A new framework based on NVIDIA GPU and Bluefield DPU helps network security providers develop AI solutions that can instantly detect network vulnerabilities

NVIDIA Morpheus application framework can provide network security partners with a set of accelerated AI technologies that can detect and prevent security threats in real time.

Santa Clara, California – GTC Conference – April 12, 2021 Pacific Time – NVIDIA today released the NVIDIA Morpheus application framework, which can provide network security partners with a set of accelerated AI technologies that can detect and prevent security threats in real time.

As a cloud native network security framework, NVIDIA Morpheus can identify, capture and respond to previously unidentified threats and exceptions through machine learning, including the disclosure of unencrypted sensitive data, phishing attacks and malware. Deploying Morpheus for secure applications can give full play to NVIDIA AI computing and NVIDIA ® BlueField ®- 3. The advantages of DPU provide users with data center protection capability from core to edge.

Huangrenxun, founder and CEO of NVIDIA, said at the GTC conference: “The zero trust security mode requires us to monitor every operation of the data center in real time. This is a major technical challenge – we must immediately sense the intrusion and discover the threat in the server without affecting the data processing performance of the modern data center. NVIDIA Morpheus combines mellanox network and NVIDIA AI technology in the server to predict the threat through real-time inspection of complete data packets, so that the threat can be eliminated as soon as it appears The purpose of extinction. “

Morpheus combines with Bluefield DPU to make each computing node in the network become an edge network defense sensor. Enterprises do not need to copy data, but can also analyze each data packet at line speed. In contrast, traditional AI security tools can only sample about 5% of the network traffic data, so the threat detection algorithm is not based on a complete model.

Through the combination of edge and AI technologies, Morpheus can analyze more security data without sacrificing cost and performance by using real-time telemetry, policy execution and operation. Developers can also use the deep learning model to create their own Morpheus AI functions on the basis of existing IP investment.

AI enabled security ecosystem

NVIDIA cooperates closely with many leading hardware, software and network security solution providers to integrate various data center security products with NVIDIA Morpheus AI framework and optimize performance, such as aria cybersecurity solutions, cloudflare, F5, Fortinet and guardicore, which are well-known in the security industry; And hybrid cloud platform service providers such as canonical, red hat and VMware.

Adam Mishler, chief information security officer of best buy, said: “Protecting the complex and changing business environment requires continuous visibility. Providing real-time and dynamic network status helps to find places where we can further strengthen management, and further improve the anomaly detection capability based on machine learning. NVIDIA Morpheus framework provides us with a flexible and scalable anomaly detection platform in the ever-changing network threat environment.

Aaron Sant Miller, chief data scientist of Booz Allen Hamilton, said: “With the rapid expansion of the attack surface, the demand for advanced detection tools at the edge has become increasingly important. To meet this demand, we choose to cooperate with NVIDIA to jointly design, develop and deliver mobile cyber AI tools. Morpheus is the foundation of our cyber precog flyaway kit tool. The cyber precog flyaway Kit is an AI based network threat hunting platform that can package data center computing into a rugged small Type GPU platform. “

Tim Tully, senior vice president and chief technology officer of Splunk, said: “Splunk is very happy to cooperate with NVIDIA to explore more solutions to use GPU accelerated in-depth learning to better serve our common customers and turn data into productivity. We expect our team to quickly create prototype systems through Morpheus framework, integrate a variety of new functions on our platform, and unload computing intensive tasks onto GPU to help our customers.”

In addition, if Morpheus is running through NVIDIA Certified Systems ™ The performance is better on the systems of certified major server manufacturers around the world. Including ATOS, Dell technology, gigabyte, Xinhua III, HPE, Langchao, Lenovo, QCT and ultramicro.


Network technology and network security developers, software partners, start-ups and computer manufacturers can now apply for a preemptive trial of NVIDIA Morpheus platform.

Leave a Reply

Your email address will not be published.