Power your AI and modern applications with FlexPod

As enterprises work towards digital transformation and operational efficiency, they need technology that can adapt to dramatic shifts across the organization and keep up with the ever-increasing demands of modern workloads and applications.

Guest Blogger:
Siva Sivakumar, Senior Director, UCS Solutions, Computing Systems Product Group, Cisco

Today’s businesses are inundated by large volumes of data – data that remains relatively untapped. This data can help uncover critical insights that can drive new and improved customer experiences and create innovative new business opportunities. As these critical insights are needed in real-time, there is a need to ingest, collect, store and process data much more quickly and efficiently than ever before. As a result, IT organizations need a platform capable of supporting modern workloads like Artificial Intelligence (AI), Machine Learning (ML), and Deep Learning (DL) as well as next generation databases and enterprise applications. This requires constant advancements in the capabilities and capacity provided by compute, network and storage resources.

Cisco and NetApp introduced our industry leading FlexPod converged infrastructure platform back in 2010. According to IDC’s Quarterly Converged Systems Tracker, over the past 2 years, FlexPod has been the fastest growing converged infrastructure platform, gaining 5 market share points. FlexPod is built on Cisco Unified Computing System (UCS) servers, Cisco Nexus switches and NetApp storage and is trusted by nearly 9000 customers worldwide. Over the past 8 years the FlexPod platform has continued to evolve, incorporating the latest technology advancements from Cisco and NetApp and supporting the widest range of workloads of any converged infrastructure platform with over 170 validated designs.

We are excited to introduce many new innovations to the FlexPod platform:

Platform for innovation

AI and ML workloads with UCS 480 ML M5 with GPU support
Power modern applications including SAP HANA, SQL 2017, Oracle 12c, VDI w/GPU
Greater performance and efficiency with UCS M5, 100Gb Nexus, and ONTAP 9.5
Ultra-low latency with end-to-end NVMe using Cisco MDS and NetApp AFF A800
Rapid technology integration including upcoming persistent memory and MAX Data

Unmatched versatility

Multi-cloud connected for cloud-based backup, DR, tiering on AWS, Azure and GCP
Edge-to-core-to-cloud orchestration with enhanced Data Fabric and InterSight
Multi-site HA with MetroCluster IP and ACI Multi-Pod

Platform trusted worldwide

Simplify support experience with one contact support for the full stack from Cisco, NetApp or partner
Continuously optimize operations with cloud-based monitoring and predictive analytics
Most trusted platform as fastest growing converged infrastructure solution

This constant influx of new technologies in the FlexPod platform ensures that our joint customers will continue to get a flexible, powerful and reliable foundation for running their most demanding workloads – a platform built with the future in mind. FlexPod is also a superior platform for private and hybrid cloud environments. By integrating FlexPod with Cisco ACI Anywhere , Cisco CloudCenter, and NetApp Private Storage for Cloud, customers can consume and optimize compute and data resources in a way that aligns with their cloud strategy, with a consistent set of data management tools for edge, private, and public clouds. Both Cisco and NetApp are committed to continued investment in the FlexPod platform, incorporating cutting edge technologies and delivering new validated designs including our latest solution for AI/ML/DL workloads.

New FlexPod AI ML solution
83% of executives believe AI is a strategic priority for their businesses today¹, while at the same time, AI is becoming increasingly complex and demanding unprecedented levels of compute power. Designing, configuring and maintaining a reliable infrastructure to satisfy the compute, network and storage requirements for these initiatives is top of mind for our customers and their IT departments. To address this emerging opportunity and its associated challenges, last Fall, Cisco partnered with NVIDIA and expanded our Cisco UCS® portfolio with an optimized machine learning computing system targeted at deep learning. The Cisco UCS C480 ML M5 Rack Server is a 4U server with 2 Intel Xeon Scalable processors and 8 NVIDIA Tesla^® V100-32G GPUs with NVLink interconnect technology. The NVIDIA Tesla^® V100 is the world’s first GPU to break the 100 teraflops barrier of deep learning performance with a whopping 640 Tensor Cores. AI models that would previously consume weeks of computing resources, can now be trained in a few hours. With this dramatic reduction in training time, a whole new world of problems can now be solvable with AI.

