ilona analytics : infrastructure

Node Requirements

Each analysis environment - real time, near real time and batch - requires one or a cluster of compute nodes depending on workload. Nodes may be physical, virtual or a docker container.

Each node requires the following external components

• A supported 64-bit operating system. One of
  • Linux: Redhat Enterprise Server, SUSE Linux Enterprise Server, Ubuntu
  • Microsoft Windows Server
  • Unix: IBM AIX, HP-UX, Oracle Solaris
• A 64-bit Java Runtime Environment (JRE) and Java Development Kit (JDK), which must be compatible with the selected operating system. One of
  • IBM JDK
  • OpenJDK
  • Oracle JDK
• A Java Transaction Architecture (JTA) compatible transaction monitor. This can be one provided by a JEE server or a stand alone implementation. The selected solution must be compatible with the selected operating system and Java environment. Supported JEE servers are
  • IBM WebSphere Application Server
  • JBoss EAP
  • Oracle WebLogic Server
  A stand alone JTA monitor is bundled as a component of ilona analytics
• A connection to an XA enabled data store. This can can be an RDBMS accessed via JDBC or Apache HBase accessed via our XA capable ilona hbase client. Supported RDMBS are
  • IBM DB2
  • Microsoft SQL Server
  • MySQL
  • Oracle Database
  • PostgreSQL
  Apache HBase is bundled as a component of ilona analytics

Additional Components

In addition to the external components above, the following are required

• Infrastructure to support the data stores accessed by the analysis environments, provisioned according to the recommendations of the vendor. See above for supported data stores.
• A workflow management system. As discussed in ilona analytics : routing and workflow this can be the system bundled as a component of ilona analytics or a third party system. Requirements for the former are the same as for an analysis environment, see the vendor recommendations for the latter.
• When analysis environments are clustered, as they invariably are for production, a cluster management solution is required. This will vary according to how and where the nodes are deployed - on physical or virtual hardware, on premises or in a cloud.