The client is a Korea’s tier-1 mobile service provider that boasts approximately 30 million subscribers across the country. It actively responds to evolution of mobile communication network technologies such as 2G CDMA, CDMA 2000 1X, EV-DO, WCDMA, HSDPA networks, while offering nationwide mobile services. Since commercialization of 4G LTE network in 2011, the company has been playing a leading role in settling mobile broadband services through an early supply of nationwide network.
The client company used to develop and run a billing data analysis system on a basis of pricey data warehouse solutions provided by an overseas solution vendor. However, due to an increase in the number of 3G and 4G LTE subscribers together with an explosive data traffic, it saw an inflow of enormous data exceeding processing capacity that was designed at an early stage. The situation made the system less stable and available, requiring a need to extend processing capacity with expensive systems as soon as possible.
Instead of increasing capacity of the existing system at high cost, the client selected IRIS, a big data database solution for processing and analyzing massive data. With the introduction of IRIS, the client has significantly saved costs by replacing some of heavy-load processes that are currently done by the existing data warehouse system. Moreover, it could maximize system performance by transferring the following functions to IRIS for faster processing—storage/search of raw data (also known as primary data) and inquiry/analysis services. The major tasks carried out by IRIS in this project are as follows.
Compressing roughly 60 billion log data records per day and store them after the data replication Creating indexes and tables of data on a real-time basis Real-time search of three months’ amount of data (nearly 5 trillion records, 1.5PB)n
This project was to increase extra capacity at low cost within several months by adopting Mobigen's IRIS. The project was successful as it also improved operational efficiency of the existing system through effective sharing of roles between the existing system and IRIS.
Backlinks to this page