Deploying AnzoGraph

AnzoGraph DB can be deployed on-premises or on any of the common cloud platforms. AnzoGraph also integrates with popular container management applications. This topic lists each of the available deployment methods, provides links to the deployment instructions for each method, and describes any differences in the features that are available with each deployment type.

Deployment Options

The table below lists the available deployment methods and provides links to the deployment instructions for each method.

Deployment Method Details
Docker Engine - Community With this option, you use Docker on Linux, Mac, or Windows 10 to deploy AnzoGraph as a single container instance. Docker does not support cluster configurations.

For more information, see Docker Deployments.

Kubernetes with Helm With this option, you use Helm to deploy AnzoGraph on a local or remote Kubernetes cluster. You can deploy AnzoGraph as a single instance or multiple instances in a cluster.

To deploy AnzoGraph with Kubernetes Minikube, you must use Minikube Version 1.10 or later. The Linux kernel that ships with earlier Minikube versions is not sufficient for running AnzoGraph DB.

For more information, see Kubernetes Deployments.

RHEL/CentOS Tarball With this option, you deploy AnzoGraph by unpacking a tarball onto a server with a RHEL or Centos 7.4+ operating system. You can use cloud instances or on-premise servers, and you can deploy AnzoGraph on a single server or on any number of servers in a cluster.

For more information, see RHEL/CentOS Tarball Deployments.

AWS CloudFormation Services With this very advanced option, you use Amazon's CloudFormation service to deploy a fully automated and managed AnzoGraph stack. You can deploy AnzoGraph on a single EC2 instance or multiple instances in a cluster.

For more information, see AWS CloudFormation Deployments.

IBM Cloud Pak With this option, you use Helm to deploy AnzoGraph DB on IBM Cloud Pak.

For more information, see IBM Cloud Pak Deployments.

By default, all deployments include an embedded anonymous free edition license. The license does not expire, but it is valid for single server deployments only and it enforces a limit of 8 GB RAM usage.

Though AnzoGraph will not use more than 8 GB RAM in the free edition version, Cambridge Semantics recommends that you deploy AnzoGraph on a host server that has at least 16 GB of RAM available.

To increase the memory limit or remove all limitations, you can upgrade the license by registering with Cambridge Semantics. For more information, see Upgrading an AnzoGraph License.

Feature Exceptions by Deployment Type

Features such as Using AnzoGraph Analytics, application programming interfaces, and the command line interface are available regardless of deployment method. Certain features, however, are excluded in some deployments and might impact your decision when choosing a deployment method. The table below describes the features that vary by deployment method.

Feature Description Availability
Query and Admin Console A graphical user interface for running queries and performing administrative tasks.
  • Available in Docker, Kubernetes, and AWS CloudFormation deployments.
  • Not included with RHEL/CentOS tarball deployments but can be installed separately with Docker. See Deploying the AnzoGraph Frontend Container.
Data Science Functions These functions provide distribution probability, correlation, profiling, and frequency calculations such as those commonly used for statistical analysis and machine learning. For more information, see Additional Data Science Functions (Preview).
  • Available in Docker, Kubernetes, and AWS CloudFormation deployments.
  • Not included with RHEL/CentOS tarball deployments.
Geospatial Functions These functions offer advanced capabilities for developing large scale location intelligence and geospatial applications. For more information, see Additional Geospatial Functions (Preview).
  • Available in Docker, Kubernetes, and AWS CloudFormation deployments.
  • Not included with RHEL/CentOS tarball deployments.
Apache Arrow Functions

This collection of user-defined service functions supports the Arrow Flight protocol for Apache Arrow integration with leading ML and other Big Data Ecosystems including Python Pandas, Spark MLLIB and Google Tensorflow, Cassandra, Kudu, and Hadoop. For more information, see Apache Arrow Functions (Preview).

  • Available in Docker, Kubernetes, and AWS CloudFormation deployments.
  • Not included with RHEL/CentOS tarball deployments.
Data on Demand The Data on Demand service generates Open Data Protocol (OData)-based feeds that can be used to access data programmatically via a RESTful API or from third-party business intelligence tools. The service is available in the AnzoGraph Frontend Container. For more information, see Accessing Data Using OData Protocol (Preview).
  • Available in Docker, Kubernetes, and AWS CloudFormation deployments.
  • Not included with RHEL/CentOS tarball deployments but can be installed separately with Docker. See Deploying the AnzoGraph Frontend Container.
Data Toolkit The Data Toolkit service is a flexible and configurable interface that enables AnzoGraph to connect to upstream data sources and query data or extract, load, and transform the data. For more information, see Connecting to Sources with the Graph Data Interface (Preview).
  • Available in Docker, Kubernetes, and AWS CloudFormation deployments.
  • Not included with RHEL/CentOS tarball deployments.

The main selection criteria for choosing deployment options, as well as the associated server sizing and scaling, starts with the analytic applications you intend to use with AnzoGraph. Consider the data sources and size of the data you plan to load as well as your query performance requirements. For more information, see Sizing Guidelines for In-Memory Storage.

Related Topics