Greatest Hits Archives - Software Engineering Daily · Greatest Hits Archives - Software Engineering Daily

Render: High Level Cloud with Anurag Goel

·1 hr 13 min·5 clips

Crestle's single-click Jupyter deploy backed by a GPU exposed how painful deployment was.

1. Greatest Hits Archives - Software Engineering Daily centers on Render: High Level Cloud with Anurag Goel and the layer 1 versus layer 2 cloud split. 2. Anurag Goel, founder and CEO of Render, is the guest, and the Software Engineering Daily host uses his Stripe history to test the company's claims. 3. The episode asks what Render is for: a developer-focused layer 2 cloud that hides Kubernetes, VMs, SSL, and deployment setup. 4. Goel traces the cloud back to AWS's 2006 S3 launch and says S3 still exists today. 5. He says Heroku launched in 2007 on top of AWS and became the first popular layer 2 cloud provider. 6. He contrasts AWS, Google Cloud, Azure, and DigitalOcean as commodity layer 1 infrastructure with exposed primitives like VMs, load balancers, and persistent disks. 7. He argues Kubernetes is a paradigm shift because it abstracts VMs away and reduces the need to SSH into machines. 8. He says Docker and Linux kernel advances pushed more software projects, internal tools, and SaaS applications into container deployment. 9. He describes Render's product as a narrow set of workflows: Node.js web servers, static sites, API backends, Postgres, and Docker containers. 10. He says Render connects to GitHub, detects build and start commands, and handles production rollout, subdomains, SSL, redirects, Brotli compression, and HTTP2. 11. He says his perspective on deployment changed after Stripe, where a large systems team hid deployment infrastructure from product developers. 12. He says Crestle, his single-click Jupyter deploy backed by a GPU and later sold to Dock.ai, was the first app that made deployment pain obvious to him. 13. He says Kubernetes learned late and still creates gotchas even for Render, which is why many developers are intimidated by it. 14. He says Render chose early use cases by looking at common developer stacks from surveys and day-to-day habits, including Express APIs, Django, Elasticsearch, and Kafka. 15. He spent about a year after leaving Stripe testing healthcare, real-time communications infrastructure, AI, and data science before Render won out. 16. He says Stripe was the wrong place to start a cloud provider because payments was its core competency and the company was not evolving in that direction. 17. He says Heroku breaks down as applications grow because 2GB of RAM costs $250 a month and the platform lacks built-in cron jobs, private services, persistent storage, HTTP2, and zero-downtime deploys. 18. He says Netlify is focused on static sites and frontend extras like form handling and authentication, while Zite moved toward serverless and away from long-running Rails apps. 19. He says Render is meant to answer the question of how to deploy a Node server on Netlify while avoiding the stagnation he sees in Heroku and the complexity he sees in AWS. 20. He says the ideal workflow is pushing code to GitHub from a browser or iPhone and watching the build logs until the app goes live.

As heard by us

A practical look at why developers might choose Render over raw cloud infrastructure.

Render's role in the cloud stack comes through a practical developer question: if AWS already provides the raw infrastructure, why use a Layer 2 platform at all?

Read the full review in PlayNext →

Why you'd press play

You want cloud deployment without turning every app into an AWS console scavenger hunt.

Read the full recommendation in PlayNext →

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from Greatest Hits Archives - Software Engineering Daily