Architecture and Tools

DSE Version: 6.0


In this unit, we will dive deeper into the architecture of DataStax Enterprise Graph. We will take a look at the architecture from three perspectives: technologies, components, deployment. Then we'll look at the tools available for DataStax Enterprise Graph.


Now lets dive a little deeper into the architecture of DSE Graph

To do so, we're going to use this format to guide the discussion.

First we'll take a look at the architecture from 3 unique perspectives:




Then we'll look at the tools available for DSE Graph.

This is just the overview portion of this course so we will keep things high level for now.

We will start this discussion by viewing DSE Graph as a blackbox that powers graph applications.

The upper most layer of DSE Graph, the view that's accessible directly from our clients is the Gremlin APIs, Spark APIs, and proprietary graph schema and configuration APIs.

DSE Graph is able to capitalize on the robust, responsive, and resilient store layer of DataStax's version of Apache Cassandra

The middleware layer is the proprietary codebase that connects and optimized the connection between the graph APIs and storage layer.

Applications and users access data stored in DSE Graph trough the DataStax drivers, either using a string API or builder style API we call the Fluent API/

DataStax Studio is a fantastic visual tool that provides graph developers a visual way to interact with DSE Graph and optimize Gremlin.

For those that prefer a CLI approach to development, the open source Gremlin console is provided.

Under the covers, DSE Graph projects a property graph onto Apache Cassandra. Because DSE Search is integrated into DataStax's version of Apache Cassandra. At write time, data is automatically indexed by DSE Search when search indexes are configured on a vertex. DSE Graph exposes datastax proprietary DSE Search methods through the Gremlin API to leverage the power of DSE Search through Gremlin.

Here you can see a high level overview of how DSE Graph connects the API layer with the storage layer.

DSE Graph exposes both the DSE Server, for connections from DataStax drivers, and the Gremlin Server, for connections through the open source Tinkerpop drivers. For OLAP processing, DSE Analytics, is automatically engaged.

Just like a standard DSE cluster topology, users of DSE Graph can achieve workload isolation through the virtual data center concept. This means users can ensure that OLTP and OLAP workloads do not interfere with each other without the need to manually extract/move data between environment. Because DSE Graph is built on DataStax’s version of Apache Cassandra, data is replicated throughout the cluster.

Here we see a screenshot of DSE Studio, the visual tool developers use to explore their graph, interact with CQL, optimize and debug Gremlin, and with the latest release, interact with Spark SQL.

DSE Studio is a tremendous help to users creating graph databases in DSE Graph, particularly developers who appreciate a visual understanding of their graph.

For those users who come from an open source graph database background or who prefer a command line experience, the Gremlin Console is the right tool choice. The Gremlin Console exposes Gremlin to end users through a fully Open Source compatible CLI tool.

When users need to load data into DSE Graph, they have several choices.

Use DSE GraphFrames to bulk load data quickly and easily through Spark.

Create custom loading applications using the DSE driver and the Fluent Batch API.

Or use the DSE Graph Loader utility, which can import and process flat files as well as graph specific file formats like gryo.

Because DSE Graph is a part of the DSE product suite, all of the great management tools that are available for DSE work with DSE Graph. This includes Opscenter which can be used for automated deployments, backups, and other operational needs.

Now it's your turn to explore DSE Graph. Please work through Exercise 1, getting started with DSE Studio for an introduction to DSE Graph. We’ll be ready to pick up with the next section of the course when you return.


No write up.
No Exercises.
No FAQs.
No resources.
Comments are closed.