Modern applications are becoming increasingly complex – and computationally hungry – and are placing increasing demands on the CPU. VMware last week presented a solution which may go some way to solving this problem.
The CPU has in the past been used for all computational tasks, but it’s now becoming apparent that are limitations to what it can process. One way of coping with this has been to introduce hardware accelerants such as graphic processing units, field-programmable gate arrays, and network interface cards (NICs), but these solutions were becoming increasingly complex to manage. Clearly, the hardware was not keeping up with the demands of software. Something had to change.
At VMworld last week, VMware said it’s been working on this problem for about the last three years and announced the release of it solution, known as Project Monterey. This involves a re-architecting of its VMware Cloud Foundry (VCF) hardware, the hybrid cloud platform that’s used to manage virtual machines and containers.
This is about us going deep into the infrastructure. This is about us moving the industry to a more agile infrastructure.
Paul Turner, VMware
Writing in his vSphere blog, VMware’s Kit Colbert says Monterey uses the new hardware technology of ‘SmartNICs’ to deliver improved performance, zero-trust security, and operations in VCF deployments. When Monterey is released to the market, thought to be sometime next year, it will also offer customers support for bare metal operating systems and applications.
He writes that Project Monterey is going to be able to deliver peak performance by offloading network processing to the SmartNIC, improving network bandwidth and freeing up CPU cycles. This will improve the performance of the app and enable consistent operations across all apps. There will also be improved security controls, because security functions will offloaded to the SmartNIC.
Deep dive
Paul Turner, VP, vSphere product management, told reporters on a call last week that while VMware has enabled a more dynamic and agile environment in which Kubernetes can be run, the infrastructure also needed to be dynamic as well.
He said the project was named after the Monterey Canyon, which, at a mile deep and 95 miles long, was the deepest submarine canyon on the West Coast of the United States.
“This is about us going deep into the infrastructure. This is about us moving the industry to a more agile infrastructure.”
VMware is working with partners including Intel, HPE, Dell, Lenovo and Nvidia, to build infrastructure based on Monterey.
Pat Gelsinger, VMware CEO, also on the call, said, while the infusing of containers and Kubernetes as a control plane into vSHpere was ‘a big deal’, Project Monterey was re-architecting ‘the plumbing’ of what it was doing inside vShpere.
“It’s a major rebuilding and rewrite of the underlying architecture, but the magic is that an application doesn’t change. If it can run in a VM or a container it’s able to take advantage of these capabilities because it’s all hidden inside of how vSphere will operate.”
He also predicted that it will be increasingly important to be able to operate with the very large data sets of next generation apps, when, for example, AI dataset workloads become standard in the enterprise
Share