Why Bare Metal Can’t Be Operated Like the Cloud (Yet)

There is a problem with bare metal management in our industry. We lack an end-to-end solution that provides a seamless managed operating model that automates the provisioning, deployment, and management of bare metal servers located anywhere. Whether it is a private cloud running Kubernetes or Virtualization in an on-premises data center, or a couple of servers running at the edge of the network, enterprises are looking to solve the problem of automating and orchestrating the lifecycle of bare metal servers in a way that is as seamless as operating the cloud.

Changing business needs also demand agility.  Bare metal should accelerate your business, not slow it down like it does today.  What does a bare metal with cloud agility look like?

  • Bare metal provisioning and management is fully automated and deeply integrated into the overall lifecycle of applications.
  • Application developers are able to repeatedly build, test, and deploy newer versions of their applications, no matter where they are running.
  • IT teams are able to upgrade OS versions and software hassle-free when security patches and bug fixes are rolled out.
  • IT teams are also able to instantly provision more bare metal servers on-demand when the business requires infrastructure to scale or when servers need to be re-provisioned for different use cases/workloads.

While there are solutions that address different parts of this problem, what’s missing is a unified modus operandi that extends cloud-like simplicity to physical servers.

The Need for Bare Metal Orchestration and Management

A number of our customers spanning a variety of industries and use cases are asking Platform9 about automated bare metal orchestration for many reasons: high performance, low latency, avoiding the hypervisor tax, improving utilization, better integration with the container, virtualization stacks, reducing manual effort, improving agility, and more. The bare metal use cases we most often run into include Retail, Telco and 5G Edge, Private Cloud, and High Performance Computing (HPC).

Retail Edge

According to Gartner:

“By 2022, more than 50% of enterprise data will be created and processed outside the data center or cloud”.

Salient dynamics for retail include:

  • Retail chains may have thousands of stores, and will likely provision commodity servers to support operations and applications that tend to their business needs.
  • Retail edge locations are often characterized by limited connectivity to the cloud or internet, a small set of servers, and office staff that may not be technically proficient enough to operate these servers.
  • Examples of such use-cases are video surveillance and analytics, IoT processing, Point of Sale, and applications that improve the efficiency and operations at the retail edge location.

Telco and 5G Edge

  • With the advent of 5G networks, telecommunications operators are evolving from legacy network infrastructure to using virtual Radio Access Networks (vRAN) and packet core software to enable the handling of large data sets produced from IoT sensors, high-definition video content, and virtual reality (VR) or augmented reality (AR) applications.
  • These applications and their associated network traffic require a more agile approach to software deployment and server provisioning at the 5G edge of the operators’ infrastructure.

Private Cloud

  • For a variety of reasons, including compliance, data privacy, higher performance, and infrastructure control, IT teams are transforming on-premise data centers into private clouds that deliver on the promises of a self-service operating model for consumers of IT and are based on the tenets of flexibility, agility and elasticity.
  • Provisioning bare metal servers is manual and tedious, consuming precious cycles from technical experts and impacting the elasticity of their private cloud infrastructure.

High-Performance Computing (HPC)

  • Organizations running resource-intensive big data, analytics, database and AI/ML applications are pushing their IT teams to provision and manage bare metal servers for these workloads to run on, so that they don’t have to endure the performance hit of virtualization.
  • These workloads span a spectrum of industries, like EDA, media and entertainment, oil and gas, financial services, and research labs.

Bare Metal Management today is Manual, Slow, Error-prone, and Tedious

End-to-end bare metal orchestration requires a number of steps, most of which are manual:

  • Provisioning bare metal servers that just expose an IPMI interface over a network
  • Deploying OS images on them, running applications on them
  • Updating those applications based on business needs
  • Upgrading software on the servers to keep them up-to-date with security patches and bug fixes
  • Guaranteeing availability of those servers in case there is an outage
  • Re-provisioning servers when there are performance glitches or other issues

The sheer number of manual steps involved, the complexity of prerequisite knowledge required, and the risks associated with server downtime make it difficult to manage and operate bare metal servers. This is especially true with commodity servers. Business agility is impacted when bare metal servers need to be manually provisioned, upgraded and scaled. Organizations today also need the ability to repeatedly deploy applications to meet the changing demands of their business. IT operators running large environments with combinations of bare metal, virtualization, and containers need a simpler, self-service operating model.

Organizations Still Seek an Automated End-to-end Bare Metal Orchestration Solution

While there are products that solve some of these challenges, we find that organizations are seeking a complete solution that enables seamless bare metal orchestration for commodity servers– no matter where they are located– while giving them the freedom to run any workload, on any substrate, without any lock-in. To name just two examples, public cloud providers with proprietary hardware make provisioning seamless, but are expensive and lock you into their technology and hardware stacks. Standalone bare metal orchestration tools will help you build a bare metal cloud, but don’t give you an easy way to integrate with the rest of your cloud.

We’ve been listening to our customers very carefully over the past year, to understand the gap in the end-to-end experience and the pain of bare metal orchestration described above. Platform9 has a history of effectively addressing these types of customer issues with solutions that provide the simplicity of the cloud operating model, high-performance, and the power of SaaS-based management to provide a white-glove, fully-managed service to offload your operational burden at scale. Stay tuned on how we are going to tackle this issue in a similar fashion!


You may also enjoy

Mastering the operational model challenge for distributed AI/ML infrastructure

By Kamesh Pemmaraju

The resurgence of OpenStack: Addressing the cloud conundrum

By Kamesh Pemmaraju

The browser you are using is outdated. For the best experience please download or update your browser to one of the following:

State of Kubernetes FinOps Survey – Win 13 prizes including a MacBook Air.Start Now