Telco Industry Solutions

Background of the Solution

The Internet has evolved from the PC desktop era to the current mobile Internet era, with more mobile devices coming into consumers’ view. This is an era that calls for innovations in human-computer interaction (HCI). Intelligent voice technology offers a touch-free interaction method, freeing users from the hassle of frequent touchscreen taps. By connecting most devices to a local area network (LAN), “one voice command, multiple responses” voice control can be achieved.

Communication users’ demand for rack-mounted GPU servers mainly arises from the need to build new AI platforms on top of their existing equipment environments. The core objective is to construct large-scale, systematic artificial intelligence (AI) clusters to meet the requirements of daily training and application inference, with a focus on exploring the construction of GPU server-based clusters and conducting research on cluster resource scheduling and management.

Overview of the Solution

ZRJ’s GPU servers are primarily designed for medium-to-high power density data centers and standard 19-inch racks, offering flexible procurement and deployment capabilities.

This solution adopts a 2U 4-GPU server, equipped with dual-port fiber network cards that support RoCE. It enhances device information processing bandwidth and reduces latency during data transmission, making it ideally suited for deep learning scenarios in artificial intelligence (AI). It provides superior adaptability and higher performance for various user applications.