BeDataDriven is working with partner Altai Consulting to bring the web-scale technologies pioneered by Google and Yahoo! to mobile operators in emerging markets, who have often struggled to manage high volumes of data with general-purpose relational databases.
Mobile phone networks generate several gigabytes of usage logs each day that hold tremendous insight into subscriber behavior and feedback on marketing performance. Unlocking these insights, however, can be difficult with a traditional database system like Oracle or MySQL because the primary bottleneck is not computation power– but the speed at which data can be read from disk.
A cluster of on-site servers running Hadoop, an open-source implementation of the Google File System and MapReduce, and Hive, a data warehousing system developed by Facebook, can complement or replace an overwhelmed relational database for daily reporting requirements while enabling fast execution of complex, ad-hoc queries and data mining.
Opportunities in the Cloud
Cloud computing also offers opportunities to mobile phone operators. Services like Amazon EC2 and Google BigQuery (in preview) allow operators to move historical usage records to online storage, freeing up on-site resources, while still keeping this data available for data mining.
This storage is extremely cost effective: for one of our clients, storage costs for a year’s worth of call detail records, for their 5 million subs, totals $73.16/month on Amazon S3.
See our technical post on Map/Reduce for telecoms.