Author: Paul Turner, Vice President Product Management at VMware

Last year at VMworld 2020, VMware announced Project Monterey as the next major step in the evolution of cloud infrastructure for modern applications, including bare-metal workloads. The increasing velocity of digital transformation coupled with the adoption data analytics as mainstream IT technology is impacting the design and architecture of the next generation of data centers, and VMware’s answer to that is Project Monterey.

The data center is changing

Modern applications are driving new hardware requirements into data center infrastructure. The addition of function accelerators like DPUs to meet these requirements often results in the creation of application-specific silos, thus increasing the overall operational complexity of managing and scaling the infrastructure.

In a typical data center, the x86 based CPU complex tends to be a shared resource between infrastructure services and workloads. The CPU complex has come under increasing strain from the growing demand for infrastructure services, leaving fewer CPU cycles for applications. Paradoxically, a good portion of any newly added compute capacity is consumed by infrastructure services, lowering the overall efficiency of the data center.

As the perimeter of enterprise computing expands into a distributed multi-cloud architecture, the nature of security threats and vulnerabilities are changing constantly. The CPU-based security model is increasingly becoming a single point of failure for both the provider and tenant domains.

Reimagining virtual infrastructure

Project Monterey solves this problem by reimagining virtual infrastructure as a distributed control fabric through tight integration with DPUs (Data Processing Unit – also known as SmartNICs). The DPU plays two distinct roles in realizing this concept.

1. The DPU can offload and accelerate critical infrastructure services like network, security, and storage from the CPU. The newly available CPU cycles can then be used by key workloads, resulting in improved data center performance. This is a great scenario for enterprises who like their current infrastructure management model but are looking for ways to improve the efficiency and performance of their data centers.
2. The DPU can also be made an infrastructure fabric control point, on par with the x86 CPU, to scale network, security, storage and manageability functions across the infrastructure. In this scenario the DPU runs an instance of ESXi and its lifecycle is independently managed from that of the x86 CPU, enabling the following new capabilities:
   • Bare metal support: Since the ESXi instance running on the DPU is independent of the x86 CPU, the OS running on the x86 CPU can be ESXi, Windows or Linux. The same virtualized operating model that customers are familiar with, can now be extended to bare metal use cases.
   • Clean separation of provider and tenant domains: Service providers can achieve better security and a clean separation of services running on the DPU from that of their tenant’s x86 workload domain.

An Industry Change, not just a vSphere change

As you can see, the potential for an isolated execution environment for workloads is quite compelling with significant security, trust boundary and performance benefits. This would not be possible without support from our technology partners. Together with the following partners, we are working to redefine the infrastructure for tomorrow:

• Server OEMs: Dell, HPE, Lenovo
• DPU (SmartNIC) vendors: NVIDIA, Pensando, Intel

I’d like to take the time here to say thanks to all our partners for their continued support to help build the next generation infrastructure platform.

What’s new?

Since our initial announcement, we have continued to understand the key infrastructure challenges faced by our customers and work with our partners to build out key capabilities under Project Monterey. In August of this year, we kicked off the Project Monterey Early Access program which provides customers an opportunity to collaborate with VMware to test and validate their use cases in a lab environment to address and solve key pain points. This was followed by NVIDIA’s own Project Monterey Early Access Program based on servers from Dell Technologies and Lenovo.

To learn more about Project Monterey and VMware’s collaborative approach to developing a vibrant ecosystem, please register for VMworld 2021 and attend these great sessions offered by VMware and its partners:

• [MCL2379] Partner Roundtable Discussion: Project Monterey—Redefining Data Center Solutions(VMware)
• [MCL1833] 10 Things You Need to Know About Project Monterey(VMware)
• [MCL1401] Project Monterey: Present, Future and Beyond(VMware)
• [NET3382S] The Roadmap from ESXi to Project Monterey and Beyond (NVIDIA)
• [NET2790S] Software-Defined, Hardware-Accelerated Data Center with Project Monterey (NVIDIA)
• [NET2847S] Accelerate Infrastructure Functions and Improve Data Center Utilization (Intel)

Resources