.svg)
Share
Copied!
As digitalization across the industry accelerates, the stable operation of thousands of servers and hundreds of service instances has become an inevitable challenge. In particular, for telecommunications companies that need to provide real-time services on a national basis, uninterrupted operation and quick response to failures are key factors in improving the customer experience.
In order to proactively respond to these demands, LG Uplus Complete overhaul of the existing monitoring systemand operating existing legacy systems and new systems Standardize hybrid infrastructure with a single standard for integrated managementI succeeded in doing it. In this example, we will introduce in detail the background and results of the introduction of WhaTap (WhaTap), which enabled connections between various systems, along with the infrastructure operation know-how that LG Uplus has accumulated over many years.
LG Uplus is one of Korea's leading telecommunications companies with over 70 million subscribers and provides various services such as wired and wireless communication, IPTV, smart home, and enterprise solutions. In a market environment where digital transformation is accelerating, LG Uplus is within its data center Operate an on-premise environment and an AWS-based public cloud togetherFurthermore, we are continuing to promote technological innovation to improve service quality and stable IT operations.
LG Uplus processes an average of 100 million transactions per day on more than 1,000 VM and container-based systems, and is one of Korea's leading enterprise examples that operate large-scale traffic and advanced distributed architectures.
LG Uplus's customer management system is based on a monolithic architecture built 10 years ago, and has been operated on an integrated DB and UNIX server centered on Oracle RAC. However, as the number of subscribers surged and the environment had to process more than 100 million transactions per day, existing infrastructure and software structures faced limitations in terms of scalability and agility.
The integrated DB reached its scale-up limit, existing UNIX servers were less effective than the expansion, and Query-centered script-based software had problems with both application complexity and development speed because it was difficult to analyze the impact of changes. As a result, LG Uplus Targets business agility, availability, scalability, and operational cost reductionI decided to build a new system with
The new system was designed based on a microservices architecture (MSA) and hybrid cloud, and transitioned to a complex distributed structure containing various open source components and external interlocking systems. However, these structural changes soon led to new operational challenges. Due to the rapid increase in monitoring targets, there is no need to consider it in traditional monolithic environments Complex monitoring requirements such as distributed tracking, integrated log collection, and cross-component impact analysisThis is a new thing.
In order to solve increasingly complex monitoring requirements, LG Uplus promoted the introduction of a solution that can unify existing application, database, and server infrastructure monitoring tasks, and chose WhaTap Monitoring (WhaTap Monitoring). In particular, Distributed tracking function that can track call flows throughout the system by linking transaction IDsThis was a key deciding factor.
With a global transaction ID and its own MTID (Multi Transaction ID), Watap helped track call flows on a consistent basis across both legacy and new systems. At the same time, asynchronous services, such as relay processing based on KAFKA messaging, have also made it possible to link transactions through Weaving technology. In particular, the ability to insert and track transaction IDs without code modifications was highly appreciated even for legacy systems.
The hybrid cloud structure is also an area where WaTAP was able to respond naturallyThis is it. In consideration of public cloud traffic costs, collection servers were distributed within the cloud, and collection servers were arranged in departments such as architecture teams, IPTV, and authentication services in private environments to simultaneously achieve independence and integrated management.
Also, Various log collection environments are also included in the corresponding categoryIt's done. Infrastructure logs are an Elasticsearch - Fluentd - Kibana (EFK) stack, and application logs are collected in a tap without code modifications to the framework, making it possible to flexibly operate log strategies without program changes.
After the introduction of Watap, LG Uplus 'A distributed processing flow that I couldn't see began to be visible'I rated it. Even without a global transaction ID, it was possible to visualize the entire call flow using MTID automatically issued by Watap, and call relationships and bottlenecks between services can be identified in real time. As soon as an alarm occurred, the operations team moved to the appropriate location to analyze resources, transactions, and logs in conjunction with each other, greatly improving the speed of incident response.
Furthermore, the security of access rights management, such as opening a firewall, has been enhanced through a single listener server structure, and it is operated as an enterprise-wide standardized tool, such as requiring the installation of Watap, an integrated monitoring tool when applying new services.
LG Uplus Introduction of application performance management (APM) in 2018Starting with Gradually expand the scope of advanced monitoring to server infrastructure monitoring (SMS), database monitoring (DPM), and KubernetesI've done it, and now it's being offered by Watap Consider introducing various types of monitoring such as log monitoring, browser monitoring, and NMS It's in progress. This is an example of how sufficient scalability and flexibility can be secured even with a single tool (platform) called WhTap Monitoring without introducing and operating various siloed monitoring tools for each system.
Furthermore, the Weaving function of WATAP was applied to asynchronous systems based on Netty and Akka to implement monitoring that can be tracked without modifying the source code, and the monitoring server is also managed stably in on-premise and cloud environments by operating integrated monitoring servers. This is assessed as an example of unifying monitoring standards and effectively reducing operational complexity across the enterprise.
LG Uplus is business-specific through the introduction of Watap Consolidate distributed IT monitoring environments across the enterprise, and greatly improve fault response and operational visibilityI did it. By standardizing a transaction-based tracking system, we have established a structure that can quickly identify and respond to the cause of failure, and we have also secured flexible scalability across on-premises and cloud.
View the entire system flow from a single dashboardDo it, Track call relationships between services in contextwhile doing, Reduce unnecessary costs by simplifying the license structureOne example provides a great best practice for enterprise companies running complex infrastructures.
If DevOps and IT infrastructure organizations are considering a reliable operating system, I recommend that you consider how to build an integrated and flexible monitoring environment through Watab like LG Uplus.
What about companies that want to develop a monitoring system like LG Uplus and ultimately establish integrated monitoring? 👉Inquire about implementation
