AIvengers: the infinity infrastructure for virtualized AI

AIvengers: the infinity infrastructure for virtualized AI

An information system is built from a multitude of stones, that although powerful on their own, only show their true power once brought together. With the new Dell Technologies Validated Design for AI solutions, the six artificial intelligence gems are united in a virtualized infinity infrastructure. Avengers! Assemble!

Using artificial intelligence is the everyday reality for our customers who use it to extract business value from the considerable amounts of data they collect. Yet, running machine learning or deep learning algorithms often calls for implementing special systems that are deployed and managed separately from the more conventional corporate applications. These siloed resources cause complexity and stop organisations from getting the full power out of their component stones.

No alt text provided for this image

This is why Dell Technologies, in partnership with VMware and NVIDIA has brought together the six components required by AI in a fully integrated, supported and certified infrastructure able to take charge of both new generation workloads and conventional applications using unified management.



With all six stones united in Dell Technologies Validated Design for AI, you could simply snap your fingers, and complexity would cease to exist.


????????Power Stone - The VxRail orb

Based on Dell EMC PowerEdge servers and born from a close partnership between Dell Technologies and VMware, the VxRail hyperconverged platform is the orb that brings hardware and software resources together in a single circle. Fully virtualized, VxRail brings the power to accelerate application deployment in the VMware environment and simplify the management and scalability of the underlying infrastructure. Of course, VxRail is primarily a platform certified by NVIDIA. The natively installed vSphere software supports the powerful NVIDIA A100, A30 and soon to be released H100 GPUs, specially designed to accelerate AI applications.

No alt text provided for this image


????????Time Stone – The PowerSwitch portal

Dell EMC PowerSwitch switches offer open infrastructure based on market standard protocols. The hardware is uncoupled from the software so that users gain the flexibility they need to take charge of the most innovative applications. PowerSwitch is the magic portal through which enterprises will enter the world of virtualized AI.

No alt text provided for this image


????????Space Stone – The PowerScale cube

This is your data's tesseract; the cosmic cube that comprises your organisation's digital treasure. PowerScale is the world's most flexible NAS scale-out solution. Able to evolve easily from just a few terabytes to tens of petabytes, in all flash or hybrid modes with multiprotocol (NFS, SMB, HDFS, S3, NDMP, REST, HTTP and FTP) access, allowing a variety of applications to access the same data sets with different protocols. This is the ideal platform for forming a data lake for feeding your algorithms with unstructured data.

No alt text provided for this image


????????Mind Stone - The NVIDIA AI Enterprise vision

NVIDIA AI Enterprise is a suite of AI and analysis frameworks and tools, like PyTorch or TensorFlow, certified by NVIDIA to run on vSphere and vSphere with Tanzu. This way enterprises can run AI workloads on both virtual machines and Kubernetes containers. The suite can also be used on public cloud instances. Hence, NVIDIA AI Enterprise puts into practice the vision of AI without silos, managed with the same tools as the rest of the data centre?and cloud.

No alt text provided for this image


????????Reality Stone – NVIDIA GPU data flows

The NVIDIA GPUs integrated into the Dell Technologies Validated Design for AI platform will considerably speed up data flows for inference and learning processes. They have the capacity to offer enterprises a new vision of reality by allowing artificial intelligence to extract relevant information for the business from an elusive torrent of data. With the planned support for the very latest NVIDIA H100 GPU, one based on the new Hopper architecture and 4 nm chips, PowerEdge servers and Validated Designs will be able to run the most complex models with tens or hundreds of thousands of parameters.?

No alt text provided for this image


????????Soul Stone – The VMware vSphere core with Tanzu

Containerising AI workloads will allow administrators, developers, and data scientists the ability to handle the life cycles of their applications more easily, from prototyping to deployment. The orchestration and automation brought by Kubernetes will lift the technical barriers that complexified and slowed bringing the models and VMware environment into production. AI specialists will no longer need to sacrifice their precious time to ensure workload performance, availability, and security.

No alt text provided for this image


You now understand that you're about to take the full force of virtualized AI. No other infrastructure has ever the might to wield six Infinity Stones. The universe of AI lies within your grasp.

要查看或添加评论,请登录

Segundo Ramos的更多文章

社区洞察

其他会员也浏览了