VMware vRealize AI Cloud VMworld 2020

Posted by

At VMworld 2020 the new vRealize AI Cloud was announced.  This is a huge step forward for VMware to bring AI into the fold to help optimize customer environments.  The ability to create a digital copy of your environment and run ML against the copy in hundreds of scenarios is a game changer for VMware.  This new ability to ensure KPIs are being met in a continuously changing eco-system will help many customers as their datacenters expand and bring in new technologies.  Starting with vSAN optimization has been a major ask by many customers the past couple of years.  Below you will find an overview of this latest solution release from VMware.

Other VMworld 2020 announcements:

The Vision behind AI Cloud is to build reinforcement learning capabilities. This tool will not just be another tool to have another tool to support the tool of that tool you bought 6 months ago.  AI Cloud will analyze the operational aspects of a datacenter allowing you to continually tune the datacenter based on a business KPI or application SLO, which will be huge for my DevOps friends out there.

AI Cloud accomplishes this by leveraging a third branch of machine learning by learning about operating environment.  Then analyzes performance and then builds a digital twin of your environment.  It will learn and apply config changes to meet intended KPIs.  Once changes have been applied it will show you if you met the expected outcomes.

The first use case of vROPs integration with the intent of continually tuning vSAN will be a huge assist to customers who are trying to maximize the most out of their DellEMC VxRails clusters or VMware vSAN ready node clusters.  You could easily see in the future AI Cloud expanding out to networking with network insight or out to the public cloud space with cloud health.

Between traditional datacenter, private cloud, hybrid cloud, and public cloud going into multi-cloud; today’s IT eco-system has become complex with many control points and presents many opportunities for performance issues.

As workloads are moving from onprem to the cloud and in some cases running in both environments at the same time there are many interdependencies that can have a ripple effect if a single change is made to improve a single applications performance.

Speaking of machine learning techniques this is accomplished by three different fields

  • Supervised learning really broken down into two sub-categories
  • Classification – identifies A /B or Yes or No – output is categorical
  • Regression – uses X/Y axis to determine a relationship between variables with known data

Reinforcement Learning – this is where data collection and observation comes into play with continual learning for greater rewards.

AI Cloud takes optimization one step further by performing data collections on your vSphere and vSAN environment and sends it up to a data lake and compares against all workloads demands.  This is where the digital twin mentioned earlier comes into play.  The machine learning algorithm is applied in a variety of tuning scenarios to see where the largest reward with the minimal impact is going to be.  This will help prevent you from making an environmental change to benefit a single application while negatively effecting the performance of multiple other apps.

It even comes with a slick new UI built into vROPs under Optimize Performance on the home page

Outcome based UI that is easy to understand

Each wheel represents a datacenter and each dot represents clusters

The legend in the bottom left hand corner allows for easy interpretation of the screen.

This is a perfect example of how the continual learning mode is making changes every day in your environment.  These changes are easy to dig down into. 

There is also a report available if you need to perform audits.

It is my opinion that AI Cloud is one of the coolest announcements to come out of VMworld 2020.  Having been an architect at global companies, I cannot tell you how many hours were spent in strategy meetings for the purpose of this tool solves.  I’m definitely looking forward to the evolution of machine learning at VMware and this new AI Cloud offering.  As always, I hope y’all found this helpful.

Leave a Reply