HPE elastic infrastructure powered by IaaS for AI

[Rehab-Alzahrani]

How HPE can help you build a powerful AI infrastructure to facilitate advanced analytics and enhanced fact-based decision-making capabilities

---

In today’s increasingly data-driven economy, advanced analytics have become a top priority for any forward-thinking enterprise.

Unlocking the true value of the data available remains by far the biggest challenge. Data sets have become far too large for manual interpretation alone, which is why between 60 and 73% of all enterprise data ends up going unused for analytics.

To overcome this challenge, enterprises need access to the powerful hardware necessary for analyzing these data sets at scale. HPE offers an infrastructure-as-a-service (IaaS) solution that combines cutting-edge hardware, cloud-based artificial intelligence, and flexible service and delivery options to help drive your digital transformation.

Here is how HPE can help you build your AI and analytics infrastructure the right way:

Hardware: Powering AI workloads with high-end servers

AI applications, such as deep learning and advanced analytics, require powerful, specialized hardware to work efficiently with vast data sets. The demand for scalable storage systems is growing rapidly too, as data continues to proliferate. To meet these technical challenges, HPE provides powerful enterprise servers tailored to the needs of AI:

• HPE Apollo 6500 Gen10: With eight GPUs per server and a bandwidth of 300 GB/s, this deep-learning platform can be used to train AI models in hours and minutes, rather than weeks or days. This helps your researchers become more productive and greatly reduce project cycle times. Learn more
• HPE Apollo 6500 Gen10 Plus: Built for the exascale era, this cutting-edge system is powered by nVidia HGX A100 tensor core or AMD InstinctTM MI 100 GPUs to power the most complex AI workloads. Designed with maximum availability, scalability, and reliability in mind, it supports up to 16 GPUs and a wide range of operating systems. Learn more
• HPE Apollo 2000 Gen10 Plus: Bringing the power of supercomputing to data centers of any size, this multi-server platform packs extreme performance and workflow flexibility into a small space while delivering the cost and scalability benefits of a shared infrastructure. Learn more
• HPE ProLiant DL385 Gen10 Plus: Equipped with AMD EPYC 7002-series CPUs, this server offers memory speeds of up to 3200 MT/s and increased data transfer rates thanks to PCIe Gen4 capabilities. This is ideally suited to virtualization, memory-centric workloads, and high-performance computing.  Learn more
• Cray ClusterStor E1000: For all your data storage needs, the Cray ClusterStor E1000 solution handles storage in an efficient and scalable way with significantly fewer drives than many competing platforms. This frees up budgets to spend on CPU/GPU nodes to accelerate time-to-insight. Learn more

Software: Scaling AI workloads with flexible software

Software provides the flexibility you need to scale and customize your AI workloads and drive your digital transformation forward. From sandbox experimentation and AI model training and maintenance, HPE provides the software necessary to operationalize machine learning at massive scale:

• HPE Ezmeral Container Platform: It is an enterprise-grade container platform that supports both cloud-native and non-cloud-native monolithic applications with persistent data. It includes innovations from HPE’s recent acquisitions of BlueData and MapR, together with open-source Kubernetes for orchestration. BlueData has a proven track record of deploying non-cloud-native AI and Analytics applications in containers and MapR brings a state-of-the-art file system and data fabric for persistent container storage. Now enterprises can extend the agility and efficiency benefits of containers to more of their enterprise applications—running on either bare-metal or virtualized infrastructure, either on-premises, in multiple public clouds, or at the edge.. Learn more.

Services and delivery options: Unlocking value with a cloud-first experience

• HPE GreenLake:

When building a scalable AI development infrastructure, enterprises used to look to the public cloud option first, as they can pay-per-use, scale up and down whenever it needed rather than risk getting stuck with expensive proprietary and overhead of on-premises infrastructure. Also, the rapid pace of technological advancement makes them prefer a flexible and cost-effective architecture that offers access to resources on demand.

Many business-critical workloads, such as ERP and CRM, lack the agility of modern cloud services. HPE GreenLake helps overcome these challenges by bringing the speed and agility of the cloud to where your apps and data are today on-premises and off-premises, allowing you to facilitate faster digital transformation without rapidly rising costs.

You can scale up and down, and enable essential data and analytics as a service in all of your locations. By centralizing operations and insights across your cloud infrastructure with a single self-service platform, you get a unified view where you can monitor usage and retain full visibility into costs, compliance, security and more.

HPE GreenLake is a fully managed service that allows enterprises to offload monitoring and management of their on-premises and public or hybrid clouds to world-class IT operations centers.

• HPE PointNext Services: This range of services helps enterprises stay ahead of the curve by making it easier to adapt to new data sources and evolving business models. Combining the power of AI with cloud and edge computing, it frees your IT team from repetitive routine tasks so they can focus on high-value business goals. Learn more

HPE helps enterprises unlock the practically limitless potential of deep learning by delivering a highly customizable and scalable infrastructure-as-a-service platform. Learn more about our IaaS solutions for AI.

Rehab Alzahrani
Hewlett Packard Enterprise

twitter.com/hpe
linkedin.com/company/hewlett-packard-enterprise
hpe.com/sa