, and providing tailored product recommendations to enhance user experience.

The diversity of teams within this hypothetical company, each with their unique focus and specialized data services, epitomizes the common challenge in modern data management and API development. These teams all operate with distinct APIs, schemas, and methodologies. This fragmentation not only complicates the process of developing applications but also necessitates the involvement of skilled data architects and engineers to efficiently integrate and maintain these varied data sources.

Subgraphs​

In our e-commerce application example, each service above represents what we term a subgraph.

A subgraph is much more than just a single data source; it is all the metadata needed to act as a self-contained entity that can be independently authored, owned, maintained, and built by an individual team, to then be an interconnected part of a unified API.

For API authors, this means the ability to build out an API — composed of different sources — in a simple, declarative way. You can utilize unique access-control rules, authentication mechanisms, and custom business logic within your subgraphs to create a secure and robust API.

Each subgraph in our example is a modular component of the data supergraph. These components securely work together to overcome the traditional barriers posed by isolated data sources, facilitating a more fluid and interconnected data layer. This integrated approach simplifies application development across diverse data systems, enabling faster iteration and more efficient maintenance by engineering teams.

How do I build a subgraph?​

Subgraphs are composed of related data sources — be that as traditional databases, or as custom business logic — and are connected to your data layer using data connectors. Hasura DDN utilizes a number of data connectors that work with popular databases, services, and APIs out-of-the-box; you can also build your own. These subgraphs can even include your own custom business logic as TypeScript functions that return or mutate data directly via your GraphQL API. This saves you the time and effort of building and maintaining your own APIs to manage data sources and existing microservices.

After connecting your data source, you can then author your metadata using the Hasura CLI and our LSP-enabled VS Code extension. This metadata is then built on your Hasura DDN instance, where it is used to serve your GraphQL API.

Supergraph​

A data supergraph is the framework that brings together all of your subgraphs into one secure API. This orchestration allows for the creation of applications that span multiple data sources, services, third-party APIs, business logic and — most importantly — teams, seamlessly bridging their complexities into a single layer.

For those responsible for maintaining the data layer, this means you now have a single, holistic API to manage, rather than multiple disparate data sources and connections. This enables you to create CI/CD pipelines, monitor performance using industry-standard observability tools, and manage access control in a more streamlined and efficient manner.

For your consumers, this means they have a single endpoint to access all the data they need, thereby reducing the complexity of their applications. This also allows them to focus on building the best possible user experience, rather than wrangling the underlying data layer.

How do I build a supergraph?​

This happens automatically.

As you add subgraphs to your Hasura project, Hasura DDN automatically builds a supergraph of all your data sources. You have the freedom and flexibility to define relationships — including fine-grained access control — across this supergraph, connecting data sources together and making them available in ways that make sense for your application.

Before Hasura DDN​

Businesses, while modernizing, face challenges in delivering digital experiences quickly and cost-effectively. Product development teams are often bottle-necked, waiting on backend teams for secure and efficient data access. This results in manually implementing custom APIs, microservice sprawl, orchestrating security policies, wasting time and stifling innovation.

Hasura DDN basics before diagram

Next steps​

Check out our getting started guide to learn how to build your first data supergraph.

Loading...