Customer Overview

This customer is an online radio and TV broadcasting organization of an international network company. Its network TV station fully deploys multi-terminal business architecture and has built five integrated broadcast and control platforms: network TV, IP TV, mobile TV, mobile TV, and Internet TV. The company has covered Internet users in more than 190 countries and regions around the world by deploying global mirror sites and launched six foreign language channels in English, Spanish, French, Arabic, Russian and Korean, as well as five ethnic minority language channels in Mongolian, Tibetan, Uygur, Kazak, and Korean. The company has established a well-known network video company with an all-media, full-coverage dissemination system of network audio-visual public service platform.

Customer Demands

The database of the video library business system of this company is based on a MySQL Master-Multislave architecture. With the growth of the customer base and the expansion of product lines, the volume of business data has increased dramatically and the time of business logic execution has increased significantly. Facing the current situation, the customer has the following requirements:
  • 1.Improve database performance and user experience.
  • 2.Improve the hardware utilization rate. Due to the limited budget expenditure of the department, the customer wants to reduce the operating cost of the server room by improving the performance of individual machines.
  • Analysis of the application problems encountered by the customer

    The current system operating pressure is becoming more and more obvious. The main performance of the system is as follows:
  • There are database performance problems. After relevant monitoring and analysis comparison, we found that the storage device provides less than 3000 IOPS, and more than 40% of the CPU time is IO Wait. That is, the CPU spends a lot of time in an idle state, waiting for the response of the storage system.
  • The existence of performance bottlenecks in the storage system leads to a long response time of upper-layer applications, poor end-user experience, and frequent timeouts.
  • Solution Analysis

    Summary

    Deploying the Shannon Direct-IO® PCIe Flash solution brought the following improvements to the video library operations of this well-known online video company:
  • Completely eliminates high latency in data query and data update services, creating smoother user experiences.
  • Provides more stable data processing capabilities.
  • Eliminates database slow queries.
  • Improves database read and write performance by 8 times.
  • Increases CPU utilization by 40%.
  • Further reduces energy consumption and cooling requirements, improves infrastructure deployment density, and cuts down comprehensive costs.