User Tools

Site Tools


public:smart_big_data_eng.ver

This is an old revision of the document!


A Highly Wired World and an Increasing Demand for Big Data Analytics

The world is wired with ICT systems that are connected through networks, and it is becoming more difficult to operate the systems in a stable manner while maintaining high quality performance. Smartphones pervade every aspect of our life and communicate seamlessly with ICT systems while being connected to the telecommunications networks. Smartphones are now connecting users even to the other side of the earth. Going beyond that, we are faced with the era of the Internet of Things, which makes it even more difficult to operate ICT systems.

Furthermore, the way we utilize the ICT systems has also changed. Cloud storage in which data is maintained and managed remotely and made available to users over a network is mostly used these days. Along with the technology trends that promote the overall virtualization of servers and networks to realize an effective, large-scale cloud center, the complexity and difficulty that faces the ICT operating team has grown exponentially.

Mobigen has developed and provided high-performance, large-scale systems to help operate and manage high-volume ICT systems. In an effort to solve the issue of an effective operation of massive ICT systems closely related with the IoT, cloud storage, and virtualization, our R&D center aims to come up with technological breakthroughs in the following areas.


Big Data Database Cluster

Massive ICT systems produce a vast amount of Big Data. There is a wide variety of ICT systems that produce a huge amount of Big Data including high-speed telecommunications networks, smart factory systems, smart grid electric power systems, smart buildings, smart u-City, smart u-Health, and intelligent road networks. As the number of devices wired to the ICT systems grows in the era of the IoT, the data volume will also increase geometrically.

Although the price of storage media and the cost of network bandwidth are constantly going down, the range of data growth is much higher than that, and therefore, it is a real challenge to manage the growing Big Data in the first place. In order to manage massive Big Data which cannot be handled by the existing database management systems, it is critical to establish distributed database management systems and continue to develop technologies required to process high-volume data that is possible to manage only with an architecture specially designed for real-time data processing. Due to overwhelming data volume, ICT system managers find it difficult to process even very basic data. There are various Big Data platforms, such as Hadoop, but system managers find them even more complicated and far from useful. What they need the most is a tool which can help them handle a large amount of Big Data simply, as it is not easy to manage basic functions such as storage, statistics, and search, in a distributed Big Data environment. Our IRIS Big Data database platform is a solution which helps system managers with such difficulties. The IRIS Big Data database platform ensures convenience, high performance and capacity, durability and simplicity, and it will serve as an essential platform for solving the issues of ever-growing data volume and performance management.


Big Data Log Analyzer

Large-scale ICT system operation requires log collection and analysis. In particular, system components are becoming more complicated as multiple ICT systems are connected through networks and virtualized servers, virtualized networks(Software Defined Networks) and virtualized data centers(Software Defined Data Centers) are becoming more common.

In order for system administrators to understand and manage the systems, there needs to be a foundation for collecting logs forwarded by all components, storing the logs and analyzing them in a single system. While managing the logs is somewhat similar to handling documents and texts in general, it is different from general text indexing and search in that it requires immediate index creation after data entry as well as real-time statistics production. Mobigen’s distributed Log Analyzer is an optimal system which can process Big Data logs collected during the large-scale ICT systems operation with focus on real-time indexing. There is an ever increasing need for large-scale, distributed log processing as ICT operation management systems along with various security monitoring systems, factory automation systems, and logistics systems are becoming larger and inter-connected.


The Need for an Intelligent Expert System

To realize a more effective operation of the massive ICT systems, we need to have a full understanding of the systems, discover systems errors, and come up with solutions to the problems from the perspective of a systems expert. As the systems are getting more complicated these days, it is difficult to find an expert who can deal with entire systems. Therefore, the realistic alternative is to have an intelligent analysis system assist the analysis process in an automated manner, so that the burden of systems administrators can be lightened.

An automated expert system that is based on Big Data enables Big Data analysis in a wide variety of areas. These include Anomaly Detection, Time Series Prediction through time-series data analysis, and Root Cause Analysis through an analysis on the correlation among diverse variables, and this helps the large-scale ICT system administrators manage the systems in a much easier way. Mobigen has continuously conducted research on the Advanced Analytics method and provided insights on how to have a comprehensive sense of judgment about the overall systems for the ICT administrators by applying internal research results to the data produced during real work-site operations.


A Need for Service Quality Management Focusing on Customer Experience

With the advance of the recent communications network environment into the All-IP data service, the network structure is becoming more complicated and expectations for network operation efficiency and service competitiveness are growing higher than ever. In addition, corporate expenditure and budget losses for network malfunction continues to grow exponentially. Network operators endeavor to detect problems immediately in order to prevent network error or poor service quality, and to figure out the root cause of problems in a prompt manner. They do this by utilizing various types of data produced in the network in a bid to maintain entire network stability and improve service quality. In particular, they conduct real-time monitoring of end-to-end quality and check on effective uses of network resources to respond promptly and appropriately to customer complaints concerning service quality. Furthermore, they concentrate all their energies on strengthening service competitiveness by detecting potential network problems ahead of time and promoting “preemptive and proactive service management.” However, they still face challenges in managing networks due to the many problems of the real network operation environment.


Problems in the Network Operation

The Need for Network Management from the Customer Perspective

Each network equipment produces periodic data from its own standpoint. When the network is operated based on the data provided by the given equipment, there can be a large gap between the given data and the service quality that customers experience, and it takes a long time to figure out the root cause of problems, which in turn leads to complaints from customers. Moreover, as the service quality that each customer experiences varies, there is a growing need for managing the service quality individually to improve service competitiveness.

Difficulty in High-Volume, Real-Time Data Processing

The network operation systems process a wide range of data, such as real-time fault data and statistics data provided by each piece of equipment, and various log data, which show the quality of service provided to individual customers. In particular these days, the systems collect and process packets, which flow within the network to figure out the quality of the All-IP network service that individual customers experience. These systems should be operated based on a Big Data processing platform verified on a high-volume network as they need to perform a real-time processing of billions of records of network data each day depending on the network scale and the number of customers.

The Inefficient Utilization of Network Data

Network data, particularly traffic data is used to operate the network effectively. For instance, it is utilized as a basic information to provide a stable service through controlling heavy traffic users and applications. Moreover, it can be used for the estimation of the capacity of each equipment in the network and also for the establishment of marketing strategies based on the traffic patterns of customers. However, most of the operations support systems were solely designed to facilitate the network operation and thus, either a new system needs to be introduced or existing systems should be reformed in line with the needs of other departments, such as network engineering, customer care center or marketing. As such, the systems used by network operators should provide basic data that is multi-purpose and fit for use by many departments.


Promoting Services that Focus on Customer Experience

Mobigen offers solutions to support various operational tasks required for the improvement of communications service quality. We also contribute to securing service quality competitiveness and improving operational efficiency by putting ourselves in our customers??shoes. Our operations support systems process, store, and analyze various types of information collected from a wide range of wired and wireless communications networks and pieces of service equipment. Mobigen solution provide various functionality to detect errors and analyze problems occurred at each layer including individual subscribers, devices, networks, and services. In particular, we provide an Advanced Analytics environment based on our Big Data platform in which it is possible to process a vast amount of network data in real-time and realize Preemptive Monitoring, Anomaly Detection, and Root Cause Analysis.

public/smart_big_data_eng.ver.1554686398.txt.gz · Last modified: 2019/04/08 10:19 by jhnam

Backlinks to this page
  • public:모비젠-홈페이지-콘텐츠