Data domain cluster

One or more data domain clusters can be hosted in a clustered Endeca Server deployment. This topic introduces the data domain cluster and describes its capabilities.

A data domain cluster is a set of Dgraph processes that together handle requests for a given data domain. One of these Dgraph processes is responsible for handling all write operations (updates, configuration changes), while the remaining Dgraph processes serve as read-only. All Dgraph nodes in a given data domain cluster utilize the same index residing on shared storage. Each data domain cluster node is hosted on a separate Endeca Server node.

In this guide, the term Dgraph node is used interchangeably with the term data domain node. The Dgraph node is the node in the data domain cluster that runs the Dgraph process. The Dgraph process of the Oracle Endeca Server is the main computational module that provides the features of the Endeca Server, such as search, refinement computation, and Guided Navigation. For information about the capabilities of the Dgraph process, see the Oracle Endeca Server Developer's Guide.

Additionally, the term data domain is used interchangeably with the term data domain cluster.

For definitions of the data domain cluster and the data domain nodes see Data domain cluster concepts.

Data domain cluster capabilities

A data domain cluster provides the following capabilities:
  • Enhanced availability of query processing by the Oracle Endeca Server. In a cluster of Dgraphs (a data domain cluster), if one of the Dgraph nodes fails, queries continue to be processed by other Dgraph nodes.
  • Increased throughput by the Oracle Endeca Server. In a data domain cluster, you change throughput capacity by adding Dgraph nodes. This allows you to spread the query load across them, without the need to increase storage requirements at the same rate.
  • Routing of requests to the data domains. Requests to a given data domain can be submitted to any Endeca Server instance in the Endeca Server cluster, regardless of whether or not it hosts the target data domain. The Endeca Server instance receiving the request will route it appropriately.
  • Automatic allocation of the data domain leader node. The leader and follower nodes differ in the types of queries they can process, however, this is transparent to the end users of applications running on top of Endeca data domains. The allocation of leader and follower node roles is performed by the Endeca Server cluster, when you add data domain clusters. Allocation of the leader node ensures high availability of data loading.