Cloud Native

What is Gartner-defined devirtualization and how it’s RackN’s backyard | Rob Hirschfeld

0

Devirtualization signifies a transition from traditional virtualization to direct hardware utilization, aiming to enhance IT efficiency and performance and RackN aims to support customers through this transition. In this episode of TFiR: Let’s Talk, Rob Hirschfeld, Co-Founder and CEO of RackN, delves into the primary factors driving devirtualization, the challenges companies are facing, and how RackN’s solutions are facilitating the move from virtualization to devirtualization. Hirschfeld explains, “We help companies that are really looking for improved choice and optionality in bare metal infrastructure.”

Why Devirtualization is Becoming a Key Trend

  • Hirschfeld notes that Gartner has marked devirtualization as a new trend at the start of the hype cycle, indicating early industry interest in moving beyond traditional virtualization.
  • Hirschfeld points out that most companies depend on virtualization, particularly VMware, but this preference is being reconsidered as infrastructure needs evolve.
  • Hirschfeld questions the notion that all infrastructure needs virtualization, noting that Kubernetes and containerization offer more flexible and efficient management, leading to a shift from traditional virtualization.

What are the Challenges of Virtualization?

  • Hirschfeld describes virtualization as the foundational step in infrastructure deployment, often associated with hyper-converged infrastructure. This method has been crucial in consolidating and optimizing resources within a single system.
  • Virtualization gained popularity due to its ability to centralize control, facilitate integration through Application Programming Interfaces (APIs), and improve efficiency via resource oversubscription. These benefits have historically made virtualization a key component in infrastructure management.
  • Hirschfeld notes that emerging technologies and changing cost dynamics are reshaping infrastructure strategies, sparking discussions on devirtualization. Organizations are reconsidering their dependence on traditional virtualization approaches.

What are the Key Drivers Behind Devirtualization Trends?

  • Hirschfeld identifies the VMware acquisition by Broadcom as a major factor driving devirtualization, noting that it has increased costs and disrupted the existing market dynamics, prompting organizations to reconsider their reliance on VMware.
  • Hirschfeld highlights Kubernetes as a leading platform, which has led companies to reevaluate the necessity of maintaining a virtualization layer, as Kubernetes offers more direct control over infrastructure.
  • Hirschfeld discusses the challenges of virtualizing graphics processing units (GPUs), particularly for inference workloads, and notes a trend towards using bare metal for these specific tasks to enhance performance and efficiency.

Economic and Power Dynamics Shaping Data Center Strategies

  • Hirschfeld notes that the increasing power constraints faced by data centers are pushing companies to rethink and redesign their architectural strategies to prioritize energy efficiency.
  • Hirschfeld draws a parallel between devirtualization and the hybrid cloud model, stressing the value of having diverse options and the flexibility to select the most suitable infrastructure solution for different workloads.
  • Devirtualization offers enhanced adaptability and cost-efficiency, especially when integrated with Kubernetes and containerized environments, leading to better resource utilization.
  • Hirschfeld explains that devirtualization allows organizations to tailor their infrastructure management more closely to their specific needs, providing greater control and the ability to optimize performance for targeted applications.

Understanding Devirtualization Choices and Their Flexibility

  • Hirschfeld outlines options in devirtualization, including improved supply chain management and cost-effective systems, offering companies more flexibility and efficiency in deploying IT resources.
  • Hirschfeld discusses the benefits of running Kubernetes or containerized systems directly on bare metal, which improves performance and resource use, especially for efficiency-critical workloads.
  • Hirschfeld highlights the flexibility of using smaller, cost-effective machines managed by Kubernetes, allowing companies to scale efficiently without overcommitting resources.
  • Hirschfeld emphasizes that devirtualization is not necessary for all workloads, as some still benefit from virtualized environments, particularly those needing stable Application Programming Interface (APIs) and high uptime.

Managing the Transition from Virtualization to Devirtualization

  • Hirschfeld explains that devirtualization requires a significant architectural shift, challenging long-standing assumptions in IT purchasing and prompting organizations to rethink their traditional approaches.
  • The term “devirtualized infrastructure” is gaining traction within the industry, reflecting a growing interest in moving away from fully virtualized environments.
  • Viewing infrastructure as a general-purpose asset is crucial, as it can enhance utilization rates and improve Return on Investment (ROI) across various workloads.

The Role of Repatriation in Devirtualization Strategies

  • Hirschfeld explains that repatriation involves transitioning workloads from the cloud back to self-managed infrastructure, allowing organizations to regain control over their resources.
  • Hirschfeld emphasizes the critical role of a robust bare metal foundation in ensuring successful and efficient repatriation.
  • Hirschfeld notes that repatriation is not a regression but a strategic move towards enhanced control and improved cost management.

Managing the Transition from Virtualization to Devirtualization

  • Hirschfeld describes the transition from virtualization to devirtualization as a stepwise process, requiring careful planning and gradual implementation. This incremental approach helps organizations manage the complexities and risks associated with moving away from traditional virtualization.
  • The importance of piloting and proof-of-concept projects is emphasized as crucial for successful devirtualization. These projects allow organizations to test and refine their strategies on a smaller scale, ensuring that the transition is effective and aligns with their operational needs.
  • Mature workloads, with well-understood and stable resource demands, are noted as particularly good candidates for repatriation. These workloads benefit from being moved back to self-managed infrastructure, where they can achieve better control and cost efficiency.

How RackN is Leading in Bare Metal Automation

  • Hirschfeld explains that RackN specializes in bare metal automation and infrastructure management, focusing on streamlining and optimizing the deployment and operation of physical servers. This specialization allows RackN to offer advanced solutions for managing self-hosted infrastructure efficiently.
  • The importance of making bare metal automation repeatable and easy is highlighted, as it ensures consistent and reliable infrastructure management. Simplified automation processes help organizations scale their operations without the complexity often associated with managing physical hardware.
  • Rack’s platform is noted for embedding best practices and extensive expertise directly into its solutions. Hirschfeld explains how this integration helps customers operate their infrastructure more effectively.

The Growing Market for Bare Metal and Data Centers

  • Hirschfeld notes that rising interest in bare metal and data centers is largely driven by the surge in AI technologies. This shift is prompting companies to reassess the value of owning and managing their own infrastructure to meet the growing demands of AI workloads.
  • The growing importance of owning and controlling infrastructure is becoming clear to many organizations. Direct control over data centers and hardware is seen as beneficial for improving performance, security, and cost efficiency.
  • There is a noticeable increase in interest in hardware original equipment manufacturers (OEMs), AI technologies, data center real estate, and edge computing. This reflects a broader trend where companies are diversifying

Guest: Rob Hirschfeld (LinkedIn)
Company: RackN (Twitter)
Show: Let’s Talk

This summary was written by Emily Nicholls.

UAE is building an open source GenAI model called Falcon

Previous article

Linux Foundation Research highlights open source solutions for energy interoperability in Canada

Next article