CDR (or XDR generated by probes) is the valuable data source in telecom operator to analyze the behavior of their subscribers and can be converted to the market intelligence for monetization.

As Ellipsiz Communications is the CDR/XDR expert with past 15-year experience in telecom industry, the business critical requirement we learned from our telco customers is to have the solutions to process the really huge volume of CDR/XDR either for fast search or for intelligence analysis.

To respond customer’s requirement, Ellipsiz Communications has developed two solutions for CDR/XDR processing with modern big data technologies.

Billions of CDR/XDR/Log records are generated from the telecom network every day, no one record can be discarded and all of them have to be stored for future use, such as troubleshooting for network issue or subscriber’s complaint, billing verification, subscriber filtering for marketing event, authority audit, and so on.

Therefore, the tradition database approach is not suitable but a fast full text search engine is required in order to derive the result from huge volume of CDR/XDR/Log records in second.

We developed the solution based on Lucene technology and the system can be expanded with network traffic growth or longer data storage time required.

Location is the key information in telecom CDR/XDR and more intelligence and applications can be derived from subscriber’s mobility.

Unlike other location based systems or applications, the solution we proposed is to predict the subscriber’s next location in near future and the regions he stays or visits more often (so-called habitat).

The location prediction can be extensively adopted in following domains:

####Application samples of Next location prediction:

• Pre-alerting when subscriber is going to access the bad radio cell.
• Pre-arrangement of network resources when crowd is predicted somewhere
• Taxi driving suggestion when crowd is predicted somewhere.
• Location based advertising to attract the consumer before he arrives or change his original mind to the new destination.
• Abnormal route alerting when the elder or children goes the different route as usual.
• Others

####Application samples of Habitat prediction:

• Location based adverting to the residents of the region or the visitors who come the region more often.
• Pre-alerting to the residents of the region or the visitors who come the region more often before some big event will be happened in the region.
• National security
• Others

We developed the solution based on the big data technologies including Hadoop, Spark, machine learning, and so on.