Presenters

Source

Unifying the Hybrid Cloud: How VM Service is Revolutionizing VM and Container Management 🚀

Hey tech enthusiasts! Shruthi Rajashekar, an engineering manager at Broadcom, is here to shed some light on a game-changer for hybrid cloud environments. For the past decade at VMware Broadcom, Shruthi has been instrumental in developing foundational technologies like vMotion and VM service, bridging the gap between traditional virtualization and modern cloud-native infrastructure. Today, she’s diving deep into how a unified control plane for virtual machines (VMs) and container-based workloads can be achieved using VM service, a VCF offering, and why this approach is absolutely critical for platforms demanding high availability and operational excellence.

The SRE Struggle: Operational Fragmentation is a Pain Point 😩

Let’s face it, Site Reliability Engineers (SREs) are often caught in a whirlwind of complexity. One of the biggest pain points they grapple with is operational fragmentation.

  • The Problem: Containers are typically managed with Kubernetes-native tools, while VMs are handled by a completely separate stack of infrastructure tools, dashboards, and APIs.
  • The Result: This leads to operational drift. SREs have to correlate alerts from disparate systems, making troubleshooting a nightmare. They’re constantly switching contexts between different systems, and standardizing automation becomes a monumental task.

VM Service: The Kubernetes-Native Solution for VMs 💡

This is where VM service, an offering within VMware Cloud Foundation (VCF), steps in to fundamentally change the game.

  • The Shift: Instead of treating VMs as external infrastructure, VM service integrates them directly into the Kubernetes control plane.
  • The Impact: SRE teams can now manage both VMs and container-based workloads using the same APIs, workflows, and operational patterns. This dramatically reduces cognitive load and fosters a unified operational model for hybrid platforms.

Why is Hybrid Management So Tricky? 🤔

Most enterprises still rely on VMs for their critical workloads, while newer services are built on container-based infrastructure. The core challenge lies in the fundamental differences in their platform-based workflows and their completely different life cycle management models.

VM service tackles this by making virtual machines first-class Kubernetes resources. This means:

  • A VM can now be declared, deployed, scaled, and managed just like a pod or deployment, using familiar Kubernetes primitives.
  • This brings infrastructure and application operations closer together.
  • Platform teams can standardize workflows using Kubernetes, enabling smoother migrations, simplifying hybrid application architectures, and allowing organizations to modernize incrementally without forcing legacy workloads to be containerized immediately.

The SRE Operational Challenges: A Deeper Dive 🛠️

To reiterate, SRE teams face:

  • Fragmented failure modes: Inconsistencies in how issues manifest across different systems.
  • Diverse tooling: The need to master multiple tools for monitoring and incident response.
  • High cognitive load: The mental strain of managing various systems, impacting their ability to respond quickly and adhere to SLOs/SLAs.

VM Service in VCF: Bringing VMs into the Kubernetes Fold ✨

Treating virtual machines as first-class resources in Kubernetes via VM service in VCF offers significant enhancements:

  • Declarative Infrastructure for VMs: Just like Kubernetes, you describe the desired state of your VM, and the platform ensures it. Instead of manual provisioning, teams define VM specifications using Kubernetes manifests.
  • Automation & Repeatability: This dramatically improves automation for VM workloads, just like containers.
  • GitOps Ready: Enables GitOps-style workflows, where infrastructure definitions can be version-controlled alongside application code. This makes infrastructure predictable, auditable, and automatable at scale.
  • DevOps Practices: Supports continuous delivery and integration, fostering innovation and adaptability.

Simplifying Operations: Collapsing Operational Silos 🌐

One of VM service’s most significant advantages is its ability to collapse operational silos.

  • The Old Way: SREs might monitor containers with Kubernetes tools and troubleshoot VMs with vSphere or other virtualization dashboards.
  • The VM Service Way: Both workload types are observed and operated through the Kubernetes control plane. This unified interface improves operational efficiency and significantly accelerates incident response. Engineers interact with Kubernetes, regardless of the workload type.
  • Standardized Automation & Policies: Teams can apply the same automation, policies, and operational playbooks across containers and VMs. Robust security features also ensure data integrity and compliance.

The Power of a Unified Control Plane: Beyond Convenience 🦾

A unified control plane offers far more than just convenience; it fundamentally improves reliability.

  • Standardized Observability: Alerts become consistent, capacity planning becomes simple, and failure isolation becomes easy. Metrics follow the same patterns, and incident response workflows become predictable.
  • Holistic Capacity Planning: SREs can reason about infrastructure holistically, leading to more efficient resource allocation, consumption, and better forecasting.
  • Reduced Total Cost of Ownership: Easier to understand and manage.
  • Proactive vs. Reactive: Shifts organizations from a reactive firefighting model to a more proactive, reliability-based model.
  • Enhanced Failure Isolation: Reduces the impact of incidents and helps manage mixed workloads effectively.

From Fragmented Tooling to a Unified Ecosystem ♻️

Before VM service, hybrid platforms often suffered from fragmented tooling, inconsistent automation, and noisy alerting systems. Each infrastructure type demanded its own life cycle processes and expertise.

VM service standardizes these workflows by bringing VM life cycle management into Kubernetes. The impact for SRE teams is phenomenal:

  • Fewer tools to maintain.
  • Lesser context switching during incidents.
  • Faster Mean Time To Resolution (MTTR).

Teams can now operate a single, consistent infrastructure platform centered around Kubernetes.

Reliability Lessons Learned: Cloud-Native Practices for Legacy Workloads 🌱

VM service also enables reliability practices traditionally associated with cloud-native platforms to be applied to VMs.

  • Programmatic Policy Enforcement: Policies for rollout, safety, resource isolation, and security can be enforced programmatically through Kubernetes.
  • Safe Rollout Strategies: Similar to container-based workloads, safe rollout strategies can be implemented.
  • Improved Failure Isolation: Kubernetes scheduling and policy enforcement can limit the blast radius of failures.

Tailoring Observability for Hybrid Environments: Unified Telemetry 📡

  • The Challenge: In many hybrid environments, telemetry from VMs and containers flows through completely different monitoring pipelines, making correlation during incidents extremely difficult.
  • VM Service’s Differentiator: Unified telemetry collection across both workload types.
  • The Benefit: SRE teams can correlate signals (like CPU pressure, network anomalies, or latency spikes) across the entire platform holistically. This allows for earlier identification of systemic issues, reduced alert noise, and a single operational view.
  • Adaptability & Security: Businesses can adapt quickly to changing demands, and security is enhanced by comprehensive visibility into potential vulnerabilities.

Incremental Modernization, Not Disruptive Transformation 💡

The final takeaway from VM service is that it enables incremental modernization.

  • Adopt VM Service with VCF: Treat VMs as Kubernetes resources alongside your container-based workloads.
  • No Need to Containerize Everything: Organizations don’t need to immediately containerize every legacy application to adopt cloud-native operational practices.
  • Unified Platform: Existing VM-based workloads come under Kubernetes governance, while newer services continue running in containers, creating a unified platform that supports both modernization and stability.
  • Simpler Operational Model: Supports automation, reliable engineering practices, and scalability without forcing teams to rewrite entire critical legacy application sets.

Thank you for joining this insightful discussion! If you have further questions, don’t hesitate to reach out.

Appendix