A sharper edge in a multicloud world


cloud computing concept
Image: serpeblu/Adobe Stock

Cloud computing has both literally and figuratively undergone a vapor-like diffusion and circulation to now permeate almost every aspect of life and business. Had the network engineers who drew the initial swirly-edged cloud bubble diagrams to represent virtual compute instances known just how cloud-like cloud computing would ultimately be, they may have allowed themselves another pat on the back.

Today we exist in a world with private cloud, public cloud and hybrid cloud all working in unison. This diversity has driven us to understand multicloud—a combination of more than one cloud instance types, typically from more than one cloud services provider—as the new norm.

Some cloud purists even talk of polycloud, a state where large computing workloads in a single monolithic application or database are “separated out” across a multicloud deployment. Put simply, today there are many clouds in many shapes doing many jobs.

Why multicloud happened

Thinking about why multicloud happened in the first place, this diversity stems from the fact that the typical enterprise runs somewhere in the region of 500 applications to drive business.

Because the prudent business knows it needs to spread risk and grasp different cost and performance efficiencies where they can be identified, a good cloud estate usually encompasses public-private hybrid deployments that also now feature dedicated edge clouds.

Built to serve processing, analytics and storage; handle complex tasks such as data streaming and machine learning (ML); and perhaps even operate with a view to welcoming quantum-based services, edge clouds are aligned to the specific needs of devices that populate the Internet of Things (IoT) and the new age of embedded device engineering.

SEE: What is multicloud architecture? (TechRepublic)

The diversification of cloud is clearly empowering, but as we know, there are some drawbacks; the payoff for heterogeneity and variation in this case is management headaches and infrastructure complexity. The burden of maintenance and security policy updates has the potential to make the cloud more difficult to navigate.

Keen to address the current state of the global cloud nation’s edge-centric management challenges, VMware used its Explore 2022 conference this month to announce cloud and edge infrastructure software products.

The shape of modern workloads

VMware’s in-house research suggests that 580 million “modern” workloads, which we can take to suggest are inherently complex, cloud-native, distributed and so on, are expected to run on diverse environments spanning public cloud, on-premises, edge and hosted clouds by 2024. With edge environments front of mind, the company also specifically calls out telco (telecommunications companies) clouds as a key growth area.

“In today’s highly distributed environments, customers need cloud and edge infrastructure that can empower them to scale their operations with consistency, availability and security, wherever their enterprise workloads are running—and at the lowest possible TCO (total cost of ownership),” said Mark Lohmeyer, senior vice president and general manager of the cloud infrastructure business group at VMware.

Newly released, VMware vSphere 8 and VMware vSAN 8 are major new compute and storage solutions.

The company claims that VMware vSphere 8 will usher in a new era of heterogeneous computing by bringing DPUs (data processing units) into the fold along with CPUs (central processing units) and GPUs (graphics processing units). Built to handle dedicated data processing tasks, DPUs are engineered with specific hardware acceleration engineering throughout their circuitry to provide more efficient and faster workload execution. Used in this context, VMware says they meet the throughput and latency needs of modern distributed workloads.

Previously known as Project Monterey, VMware’s work to harness DPUs is a result of collaboration with technology partners AMD, Intel and Nvidia as well as OEM (original equipment manufacturer) system partners Dell Technologies, Hewlett Packard Enterprise and Lenovo. By running infrastructure services on DPUs and isolating them from the workload domain, vSphere on DPUs is said to boost infrastructure security.

Terrific traffic transports

Additionally, NSX Distributed Firewall, which is now in beta, will offload to DPUs to scale customers’ security operations by securing east-west traffic—traffic between devices in a data center—at line rate without the need for software agents in a given IT deployment.

We also know that vSphere 8 will accelerate AI (artificial intelligence) and ML applications by doubling the virtual GPU devices per virtual machine (VM), delivering a 4x increase of passthrough devices and vendor device groups, which allow the binding of high-speed networking devices and the GPU.

For DevOps (development and operations) teams, vSphere 8 includes Tanzu Kubernetes Grid 2.0, a technology that expands Kubernetes capabilities to include multi-availability zones for improved resilience as well as simplified cluster life cycle and package management. Additionally, vSphere 8 customers will benefit from the new Cloud Consumption Interface service to use cloud infrastructure and Kubernetes resources for VMware Cloud environments in an open, programmable way.

SEE: iCloud vs. OneDrive: Which is best for Mac, iPad and iPhone users? (free PDF) (TechRepublic)

At VMware Explore 2022, VMware is introducing new and enhanced edge solutions purpose-built for edge-native apps and their performance, autonomy and latency requirements. They include VMware Edge Compute Stack 2, a fully integrated edge platform to operate modern, existing and future edge-native applications.

VMware Edge Compute Stack 2 will help customers address the needs of simplicity and scale at the edge. New capabilities will include support for smaller cluster sizes to run containers efficiently on smaller COTS (commercial, off the shelf) hardware. Additionally, the release will also offer higher performance with GPU passthrough support to enable AI/ML use cases.

VMware also showcased its Managed Edge Solution with NTT. VMware and NTT are announcing an expanded partnership that will bring to market a fully-managed edge compute solution with private 5G connectivity, delivered by NTT across its global footprint and powered by VMware Edge Compute Stack. The joint offering is said to give enterprises the ability to deploy, manage and monitor applications closer to the edge than ever before.

Edge-native apps

According to VMware, edge computing is the next evolution of the distributed digital enterprise with the fastest growing segment of workloads and worldwide spending expected to be a figure approaching £150 billion ($174.79 billion USD) in 2022.

VMware’s edge strategy is designed to bring together products across VMware that enable organizations to run, manage and secure edge-native apps across multiple clouds at both the near edge for consumer-level IoT devices and micro data centers and far edge for smart retail devices, connected medical devices, AI machines and wider/smaller IoT devices such as sensors.

We know that IoT and edge computing are not wholly interchangeable terms. Generally, we think of IoT as related to the device and edge computing as the process via which we can perform compute actions for tasks such as aggregate analytics, trend analysis and outlier identification.

The fact that we also talk about edge devices, edge servers and now edge clouds underlines the way IT workloads are moving outwards in so many places, which is precisely where VMware is focusing its efforts in this space. The era of cloud-native software application development has now given rise to the creation of edge-native apps.

The edge just got sharper, safer, smarter, swifter and hopefully slicker too.



Source link