Platform Engineering: DevOps Rebrand or a Meaningful Step Forward?
In the ever-changing landscape of software development, Platform Engineering has emerged as a new function within engineering teams. 20,000 people now have this title on LinkedIn, compared to 220,000 in closely adjacent DevOps roles. Platform Engineering represents the formalization of internal processes and tools engineering teams often build in-house to enable their developers. We at Next47 look forward to working with founders productizing the best practices from successful teams and bringing them to the broader software community.
Platform Engineers administer tools, targeting points of interaction between developers and DevOps that historically have not scaled well. These tools help developers deploy software in a way that is more self-serve, automated, and adherent to DevOps best practices. They also free up DevOps teams from repetitive, incremental work that creates bottlenecks. Platform Engineering done right means providing the “golden path” to production for developers.
We can draw a comparison between Platform Engineering and the original DevOps movement. Both aimed to change the workflow, responsibilities, and org structure of engineering teams to improve collaboration.
There were varying interpretations of the DevOps movement, and varying degrees of success where it was implemented. In some cases, establishing more autonomous, project-focused teams also led to a sprawl of different tools and workflows within an organization. In other cases, teams did not substantially change their org structure, and the DevOps role did not represent a major shift from the Ops role. These mixed results have led to some skepticism around the promises of Platform Engineering. Some dismiss Platform Engineering as little more than another title change for DevOps professionals. Another common objection stems from the dislike of opinionated abstraction layers. There will always be some engineers that prefer working with the flexibility and complexity of native, underlying systems like Kubernetes. While these objections are valid, the Platform Engineering framework can be tailored to benefit a variety of engineering profiles, from new hires to highly skilled full-stack developers.
We believe we are just in the early innings of Platform Engineering, and we’re seeing software companies experimenting to find the right tools and internal practices. In our market map, we aim to answer what exactly this “Platform” consists of. We have identified six categories of startups that closely align with Platform Engineering principles, and represent areas where developers are beginning to take more ownership of DevOps responsibilities. These categories are: Internal Developer Portals, Infrastructure as Code, Platform Orchestration, Environments, Deployment, and Incident Management.
One important aspect of Platform Engineering is the concept of Internal Developer Portals, which provide a modular SaaS interface for developers.
A developer portal typically includes the following key features to optimize Platform Engineering:
- A service catalog: This provides developers with an overview of all projects, services, deployment status, documentation, ownership, on-call schedules, and incidents.
- Repository scaffolding/project generation tool: For example, a tool like cookiecutter (https://github.com/cookiecutter/cookiecutter) allows developers to create new services directly within the portal.
- Custom toolchain integration: Integrating your chosen toolchain into the developer portal offers a unified one-stop experience. Possible integrations include Kubernetes, CI/CD service providers, and a secrets manager.
An excellent example of such a portal in the open-source realm is Spotify’s “Backstage“. This tool offers a unified dashboard for infrastructure tooling, services, and documentation, making it a go-to solution for many organizations seeking a centralized hub for their developers.
IaC transforms infrastructure provisioning into a code-based process, ensuring environments are reproducible, scalable, and maintainable. Paired with CI/CD, this allows for seamless, automated pipelines that take code from development to deployment with minimal manual intervention.
Looking forward, we can anticipate a deeper integration of AI and machine learning within IaC and CI/CD processes, enabling more intelligent automation, predictive failures detection, and even automatic code optimizations.
Another crucial element is Platform Orchestration, which simplifies the transition between local development environments and cloud-based infrastructures. Firstly, it acts as a bridge, integrating new software with legacy systems for seamless adaptability. Secondly, it ensures efficient operations by automating scalability. This streamlined scaling pairs with the ability to codify and automate processes, letting the human workforce concentrate on designing and delivering services.
At scale, hosting dev environments can cost millions per year. Some in the environments category are focused on minimizing collisions between developers in these shared environments, reducing cost, and improving developer productivity. Many offer isolated, ephemeral environments that spin up and down on a pull request basis. There are a variety of approaches. Some focus on local development, others are in the cloud or the browser. These vendors are evaluated on their ability to simulate production resources accurately, quickly and efficiently.
Deployment & Hosting
Deployment and hosting enables developers to directly connect their code to the infrastructure, ensuring efficient deployment while providing monitoring capabilities and multi-region support.
When a developer pushes code, they’re not just deploying; they’re also ensuring that the code is hosted in optimized environments, ready for user access. These days companies like Vercel, Railway, and Koyeb help you deploy things worry-free with a cloud-native solution.
In the multifaceted world of Platform Engineering, where numerous services often operate in tandem, incidents are, unfortunately, a given. Whether it’s a service outage, a security breach, or an unexpected spike in traffic, these events can disrupt the seamless operation of applications. This is where incident management tools step in.
The right tool can help shorten the downtime and offer insight into the root cause. Learning from incidents shouldn’t be restricted just to post-mortem meetings, but should also be a pervasive aspect of a company’s culture. One of the leaders in this category is Incident.io, which helps you approach your incident with prominent structure and unify data across the company.
A Guide to Disruption
We identified a few important considerations for startups in the Platform Engineering space:
- Pricing Power: There are not enough platform engineers today to price based on that headcount alone. Platform Engineering startups should target DevOps budget as well. Looking at the 220,000 DevOps professionals on LinkedIn, $4,000/year/DevOps headcount is an ideal long-term target to reach a $1,000,000,000 addressable market.
- Collaborative Hotspots: Founders should identify areas where developers face bottlenecks due to limited DevOps capacity. Our landscape identifies a number of these areas, and startups can aim to address multiple categories.
- ROI & Efficiency: Prioritize speed for smaller teams, and control and consistency for enterprise customers. Some companies like Rancher have successfully targeted traditional enterprises with large engineering departments, while others like Vercel have targeted smaller teams and individuals.
Platform Engineering promotes a future where developers can autonomously provision new applications without being infrastructure experts. To achieve this vision, it will require upfront effort from platform engineers and DevOps teams to define best practices that empower dev teams. We expect the Platform Engineering landscape to grow dramatically, and are eager to hear your thoughts on where this industry is headed. To get in touch, find us on twitter @coltondempsey and @whoiskatrin.