DDN BLOG

How DDN EXAScaler and NVIDIA BlueField Keep Data Safe for At-Scale AI

AI is allowing businesses to analyze and extract value from data in ways that were never possible before. While AI makes this data more valuable, it is also a target for potential cyber breaches.

Trying to secure the large datasets associated with AI workloads while maintaining accessibility for authorized users for all stages of the workflow can introduce a web of complexities that is challenging to solve. DDN, collaborating with NVIDIA, is developing ways to integrate the strongest security measures, while keeping costs in check, eliminating data silos and keeping configurations as simple as possible.

Next generation architectures must feature integrated security

The challenge of protecting data is that security measures are often in direct opposition to requirements for fast AI development. Developing a data strategy that supports enterprise scale AI workloads starts with a Storage-as-a-Service approach that delivers cost-effective centralized infrastructure that meets the performance and scalability needs of complex AI applications and datasets.

This cloud-like model for on-premises deployments, aligns well with the announcement NVIDIA made today regarding their cloud-native, multi-tenant NVIDIA DGX SuperPOD architecture. NVIDIA and DDN are both seeing the need to supply the security needed to make scalable architectures sharable between business units or organizations and the infrastructure management to deliver AI services quickly and easily. Our experience delivering systems and collaborative engineering relationships will continue to benefit customers deploying AI at any scale, from individual systems to the largest clusters on the planet.

To serve the competing interests of making data hungry applications available to a wide variety of users and stakeholders an architecture is required that natively includes strong multi-tenant protection without compromising capabilities or creating management overhead. Next generation capabilities like this are being developed by DDN in our EXAScaler Software Defined platform, which will serve as Intelligent Infrastructure for the AI era.  When it is integrated with NVIDIA BlueField data processing units (DPUs), customers automatically get a strict security barrier that separates the application from the storage environment, gaining much stronger protection than any storage and network isolation available today.

What is EXAScaler Software Defined?

DDN EXAScaler Software defined is the next generation intelligent infrastructure for the AI era. 100% software and completely portable, it is uniquely capable of serving all IT and at-scale workloads with on-demand performance. Designed to simplify all aspects of management, EXAScaler radically simplifies deployment to adapt to changing environments and applications demands without user intervention. Everything is service oriented and available to users to be able to dial-in their resources via self-service, with QoS to protect mission critical applications.

The world’s first hybrid elastic IO engine supplies breakthrough performance that automatically places and accelerates data based on IO pattern and size. With all the intelligence needed built right in, EXAScaler delivers on the promise of automated on-demand infrastructure with the minimal administrative overhead of cloud regardless of deployment type – on premises, in the cloud or hybrid configurations. Click here to receive a detailed briefing on these and other AI-boosting capabilities and availability.

What Is BlueField?

The NVIDIA BlueField DPU delivers a broad range of advanced networking, storage, and security services for complex workloads like AI. By integrating an array of Arm cores with their Connect-X network adapter and a PCIe switch, NVIDIA can offer hardware acceleration engines with full data center infrastructure on chip programmability. This infrastructure is ideal for a software-defined storage implementation like EXAScaler Software Defined.

Imagine if you could immediately eliminate several barriers to AI implementation.

Security and privacy concerns are often top of mind for IT teams deploying AI within organizations. Other hurdles include the anticipation of complexity that might be beyond the IT teams current experience and how to efficiently handle the volume of data required and the widespread need to access this sensitive information. Creating a cloud native service model and automating much of the management and deployment tasks goes a long way towards intelligently empowering IT teams with the tools needed to manage these powerful and scalable infrastructures like DGX SuperPOD.

EXAScaler Software Defined on BlueField eliminates the need for dedicated storage controllers and allows storage to scale on-demand and simply in proportion to the network. Network management complexity? Gone. Storage management? Automated. Security? Addressed by the inherent advantages of the BlueField and EXAScaler integration.

With EXAScaler deployed on BlueField, IO data now inherits all the benefits of BlueField’s network isolation, significantly reducing the number of attack vectors available to a malicious actor. Right away, hacked user accounts, rogue users, man in the middle attacks, hacked root accounts and other avenues for malicious activity are eliminated. All of this is accomplished while facilitating performance and easy management for data-intensive workloads.  This no-compromises approach means that centralized data services can be deployed at the scale required to deliver breakthrough economic benefits while adhering to the protection and data governance requirements found in healthcare and life sciences, government, financial services and other security sensitive industries.

To see a demo of these integrated capabilities, watch GTC session SS33004.

For more information on DDN and our other GTC sessions and to register for GTC, visit the GTC Registration Page.

To learn more about the integration and availability for your needs, please contact sales@ddn.com.

  • Kurt Kuckein
  • Kurt Kuckein
  • Vice President, Marketing
  • Date: April 12, 2021

This site uses cookies. Please see our Privacy Policy to learn more about how we use cookies and how to change your settings if you do not want cookies on your computer. By continuing to use this site without changing your cookies settings, you consent to the use of our cookies. Privacy Policy