Here at VMworld 2019 (San Francisco, CA), we are very excited to share the first previews into Project Magna’s adaptive optimization and self-tuning engine.
This was first mentioned last year by VMware CEO, Pat Gelsinger, where he envisioned the true AI/ML ’Self-Driving Data Center’. Modern day applications are dynamic in nature and requires that the underlying infrastructure be continuously, automatically and adaptively optimized to deliver the expected performance and SLA guarantees of today’s businesses.
Project Magna is the first instantiation of ‘Self-Driving Data Center’ vision, beginning with VMware vSAN. This is a SaaS-based solution which uses reinforcement learning where it will collect data, learn from it, and make decisions that will automatically self-tune your infrastructure to drive greater performance and efficiencies.
The Magna engine runs by itself, manages itself, makes performance tuning configurations by itself and there are guard rails within the ML algorithms that will not decrease performance by any means. It either optimizes your desired read and write KPIs or stays put.
Types of Artificial Intelligence
I do want to touch on the different types of Artificial Intelligence and share how Project Magna uses Reinforcement Learning to analyze your performance KPI’s and compares it against other vSAN clusters running similar application workloads.
You’ll typically see technologies that use Machine Learning or Deep Learning and its algorithms are dependent on using existing datapoints to correlate patterns to forecast outcomes or predict what next action steps to take. But Reinforcement Learning combs through your data and runs thousands of scenarios that searches for the best reward output based on trial and error on the Magna SaaS analytics engine. And this is automatically and continuously done across your vSAN clusters to ensure it’s always using the best settings to maximize throughput and minimize latency of your modern hyperconverged infrastructure.
Automatic Optimization
As you might already be familiar with vRealize Operations, one of the key tenants of the self-driving data center is ‘Continuous Performance Optimization’ and this is where you’ll navigate to configure Project Magna’s tunables. You’ll be able to select all or specific vSAN clusters that you want to apply the self-tuning to.
Choose the optimization goal for your infrastructure and enable Magna for:
- Read Performance
- Reduce read latency and increase read throughput
- Headroom will be at least 10% of host memory
- Write Performance
- Reduce write latency and increase write throughput
- Read performance is maintained at a minimum
- Balanced
- Optimization to reduce both read and write latency based on workload requirements
Visualize your latest performance index readings or search for a specific time frame and compare your current KPIs, the industry averages, and the industry averages with Magna enabled. Once you’ve enabled the Magna optimization to your vSAN clusters, it continuously adapts to the changing needs of your applications. With a mouseover of the buttons on the graph, you’ll learn exactly what actions were taken and when.
In the screenshot below, we see that Magna increased vSAN cluster vc2c1 read cache size by 50gb on July 23rd for the desired KPI: latency reward.
Learn More…
I invite you to learn more about Project Magna if you’re here at VMworld. Here are the breakout sessions that will outline the strategy, performance results, and use cases for Project Magna – so I highly recommend you checking them out!
- HCI1620BU – Artificial Intelligence and Machine Learning for Hyperconverged Infrastructure
- Monday, Aug 26th – 4-5pm
- MLA2021BU – Realize your Self-Aware Hybrid Cloud with Machine Learning
- Monday, Aug 26th – 5-6pm
- HCI1650BU – Optimize vSAN performance using vRealize Operations and Reinforcement Learning
- Wednesday, Aug 28th – 11-12pm
For any other questions or more information, please don’t hesitate to email us at: [email protected]
