Wilbert A.McClay
Northeastern University, USA
Title: Mobile data and individual client diagnostic acquisition from pgws call analysis utilizing spark and the hadoop ecosystem
Biography
Biography: Wilbert A.McClay
Abstract
Individual Client Diagnostics, is an account specific event log to diagnose customer experience from PGWS call analysis. The meaning of a PGW call is to perform policy enforcement, packet filtering for each DirecTV user, charging support, lawful interception and packet screening. Another key role of the PGW is to act as the anchor for mobility between 3GPP and non-3GPP technologies such as WiMAX and 3GPP2 (CDMA 1X and EvDO). Thus, the acquisition of the PGW calls within the Kafka API demonstrates the objective of the DirecTV Mobile Diagnostics project. The Spark Dstream is connected to a HBase/Websocket to IMAP with a projection into HBase. The big data analytics and diagnostics component will rely heavily upon the aggregation windows return, which prescribes how IMAP wants the PGWS call data lucidly defined. The Spark analytics machine learning analysis post Kafka API inspection of key-value pair analysis from IMAP PGWS Request and Response Data demonstrate the following statistical results utilizing Spark machine learning library (MLlib), SparkR and SparkSQL, projected into HBase as a column family architecture.