• Enterprise Search
  • Automotive & Manufacturing

POSCO: Moving to Elasticsearch for a More User-centric Enterprise Search Experience

AT A GLANCE

  • 40,000
    Enterprise Portal system users
  • 85
    million documents, with approximately 2 million updates and deletions
  • 53%
    faster than the previous solution despite managing a larger volume of data

A user-centric focus

By leveraging Elasticsearch, POSCO’s Enterprise Portal system has gained the flexibility to provide user-centric services.

A flexible product stack

The Elastic Stack’s various products and features help POSCO to respond to user demands.

Frequent upgrades

Periodic upgrades allow POSCO to meet new demands from users.

Products in use


Company introduction

POSCO Group’s Enterprise Portal (EP) system is a business management system and corporate portal that acts as a gateway for accessing the internet and POSCO’s company intranet, where employees can access company-specific documents and applications. The EP Operations team is responsible for the construction and operation of the system’s search functionality and provides services to 40,000 employees of POSCO and their subsidiaries (about 20,000 POSCO employees and 20,000 employees of POSCO subsidiaries).

The EP Operations team within POSCO’s IT Business Department found that replacing their legacy search solution with Elasticsearch had a positive impact on users’ satisfaction. The team is extremely pleased with the Elastic Stack’s various products and features, which have allowed them to manage and stabilize large amounts of data. The Operations team plans to continue to leverage even more aspects of the Elastic Stack in future projects, such as developing a system that detects server failure and a search system that delivers personalized results to users.

POSCO’S JOURNEY WITH ELASTIC

Making the case for Elasticsearch

To meet the diverse needs of users, Project Manager Jin-ha Jung and his team participated regularly in Elasticsearch Meetups in Seoul to learn more about Elasticsearch and to identify customizable search solutions that were suitable for POSCO’s system. After finding out more about open source software, Elastic’s products, and their practicality during the meetups, the team decided to replace their existing Google Search Appliance (GSA)-based search engine with Elasticsearch.

"I think it was a good decision to make the change, since I became hopeful that we could make enhancements for users after trying Elasticsearch."

Eui-young Hwang, EP Operations team, IT Business Department | POSCO

Responding promptly to various user demands

The EP system’s existing GSA-based search engine was unable to meet the diverse needs of users. Therefore, the EP Operations team within POSCO’s IT Business Department decided to make the switch to the Elastic Stack. Their new Elasticsearch-powered search functionality is now in operation after a pilot program and is able to quickly respond to users’ various needs.

Furthermore, in addition to stabilizing system operation, Elasticsearch is improving the team’s technical and operational capabilities, enabling them to provide more advanced user-oriented improvements.

Controlling and stabilizing various types of data

The EP Operations team uses specialized functionalities within Elasticsearch, such as aggregations, to meet the diverse search requirements of several linked systems.

To secure their indexed data, the team is using the Elastic security features included with the Platinum subscription to secure and stabilize massive amounts of data.

Taking advantage of the Elastic Stack's scalability, periodic version upgrades and version management, and flexibility in distributed environments

Through the various products that make up the Elastic Stack, such as Logstash and Beats, the EP Operations team has been able to provide more services, more easily, for their users. Furthermore, periodic upgrades to Elasticsearch and other products within the Elastic Stack have allowed the team to keep up with changing user demands.

How the EP Operations team uses the Elastic Stack

  • Elasticsearch: main search engine
  • Kibana: user interface and data stability monitoring
  • Elastic security features: authentication of users and data to protect indexed data
  • Logstash and Beats: collecting log file data for system failure protection purposes

Achieving positive results by replacing a legacy search engine

The EP Operations team conducted a proof of concept with Elasticsearch for four months, and it took about a year until the system was in service, not counting the time spent for the proof of concept. Most of that time was spent extracting and transferring the huge amount of data.

Now, with Elasticsearch under the hood, the EP system’s search functionality produces highly accurate results in Korean and stably and rapidly processes large amounts of data. Documents are available for searching immediately after being indexed, and users are highly satisfied with the system’s processing speed.

"Users are most satisfied with the speed of the system, which allows documents to be searchable almost immediately after they have been indexed."

Jong-chul Kim, Manager, EP Operations team, IT Business Department | POSCO

The EP Operations team within POSCO’s IT Business Department is planning various projects, such as collecting log data using the Elastic Stack’s machine learning features and Kibana to build a failure prevention system that detects server failure in advance. They also plan to construct a customized search system that yields individualized search results filtered based on the user’s preferences. The various products and features within the Elastic Stack will help POSCO to implement these projects.