Skip to content

Latest commit

 

History

History
62 lines (61 loc) · 3.4 KB

README.md

File metadata and controls

62 lines (61 loc) · 3.4 KB

CDR Aggregation Demo

The idea of the demo is to introduce the idea of CDR post-processing of CDR streams coming out of Call servers. We would create a Pipeline that reads from a Map Journal, applies stateful transformation and then pushes an aggregated CDR into another Map.

Scenario: A calling number is dialing the called number but the call server is actually connecting the caller to a real destination number. In the first three attempts the destination numbers are busy but the 4th attempt answers and call server connects the calling number to this 4th destination. The final CDR sent to mediation should be the one connected to this 4th destination number but the final cdr should mention other three destinations as well. img.png

Deployment

  • Start one or more Hazelcast members hz start and also Management Center hz-mc start
  • Compile and package ./gradlew build the code which has the Pipeline
  • Deploy the jar to the cluster hz-cli submit --class com.hazelcast.examples.cdr.CDRRead build/libs/cdrstreaming-1.0-SNAPSHOT.jar

Demo steps

You can demo the SQL in Management Center or any other SQL client. I used DBeaver.

  • Create the Mapping for the Maps
CREATE OR REPLACE MAPPING "cdr" (
  __key VARCHAR,
  "ID" VARCHAR EXTERNAL NAME "this.ID",
  "calledNumber" VARCHAR EXTERNAL NAME "this.calledNumber",
  "callingNumber" VARCHAR EXTERNAL NAME "this.callingNumber",
  "destinationNumber" VARCHAR EXTERNAL NAME "this.destinationNumber",
  "route" VARCHAR EXTERNAL NAME "this.route",
  "status" VARCHAR EXTERNAL NAME "this.status",
  "time" VARCHAR EXTERNAL NAME "this.time" )
TYPE IMap OPTIONS ( 'keyFormat' = 'java',
  'keyJavaClass' = 'java.lang.String',
  'valueFormat' = 'compact',
  'valueCompactTypeName' = 'com.hazelcast.examples.cdr.CDR' );
---
CREATE OR REPLACE MAPPING "cdr-final" (
  __key VARCHAR,
  "ID" VARCHAR EXTERNAL NAME "this.ID",
  "calledNumber" VARCHAR EXTERNAL NAME "this.calledNumber",
  "callingNumber" VARCHAR EXTERNAL NAME "this.callingNumber",
  "destinationNumber" VARCHAR EXTERNAL NAME "this.destinationNumber",
  "route" VARCHAR EXTERNAL NAME "this.route",
  "status" VARCHAR EXTERNAL NAME "this.status",
  "time" VARCHAR EXTERNAL NAME "this.time" )
TYPE IMap OPTIONS ( 'keyFormat' = 'java',
  'keyJavaClass' = 'java.lang.String',
  'valueFormat' = 'compact',
  'valueCompactTypeName' = 'com.hazelcast.examples.cdr.CDR' );
  • Now add three CDR to the cdr map using SQL. Each entry will trigger the Pipeline but it would not generate the final CDR as it is waiting for the ANSWER status CDR.
sink into cdr  values ('1', '1','447863499470', '449301857280', '447205189201', '', 'BUSY', '2020-10-10 10:10:00');
sink into cdr values ('2', '2','447863499470', '449301857280', '447205189202', '', 'BUSY', '2020-10-10 10:15:00');
sink into cdr values ('3', '3','447863499470', '449301857280', '447205189203', '', 'BUSY', '2020-10-10 10:20:00');
  • You can now see the records in the cdr map and also in the cdr-final map. The cdr-final map is empty.
SELECT * FROM cdr;
SELECT * FROM "cdr-final";
  • Now add the 4th CDR to the cdr map with status ANSWER. This will trigger the Pipeline and generate the final CDR.
sink into cdr values ('4', '4','447863499470', '449301857280', '447205189204', '', 'ANSWERED', '2020-10-10 10:25:00');
  • Now when you query the cdr-final map you will see the final CDR with all BUSY destinations as route.