Hello VMware, and Welcome to the Big Memory Party!

At VMworld 2021, VMware revealed Project Capitola – a technology preview of software they are developing to aggregate heterogeneous memory tiers into a uniform, virtualized memory layer for vSphere environments.  VMware is a recognized industry leader with thousands of enterprise customers across the globe. When VMware announces they are investing R&D resources into a new area, the industry pays attention.  This announcement brought the spotlight onto software-defined Big Memory.

For the past four years, we at MemVerge have been telling the world that software-defined Big Memory is a new foundation for applications to handle the increasing demands from data-centric applications. The world is instrumented, data is generated at accelerating volumes, and applications such as AI/ML and bioinformatics need access to more memory to process bigger data faster. New memory media (such as Intel Optane Persistent Memory) and new memory buses (such as CXL) are emerging.  All you will need is a layer of memory virtualization software to turn new memory hardware into a pool of scalable and flexible Big Memory.  The evidence is clear and the case is compelling, but as a startup, the reach of our Big Memory message has been limited.

So, when we hear from the VMworld sessions that VMware is predicting a Big Memory transformation in cloud infrastructure that combines the new memory hardware innovations with memory virtualization software, we are delighted!  Our views are very much aligned, and it is a great validation to hear the same message that MemVerge has been championing from an industry leader on the big stage of VMworld.  Big Memory is coming to a cloud data center near you!

Creating a robust and high-performance software-defined memory layer is not easy.  The first version of Project Capitola software is expected to become available in 2H 2022.  It will be part of the vSphere software, and will not expose the persistence capabilities of Intel Optane Persistent Memory.  MemVerge’s flagship product, Memory Machine^TM  is now on Release 2.0 and it supports bare metal, various hypervisors and major public cloud deployments.

We have learned a lot from our customers as they deploy our software in business-critical environments.  We have learned that aggregating disparate memory types into a memory layer abstraction is just the beginning. The transformative benefits derive from delivering advanced data services by leveraging the characteristics of the underlying memory technology. For example, MemVerge ZeroIO^TM in-memory snapshots take advantage of the underlying persistence capabilities of Intel Optane Persistent Memory, and have helped one of our bioinformatics customers reduce the runtimes of analytical pipelines by over 60%.

Twenty years ago, VMware was a small startup with a big vision of virtualizing every x86 server.  Twenty years later, its vSphere software runs across enterprise data centers around the world.  Today, MemVerge is a small startup with a big vision of petabyte-pools of disaggregated memory available to every application across the multi-cloud.  VMware has been a role model for us at MemVerge.  It is our pleasure and privilege now to welcome VMware to join us as a fellow pioneer in this Big Memory mission.