On-premises Integration Server

An on-premises Integration Server is any Integration Server that is configured to share service metadata with IBM® webMethods Integration and enables the execution of those services by integrations that are defined in IBM webMethods Integration. The service metadata that is uploaded from an on-premises Integration Server provides the accounts and applications that are required to create integrations in IBM webMethods Integration.

IBM webMethods Integration has environments and each environment has a different endpoint. You can configure one on-premises Integration Server to an environment.

Important: On-premises Integration Server version 10.3 or later with the latest fix works with IBM webMethods Integration for hybrid connectivity.

A group of services that you share with IBM webMethods Integration is called as an application. Applications are created on the on-premises Integration Server and uploaded to IBM webMethods Integration. IBM webMethods Integration can run any service that is hosted by an on-premises Integration Server. When you share services in an application, you are sharing the metadata for the service. For Integration Server services, the metadata you share is the service name, service signature, display name, and service comments. Users of IBM webMethods Integration can then create integrations that start services that are defined in the applications.

Size Limit for Hybrid Integration messages

The maximum size limit of an encoded service output to be transported over Hybrid Integration is 50 MB. The following encoding limits are enforced:

  • java.lang.String uses a fixed width encoding of 2 bytes per java.lang.Character. In general, all data types excluding java.lang.Integer and java.lang.Byte are converted to java.lang.String before encoding, for example, a java.lang.String must be less than 25 MB, however, a byte must be less than 50 MB.

  • Each field or value within the Service Output is framed with metadata during encoding, and this is added to the encoded data.

  • The Service Output being transported is encoded into a single cumulative piece, so consideration for the number of fields and their cumulative sizes are needed.

Using multiple on-premises webMethods Integration Servers to process requests from IBM webMethods Integration

Multiple on-premises Integration Servers can receive and process requests from IBM webMethods Integration. The on-premises Integration Servers can be in a clustered environment (stateful or stateless) or in a nonclustered environment. When multiple Integration Servers are configured to receive and process requests from IBM webMethods Integration, requests are distributed to each Integration Server in a round-robin fashion.

To use multiple on-premises Integration Servers to process requests from IBM webMethods Integration, the following conditions must be true for all the on-premises Integration Servers:

  • Configuration for IBM webMethods Integration tenants, accounts, and applications are the same across all the Integration Servers.
  • The packages containing services to be called from IBM webMethods Integration are the same across all Integration Servers. If you change a package on one of the on-premises Integration Servers, you must propagate the changes to the other on-premises Integration Servers.
  • Accounts and applications must be uploaded once from one of the Integration Servers, that is, you do not need to upload the accounts and applications from every Integration Server.
  • Accounts must be enabled on all Integration Servers to receive and process incoming requests from IBM webMethods Integration.