Business Problem

Our customer for this chapter is a telecommunications carrier, with millions of subscribers, and multiple interconnections to other networks and carriers to provide worldwide access for its subscribers. They handle millions of calls each day on their network. One of the challenges they constantly face is understanding their customer's calling patterns so that they can plan and optimize their network capacity. They analyze network usage based on time of day and the origin and termination locations of each call. The marketing group also reviews how the various calling plans they offer are being used.

Problem Statement

Many of the business problems our customer is experiencing would benefit from the kinds of BI capabilities that you have seen in previous chapters, but there is one huge technical problem that first needs to be overcomethe volume of data in a utility company's data warehouse can be staggering. The client has reduced the volume of data by loading filtered and summarized data into relational data mart databases for specific business applications. The business still finds itself constrained in many ways:

• The summarized level of information leads to limitations in the kinds of queries that can be answered meaningfully, because some of the dimension attributes are no longer available. For example, data summarized to the month level cannot be used to predict daily traffic levels.

• Data points from multiple systems are summarized and consolidated separately, leading to silos of information in the various data marts.

• The users cannot get timely information on events such as the effect of promotions on network traffic. As the database grows in size, increasing maintenance time is reducing availability of the system.

• Queries that span or summarize long time periods or many geographic areas are becoming too slow as the amount of data increases.