AWS’ fourth serverless database service, Amazon Neptune Serverless is a secure, scalable, fully-managed database service that gives users the ability to build and run graph applications.
Taylor Riggan, senior graph architect at Amazon Web Services for Amazon Neptune, joined DBTA’s webinar, “Deep Dive Into Amazon Neptune Serverless Database,” to discuss Neptune’s capabilities in-depth.
Today’s modern cloud-based development environment is driven by loosely coupled micro-services and purpose-built data stores, explained Riggan. With a graph database, you can quickly and easily create and navigate relationships between data. Most commonly, these sorts of databases are used for fraud detection, social networking, and recommendation engines. Graph databases allow you to explore the connections and patterns within highly connected data, ultimately providing dynamic insights.
“If you’re coming from a traditional relational database background, you may think of this as queries that need lots of joins or require a cursive data access pattern, or a common table expression to be able to express in SQL,” said Riggan. “Graph databases are designed to support those access patterns in a much simpler fashion. They have query languages that help you express those patterns in an intuitive way, versus having to handle all of that in traditional SQL.”
The use cases for graph databases are growing by the day, Riggan explained. The common uses for Amazon Neptune are divided into categories, including knowledge graphs, identity graphs, fraud graphs, and security graphs. Though they are technically use cases, Riggan emphasized that it is more like a pattern for graph-based development.
Knowledge graphs link disparate and heterogeneous data sources together to discover hidden connections; identity graphs link all related devices and ids, enabling unified profile creation, targeting, and personalization; fraud graphs locate and prevent fraudulent patterns of transactions as they are occurring; and security graphs detect threats to your environment such as unwanted user access to applications or exposed resources.
Amazon Neptune is equipped to manage all of these graph-based patterns, operating on four primary tenets that define its operation: open standards to prevent framework lock-in, fast performance at scale, reliable functions with replicas, and easy query build.
AWS continues to heavily innovate upon Neptune, Riggan highlighted. Some improvements include max capacity increasing to 128TB; the launch of general availability of openCypher inside Neptune; the launch of global databases for replicating data across regions; Python integration for graph analytics; Amazon Neptune ML; a free trial option of Neptune; and much more.
Cost expenditure and scale has become an increasing concern for enterprises, and in turn, a concern for Neptune.
“With Neptune, we’ve seen a drive to help customers cost-optimize their workloads and help with scaling as their workloads grow over time,” said Riggan. “In the past, users would have to define a certain size instance in your Neptune clusters. If you wanted to scale, you’d have to change those instances manually.”
The answer to these growing concerns is Amazon Neptune Serverless, a serverless graph database that allows customers to scale workloads automatically. Workloads with dynamic demands are best suited for Neptune Serverless, as the database automatically scales to peak demands for accelerated adaptation to customer needs and streamlines application build, all while lowering costs with a pay-for-value billing model that optimizes resource utilization.
As CPU utilization peaks, “Neptune Serverless scales up immediately, and it does scale really fast—and then it also scales back down as it notices that it no longer needs that additional capacity.”
To learn more about Amazon Neptune and its serverless iteration, you can view an archived version of this webinar here.