Network security in modern datacenters is primarily focused on the inbound/outbound packet flow, often referred to as north-south traffic. However, the growth of cloud-native applications has driven an explosion of east-west network traffic within a datacenter where applications can create hundreds of thousands of network connections among virtual machines and containers. As a consequence the ability to track, monitor and secure a datacenter in a timely manner has risen above that of any individual or team. To combat this challenge, Red Hat and NVIDIA are working together to protect networks from breaches via real-time AI security analysis.
NVIDIA’s Morpheus AI application framework is designed to handle a variety of complex security tasks and policies allowing users to develop and deploy AI-enabled security applications efficiently. Morpheus AI provides several pre-trained models, including one with the ability to immediately recognize many types of sensitive personal information, like public cloud or GitHub user credentials, private keys, passwords, and credit card numbers. This pre-trained model enables AI to search through network packets for patterns associated with these credentials and flag exposed data to the enterprise security team.
Morpheus AI framework represents a notable advance towards the concept of a ‘self-driving datacenter’ where applying AI and machine learning to IT operations, infrastructure security, development, and DevSecOps can help manage the growing complexity of modern cloud-native applications and the underlying platforms. We also see this as a key area of emerging innovation where we are engaging the broader ecosystem and Open Source community.
Red Hat recognizes the need to develop advanced solutions for network security and automation. Through the Red Hat Developer Program developers can use free Red Hat Developer subscriptions to run the Morpheus AI framework. The Red Hat Developer subscriptions include Red Hat Enterprise Linux, as well as related management tools.
Security applications developed with Morpheus AI and running on Red Hat Enterprise Linux can provide a new level of security visibility into critical data in packets. This can empower organizations to automatically identify and enforce stronger security policies regarding particular data, helping to reduce the potential and impact of a data breach.
Additionally, customers deploying this framework with NVIDIA BlueField-2 data processing units (DPUs) will gain unmatched visibility into identifying sensitive data moving across their datacenters and the enterprise. DPUs are designed to accelerate software-defined datacenter services that would otherwise require a fleet of expensive hardware appliances, and represent a broader shift in system architecture from a PC-centric design to a cloud native, service-centric, composable compute model. Network security is just one area where DPUs provide the ability to implement critical datacenter services by offloading complex tasks from the server to the domain-specific piece of hardware with software defined compute capabilities.
To take advantage of the rich set of features available with BlueField DPUs, users need the appropriate support in enterprise software that is consistent and compatible across different platforms. As an enterprise leader in open source software, Red Hat centers its support for BlueField-2 DPUs around enabling an open ecosystem with Red Hat Enterprise Linux and Red Hat OpenShift Container Platform. That ecosystem can be easily extended to BlueField-3, NVIDIA’s next generation of DPU devices.
At Red Hat, we aim to maximize collaboration, facilitate new product adoption, and create operational consistency across hybrid cloud environments by enabling an open ecosystem that accelerates innovation while providing access to the latest hardware innovations and helping our customers standardize their infrastructure.