Juniper Networks, a network equipment and system vendor that began to be acquired by HPE in January, launched its new AI-native networking platform this week, with the goal of comprehensively improving operations and end-user experience with artificial intelligence, while reducing network operation and maintenance costs, and ensuring that the network connections of devices, users and applications are reliable and secure. The AI-native networking platform's AI-native network digital experience twin, Marvis Minis, can use artificial intelligence to proactively discover and automatically fix problems.
Juniper Networks designed the AI-native Networking Platform from the ground up to connect campus, branch, and data center networking solutions, a general AI engine, and Marvis Virtual Network Assistant (VNA) for deep analytics, automated troubleshooting, and seamless end-to-end network assurance. Officials mentioned that compared with traditional solutions, the AI-native networking platform can reduce operating expenses by up to 85%, eliminate 90% of network trouble tickets, and reduce network incident resolution time by 50%.
In the AI-native networking platform, Juniper Networks has introduced 2 enhancements that make AI available for IT operations to the point where it is possible and reliable. Marvis is an AI-native VNA developed by Juniper Networks using Mist AI, with functions such as proactive recommendation, automated operation, and conversational interface.
The new feature, Marvis Minis, is an AI-native network digital experience twin that proactively simulates end-user traffic and uses unsupervised machine learning to instantly validate network configurations, enabling early detection and automatic remediation of issues before they impact users, applications, or devices. Marvis Minis is also integrated with the Mist AI engine to provide Marvis with a new source of data.
The Marvis VNA, a network virtual assistant, is now also available in the data center, running on any vendor's hardware to provide data center operations analytics. The assistant identifies and asks questions about data center cabling, configuration, and connectivity, and the Marvis dialog interface allows IT teams to quickly and easily access data center product documentation and knowledge base queries.
In addition, Juniper Networks announced enhancements to its AI data center solutions to enable users to deploy AI training and inference clusters faster and more flexibly. The solution is based on the Spine-Leaf data center architecture, based on QFX switches and PTX routers, and operated by Juniper Apstra, Juniper Networks' network automation and operations system, plus new features such as the new Marvis VNA, which dramatically reduces the complexity of network design, deployment, and troubleshooting for AI data centers.
The company is also expanding its network automation processing system, Juniper Apstra, to enhance its ability to handle AI traffic over the Ethernet network, while the new PTX router offers high-density 800GE capacity, and the new QFX switch has twice the capacity of the previous generation. Both PTX and QFX platforms support up to 800GE connection port density, and the necessary AI protocols.