We’ve now integrated this powerful Cisco UCS C480 ML M5 Server into FlexPod to deliver AI/ML/DL capabilities in a leading converged infrastructure platform. The FlexPod Datacenter for AI ML is designed to enable customers to easily utilize the platform’s extensive GPU capabilities for AI/ML workloads while reducing the time and the resources required for successful deployments with FlexPod’s simplified architecture. The Cisco UCS C480 ML M5 server can be managed alongside existing FlexPod infrastructure and data sources. IT staff can manage the server with the familiar tools they already use to administer their FlexPod systems, reducing costs and administrative overhead and eliminating infrastructure silos.

FlexPod Datacenter for AI/ML Solution Architecture

The FlexPod Datacenter for AI/ML solution delivers the following benefits:

Seamless integration of the Cisco UCS C480 ML server into the FlexPod environment
Highly configurable and scalable GPU-rich platform to support varying customer workload sizes and use cases.
Simplified operation – Unified platform management across all UCS servers using UCS Manager
Ability of the AI platform to utilize high-speed and highly available switching capability to access storage systems
Facilitate AI/ML deployments with guidance from Cisco and NetApp

Customers can now power their most demanding AI applications with a FlexPod platform that is optimized for analytic workloads. For customers with existing FlexPod deployments, they can add the Cisco UCS C480 ML M5 computing engine and tap into the valuable data already stored on their FlexPod infrastructure to gain valuable insights.

Next generation enterprise applications and databases

FlexPod’s architecture can be easily configured and adapted for the constantly growing and changing needs of the data center. It is tested and pre-validated with unmatched versatility to run hundreds of applications and workloads with reliable performance that scales with business demands.

With the new platform technology innovations outlined earlier in this blog, FlexPod is fully equipped to power the latest generation of enterprise applications and databases. New tested configurations have been released for Oracle 12cR2 RAC and 3D Graphics Visualization with VMware and Nvidia. Additional designs are also coming over the next few weeks offering a new validated SQL 2017 design and a reference architecture for SAP vHANA. FlexPod can handle the most demanding database workloads and significantly improve database response times by taking advantage of NetApp’s AFF A-Series All Flash Arrays and the high performance of Cisco UCS M5 servers. These gains are further amplified when combined with the scale and efficiencies of a Cisco ACI fabric that enables software-defined, policy-based, automated deployment of enterprise applications. The FlexPod validated design library can be found here.

Learn more about FlexPod
See what enterprise IT professionals are saying about FlexPod on IT Central Station
Visit https://cisco.com/go/flexpod or www.flexpod.com to learn more about FlexPod.

FlexPod @CiscoLiveEurope

If you are attending Cisco Live EMEA in Barcelona Jan 28 – Feb 1, NetApp is a platinum sponsor.

Read the NetApp blog
Visit the NetApp booth (P2) in the World of Solutions to see FlexPod
Visit the Cisco booth to see the new UCS C480 ML M5 server for AI workloads that is now integrated with FlexPod
Attend these sessions:
- Is your IT Infrastructure ready for Artificial Intelligence and Machine Learning – BRKPAR-2999
  Monday, Jan 28, 2:00 PM – 2:40 PM
- End-to-end NVMe architectures: Implementation instructions and performance expectations – BRKPAR-299, Wednesday, Jan 30, 10:45 AM – 11:45 AM
- Practical applications of a Hybrid, Multi-cloud architecture – BRKPAR-2996
  Wednesday, Jan 30, 4:30 PM – 6:00 PM
- Managed Private Cloud for Practical applications on a Hybrid, Multi-cloud architecture – BRKPAR-2998, Thursday, Jan 31, 2:00 PM – 2:40 PM