With the rapid advancement of emerging technologies like 5G, IoT, and AI, the demand for agile, scalable, and efficient software development has never been greater.
As Korea’s pioneering mobile and internet provider and a key subsidiary of LG Corporation, LG U+ has built a reputation for excellence in mobile, home, and corporate customer services. Recently, LG U+ expanded its offerings to include AI services, implementing a business-to-business AI strategy focused on infrastructure, data, and platforms to drive customer growth.
At MongoDB.local Seoul 2024, Boeun Jin, a Software Engineer from LG U+’s Cloud Platform Development Team, showcased the company’s innovation journey with MongoDB Atlas.
To streamline internal processes and to modernize its applications, LG U+ developed the Uplus Cloud Management Platform (UCMP). This internal developer platform is designed to manage cloud and service environments, monitor security risks, and ensure smooth service deployment.
As public cloud adoption in South Korea surged, LG U+ recognized the need to enhance its infrastructure’s security capabilities to handle the vast amounts of data processed through its cloud services. To address this, LG U+ moved to MongoDB Atlas, the multi-cloud developer data platform powered by MongoDB's flexible document data model.
Before deploying MongoDB Atlas, LG U+ relied on relational databases (RDBMS). However, as the volume and complexity of the unstructured data in LG U+’s cloud environments grew, the company’s previous RDBMS solution proved increasingly inadequate. For example, external systems like Prowler, used for making data security scans, required constant schema updates to keep pace with the evolving structure and format of new releases, a process that was very time-consuming.
For infrastructure security, LG U+ managed scan results from 350 AWS accounts, each generating up to 2,500 results per vulnerability category. With scan results retained for three months, this resulted in the management of 10 million scans.
“A relational database would struggle with these changes efficiently,” said Boeun Jin. “We would spend considerable time and resources addressing schema updates every time the data format changed, due to the rigid structure of relational databases. In contrast, MongoDB's document model is ideal for managing unstructured data, as it allows for direct storage and retrieval without the need for schema updates. Its flexibility in handling complex and hierarchical document structures, along with its scalability to manage large volumes of data without compromising performance, were key factors in our decision.”
One of the most notable improvements LG U+ noticed after adopting MongoDB Atlas was related to the company’s data query management. The team used MongoDB's aggregation pipeline to handle various millisecond real-time APIs for infrastructure security to efficiently filter, group, and process large data sets.
For example, queries that aggregated security scan results by vulnerability initially took over 13 seconds to process across 870,000 data records, due to inefficient grouping and aggregation operations. With MongoDB Atlas’s aggregation pipeline, LG U+ saw a 99.11% reduction in query execution time. This was achieved by optimizing field separation and tuning queries according to their relevance to the aggregation method.
“Although new to MongoDB, our team successfully launched UCMP's infrastructure security capabilities within just three months. MongoDB’s low learning curve allowed us to customize MongoDB Repository and Mongo Template to our internal system environment and quickly write complex queries,” explained Boeun Jin.
“Most importantly, the ease of configuring the initial infrastructure enabled us to bring it to production swiftly. By creating replica sets of UCMP within the MongoDB cluster, we ensure high availability and reliability, and that the system auto-scales with CPU usage,” she added.
Looking ahead, LG U+ plans to further explore MongoDB Atlas’s features to streamline costs, optimize queries, and enhance data management.
“For cost efficiency, we aim to leverage MongoDB Atlas’ functions and triggers to manage traffic spikes and scheduled batches effectively. In addition, we expect to optimize our database by monitoring query execution time and referring to index usage recommendations from the Performance Advisor, which provides insights to improve query performance,” said Boeun Jin.
Head over to our product page to learn more about MongoDB Atlas.