Pulumi: Program the Infrastructure with an Actual Programming Language
Welcome to The New Stack Context, a podcast where we discuss the latest news and perspectives in the world of cloud native computing. For this week’s episode, we spoke with Joe Duffy, founder and CEO of Pulumi, and Sophia Parafina, Pulumi’s technical marketing manager. In this convo, we delve into the recent Pulumi 2.0 release, which allows teams to reuse code, apply policies and do integration testing of infrastructure the same way they do for application development, a concept known as “architecture as code.”
TNS editorial and marketing director Libby Clark hosted this episode, alongside founder and TNS publisher Alex Williams and TNS managing editor Joab Jackson.
Episode 114: Program the Infrastructure with an Actual Programming Language
We covered the 1.0 release of Pulumi last September, revealing that Pulumi offers YAML-free infrastructure management. It provides the ability to use standard programming languages such as JavaScript, TypeScript, Python, and Go to deploy and manage IT resources. The new releases expand this platform by providing a way to enforce best practices around security, compliance, and cost management.
There are many advantages to programming your infrastructure using the same languages you use for your applications, Duffy pointed out in the podcast.
“You’ve got classes, you’ve got functions, you’ve got package managers; All the great things we’ve learned over the past several decades in application development out can be applied to infrastructure development as well,” he said.
He provided some examples: “If you’re setting up a network in AWS, you’re probably going to want to do a subnet for each availability zone just because you want to spread traffic appropriately. Well, okay, so [with] most YAML tools, what you do is you copy and paste the same 50 lines of YAML that set up a single subnet in one availability zone, you copy and paste that N times. In a language, you got a FOR loop. You say, for each availability zone ‘do this,'” he said. “If your development environments are different from production for some reason — maybe the number of nodes is different or you need to mock some services because they’re too expensive to run in development — Well, now you’ve got IF statements.”
Then, later in the podcast, we discuss a number of exciting podcasts and news stories on The New Stack. We delved into the developer success of Glitch. We ponder the disconnect between developers and security professionals. We discuss the importance of good telemetry to keeping microservices running. And we marvel at the big changes that both eBDF and io_uring will bring to Linux programming.
