2542828 - SocketTimeoutException error when SAP Analytics Cloud uses Direct type Live Universe Connection (LUC) to connect to a very complex universe on BIP

This KBA relates to the legacy version LUC 2.x only. It is heavily recommended to upgrade to LUC 3.x for Full Support and for a much simpler deployment process.

• SocketTimeoutException error when SAP Analytics Cloud uses Direct type Live Universe Connection (LUC) to connect to a very complex universe on BIP
• Detailed error message in SAC

{"Messages":[{"Type":2,"MessageClass":"javax.ws.rs.client.ClientException: javax.ws.rs.client.ClientException: java.net.SocketTimeoutException: SocketTimeoutException invoking http://YourBIPServer.com:6405/biprws/raylight/v1/documents/12937/dataproviders/DP0/specification: Read timed out (local port 56245 to address 10.47.x.y (YourBIPServer), remote port 6405 to address 10.47.x.y (YourBIPServer)\r\n\tat org.apache.cxf.jaxrs.client.WebClient.doResponse(WebClient.java:1121)\r\n\tat org.apache.cxf.jaxrs.client.WebClient.doChainedInvocation(WebClient.java:1072)\r\n\tat org.apache.cxf.jaxrs.client.WebClient.doInvoke(WebClient.java:878)\r\n\tat org.apache.cxf.jaxrs.client.WebClient.doInvoke(WebClient.java:852)\r\n\tat org.apache.cxf.jaxrs.client.WebClient.invoke(WebClient.java:312)\r\n\tat org.apache.cxf.jaxrs.client.WebClient.put(WebClient.java:330)\r\n\tat com.sap.csina.cloud.utils.HttpRestClient.PutTextXml(HttpRestClient.java:453)\r\n\tat com.sap.csina.cloud.boe.WebiRestClient.createWebiDocumentDataProviderSpecification(WebiRestClient.java:803)\r\n\tat com.sap.csina.cloud.boe.WebiDataProvider.UpdateDataProviderWithAllObject(WebiDataProvider.java:208)\r\n\tat com.sap.csina.cloud.boe.WebiUniverse.loadPrompts(WebiUniverse.java:182)\r\n\tat ...

• It errors out when SAC tries to establish the connection and analyze a very complex universe from BIP. In this case, there are more than 1500 objects in the universe.
• When you create a web intelligence and test a query with all objects from the same universe, it successfully returns the result in about 10 minutes.
• Note: This issue does not occur with not so complex universe. 

• SAP Analytics Cloud
• Direct type Live Universe Connection (LUC) 2.x
• SAP BusinessObjects Business Intelligence Platform 4.2 SP4 or later
• UNX universe

1. Log on SAP Analytics Cloud tenant
2. Create Direct type Live Universe Connection
3. Create a Model and select the LUC connection created in previous step
4. In the Data source dropdown list, select a very complex universe, to create a model.
5. The above SocketTimeoutException error is thrown.

The data model underlying SAC is a multi-dimensional cube, that is a set of homogeneous measures all dependent on a (limited) number of Dimensions. This model as such can’t accommodate the complexity of large Universes with multiple fact tables and contexts.

On one hand, the SAC UNX is designed to accommodate relatively small sets of measures and dimensions; and more importantly, it can’t handle incompatible objects like measures that depend on different dimensions. For this reason, one can’t directly expose a Universe that has contexts.

This works well for small “datamart-like” universes, but this is indeed a problem for large, complex “datawarehouse-like” universes meant to capture a large domain of corporate data. As a guideline, we should limit the maximum number of objects in LUC to 100.

Workaround:

As of now, we recommend customers creating multiple derived universes on top of such universes, so as to expose a set of “datamart-like” universes each dedicated to a family of business questions. These universes are then easily mapped into SAC models.

The list of limitation on SAC consuming universe is published here:
https://help.sap.com/viewer/c1834e734a0348768cc3ec0161f06223/4.2.4/en-US/c9632d24952543d690663afe8087f03f.html

BI, 4.x,  Cloud for Analytics, Cloud4Analytics, CloudforAnalytics, C4A, BOC, SAPBusinessObjectsCloud, BusinessObjectsCloud, BOBJ, BOBJcloud, BOCloud., BICloud, BO Cloud, connecting, universe, msu, unx, multiple, datafedarator, data , KBA , LOD-ANA-BI , Business Intelligence Functionality, Analytic Models , Problem