(Jan 2001 - Jan 2002)
Web Reporting System
and deployed a major upgrade to the streaming media traffic reporting
system for this large Content Delivery Network (CDN).
On average, over 10 million records were processed daily.
This system provided statistical summaries of streaming media activity
on the iBEAM network. We collected raw data from multiple points
on the streaming network and generated predefined reports summarizing
the activity over a variety of time periods (daily, weekly, monthly,
was used by iBEAM to generate customer reports, provide network
usage for customer billing, and provide diagnostic information on
the operation and efficiency of the network. More:
Acquisition - The system acquired the data daily, collected and
consolidated from hundreds of geographically dispersed points,
incorporating large data centers (New York, Oklahoma and California)
and installations at local ISP's.
and Cleansing - Raw data was parsed, transformed and loaded into
the database, continually. During this process, any records that did
not meet the minimum criteria for completeness were rejected and flagged
for further analysis.
- Data records were further reorganized into staging tables to aid
in efficient summarization. The summarization processes generated
data cubes with dimensions time, media type, streaming rate
- Reports were customer-accessible over the web, following a standard
format for all users. Custom reports were created on demand, taking
further advantage of the multiple dimensions and detailed data.
For TeleGlobe Canada (contract via SelectStar, 1998 - 2000)
Traffic Analysis and Reporting System (Phone Switch Data)
Using iBuilder, an OLAP application platform, we developed a full
production query and reporting system that provided web-based analysis
of TeleGlobe customer phone usage, updated on a daily basis.
were able to analyze both their usage and costs geographically, organizationally
and historically. The end-user application also provided ad hoc query
and drill-down functionality.
Data Warehouse/OLAP system provided production bulk data processing
facilities: Raw data was collected on a daily basis and "cleansed"
prior to being inserted in the database. Old data was archived and
purged from the database on a customer-by-customer basis, according
to individual customer policy.