Sep 2024
|
Paper
Morpheus Data Reference Architecture for 7.x
By Jason Bloomberg
Principal Analyst: Intellyx
The common theme across the first generation of cloud management platforms (CMPs) was command and control. The cloud operations (CloudOps) team used CMPs to keep the lights on while sticking to the budget, turning various knobs as needed to ensure their organization’s cloud infrastructure kept humming along.
What’s good for CloudOps, however, isn’t necessarily best for developers. The old days when dev would throw code over the wall to ops are long gone. Today, developers require tooling that enables them to work hand-in-hand with ops as they develop and deploy the software so essential to the business.
First-generation CMPs, however, fall short. Such platforms provide a central control point and source of truth for CloudOps but lack the features that support the developer experience necessary for modern software. Additionally, the increasingly heterogeneous nature of IT has meant hypervisor or cloud-specific tools have driven the need for a more agnostic approach than these early CMPs provided.
The opposite extreme from CMP’s ops-focused command and control are the loose assemblages of development and CI/CD tools we call the DevOps toolchain.
DevOps toolchains vary from one development shop to another, and often within large shops as well from project to project. Take some Git-based tool, add one or more infrastructure-as-code products, mix in some value stream management, and top it off with deployment tools. Voila! You have a DevOps toolchain.
These toolchains have one primary benefit: flexibility. Members of the development team can pick and choose the tools they want to use, ideally empowering them to achieve greater levels of productivity and job satisfaction.
However, such flexibility comes at a cost. Toolchains can be brittle, expensive to maintain, and lack the governance policy that enterprises demand.
Organizations thus face a dilemma. Legacy CMPs offer command and control without developer flexibility, while DevOps toolchains bring flexibility to the table at the expense of increased complexity and risk. The solution is somewhere in the middle: platform engineering.
Platform engineering seeks to balance two conflicting priorities: the self-service capabilities that give developers the flexibility they require while supporting the automated operations and rigor necessary to meet the organization’s cloud management needs.
The goal of a platform engineering team is to enable product and development groups to move faster via an internal developer platform (IDP) that provides a curated set of tools, capabilities, policies, and processes as a single, coordinated framework architected for self-service.
Platform engineering is the discipline of building such a framework that supports the DevOps toolchain requirements while also enabling the enforcement of policy controls in a way that is sustainable for the organization.
Such practices, however, should never be written in stone. Instead, one of the most important goals of the platform engineering effort is to make the best approach the easiest and most intuitive approach, without actually mandating it.
A common metaphor for describing this approach to platform engineering is a ‘golden path’ or a ‘paved road.’ There will always be other paths to the goal of deployed, quality software, but one road is the paved one, and hence the easiest and best-used path.
IDPs are still a relatively new innovation, and as a result, the organizations that have been implementing them have largely built them in-house. Given the complexities involved, such a challenge is not for the faint of heart. Only the most mature software development shops are up to the task.
Fortunately, there are unified platforms like Morpheus that provide the basis for an IDP by offering next-generation CMP capabilities while also supporting DevOps toolchains that meet various development team requirements.
Leveraging an orchestration framework to support platform engineering efforts addresses do-it-yourself IDP challenges while also supporting day-2 operations and service management needs of IT. In other words, this approach extends platform engineering to encompass platform operations.
Morpheus offers all the features of a modern CMP, but at its core, it is a self-service developer platform that was designed by (and for) platform engineering and operations teams.
In particular, it offers self-service provisioning of VMs, containers, clusters, and apps into any public or private cloud with all the guardrails that enterprise ops, security, and finance teams typically require.
Morpheus also supports a fully customizable instance catalog, facilitating the on-demand deployment of operating systems, databases, and other common infrastructure components into any environment, from bare metal to virtualization stacks, to Kubernetes and even cloud-managed services. Catalog items can also represent triggers for automation and infrastructure-as-code execution.
Perhaps most importantly, Morpheus offers over 90 built-in integrations with numerous CloudOps and DevOps tools, enabling organizations to quickly implement their own DevOps toolchains while simultaneously supporting self-service CloudOps in a way that is less brittle than hand-crafted home-grown platforms.
While people have called Morpheus a CMP, it’s actually much more. It is a next-generation framework for supporting platform engineering and operations teams, enabling organizations to build IDPs and unify cross-cloud CloudOps without having to start from scratch.
The techie’s most important tool for dealing with overweening complexity is abstraction ‘ and building full-lifecycle abstractions is at the core of Morpheus.
The platform supports each step in the lifecycle ‘ control, provision, connect, configure, deploy, operate, and scale ‘ in spite of the fact that each step involves many different pieces of technology and various orchestrations and workflows.
Such abstraction-driven simplicity is inherently opinionated, favoring particular ways of doing things that form the ‘paved road’ so essential to the operation of an IDP.
The flexibility to veer from the road is always available ‘ abstractions only hide complexity rather than eliminating it ‘ but with the help of Morpheus, the development and operations teams are better able to achieve the proper balance between the control of a CMP and the flexibility of a DevOps toolchain.
Copyright Intellyx LLC. Morpheus is an Intellyx customer. Intellyx retains final editorial control of this article. No AI was used in the production of this article.
