Guest: Rob Hirschfeld (LinkedIn, Twitter)
Company: RackN (Twitter)
Show: Let’s Talk

According to Rob Hirschfeld, CEO of RackN, Infrastructure Pipeline is a connected set of principles that run to the end, in order to configure infrastructure. Some even consider it to be CI/CD for infrastructure.

Infrastructure Pipeline fits into the Infrastructure as Code (IaC) by making it easier to connect 10, 15, or 20 different systems together and being able to automate that process, or (as Hirshfeld describes) “sort of coupling together or gluing together all of the other infrastructure automation pieces that they had to do from an end-to-end perspective, that is what was game changing.”

Another important thing Infrastructure Pipeline adds is the ability to use the same components and pipelines over and over by simply adding extensions so they can map to a specific environment to easily reuse it.

As to what projects are available to help with IaP, Hirschfeld mentions Digital Rebar, Terraform, and Ansible.  Hirschfeld also addresses the always-important idea of best practices when he says, “This is really about connecting systems together and I think one of the common mistakes people make in looking at an infrastructure pipeline, just like we used to do with CI/CD pipelines, is assume that you have to have an end-to-end process to get started.”

Hirschfeld continues, “Really what we’re talking about doing is connecting together more and more things so that an infrastructure pipeline could start in the middle of your process.” Hirschfeld also has this bit of advice to add, “People really shouldn’t think that they’re not succeeding at an infrastructure pipeline if they only have two components automated; they should look at it as, ‘How do I improve it?'” He then concludes, “And that act in itself is a big lift because you might find that adjacent teams aren’t talking the same language or using the same tools, or if they’re not, they don’t have to use the same tools, but they might not be handing off information to each other in ways that are smooth or designed to be adaptable.”

The summary of the show is written by Jack Wallen

Topics we covered in this show include:

• Intro to Infrastructure Pipeline
• So if you look at IaC, where does the pipeline come into the picture?
• Can you also kind of elaborate the importance of using Infrastructure Pipeline?
• Is this Infrastructure Pipeline a concept, a process, or is it also a stack of technology?
• What kind of projects are there to help with the Infrastructure Pipeline?
• How much discussion are you seeing already going on around in future plant, or is this concept relatively new?
• If somebody wants to get started with an Infrastructure Pipeline, what resources are there? What is the best place to look at? Can you share some best practices too?

[expander_maker]

Swapnil Bhartiya: Hi, this is host, Swapnil Bhartiya and welcome TFiR: Let’s Talk. And today we have with us once again, Rob Hirschfeld, co-founder and CEO of RackN. And today we are going to talk about infrastructure pipeline. Rob, first of all, it’s great to have you back on the show.

Rob Hirschfeld: It’s a pleasure to be here. Excited to talk about something as exciting revolutionary as infrastructure pipelines.

Swapnil Bhartiya: Yeah, what exactly it is?

Rob Hirschfeld: So an infrastructure pipeline is a connected set of processes that run end to end, to configure infrastructure. From its configuration or a request for it, all the way through to it working and then importantly through day two. A lot of people like to think of it like a CICD pipeline for code, but for infrastructure. And the consequences and impacts are profound. People are talking about value stream mapping and how we actually connect together a whole bunch of silos or repeat patterns across an organization, or potentially across the industry in the same ways that CICD pipelines have revolutionized how code is delivered.

Swapnil Bhartiya: Right. So if you look at infrastructure as code, where does the pipeline come into the picture?

Rob Hirschfeld: So infrastructure as code is part of the modularity and the way that you can build a pipeline together, but infrastructure as code has really missed the idea of the end to end operations. It would be like saying a build system or a linting system was sufficient to build a CICD pipeline. Infrastructure as code provides parts of the system and a philosophy that makes things easier to connect together, but it isn’t the end to end connected pipeline that actually joins all of those operations together. Both are important in these concepts and we’ve really seen the pipeline concept as elevating infrastructure as code to the next level.

Swapnil Bhartiya: Can you also kind of elaborate the importance of using infrastructure pipeline? You did touch upon what it does, but if you can talk about the importance of why people should care about it.

Rob Hirschfeld: I’d be happy to. We have customers deployed at tens of thousands of machines with global scale that can do that only because they’ve been using infrastructure as code processes that we’ve been working hard to enable in RackN. And in those processes, they literally can define everything about their data center as infrastructure as code completely defined configurations. And then as machines are delivered, or as they’re reset, they follow an infrastructure pipeline through every step of the process. From discovery and inventory [inaudible 00:02:33], operating system install, post operating system configuration security, even registration into their in inventory databases and pulling down credentials and checking in with there are existing systems of record. So if you think about it, that end to end process that connects in to 10 or 15 or 20 different systems, being able to do that as its own automation, sort of coupling together or gluing together all of the other infrastructure automation pieces that they had to do from an end to end perspective, that is what was game changing.

And then the thing that’s been even more exciting for me about infrastructure pipelines is because it is a standardized process that they’re not the only customers using that infrastructure pipeline. For us, our other customers in different industries, like our banking pipeline and our telco pipeline, and our media pipeline, oil and gas pipelines, they’re actually using the same pipelines or the same components of pipelines over and over again. So rather than having to figure out how to do all this stuff, they literally lift the pipeline, they add some extensions to make it map to their environment and then they keep using it. And as we improve it, that same pipeline can then be replicated and improved really across the industry. Its dramatic improvements.

Swapnil Bhartiya: Is this infrastructure pipeline a concept, a process, or it also a stack of technology. When you say it can be replicated-

Rob Hirschfeld: It has elements of all of it. There is a missing automation layer that pipelines are filling just like CICD, the pipeline layer itself was this added piece that we had to bring in to bring all these tools together. So it is a stack of technologies because you use the tools that are already do the individual configuration or provisioning, or integrations or security, or monitoring. You don’t replace those tools, but we do need to connect them together. So there is a stack component, but we really don’t want to underestimate the process and the people pieces here, because the other thing that we’re doing and just like CICD pipelines, we’re connecting together parts of an organization or silos that hadn’t communicated before.

And so in CICD pipelines, we talk about it a lot about DevSecOps or shifting left, where we’re including other groups into the process. Infrastructure pipelines are exactly the same where we can bring security into the infrastructure automation process in new ways, or connect teams together that hadn’t been able to be connected into a seamless process. That’s important for self-service, it’s important for edge, it’s important for zero touch automation. So there’s a lot of interesting things that come together from the perspective at an infrastructure pipeline.

Swapnil Bhartiya: Excellent. Now, since you compare the CICD pipeline, we do have a lot of project like Spinnaker is a very good example there, also the is [inaudible 00:05:25] foundation there. So if we look at, in three pipeline, can you name some projects that are there to help that?

Rob Hirschfeld: What we’ve done with digital rebar, we have not found a lot of other people in industry replicating quite that. We see a little bit with GitOps, where we’re using git to start a process, but really carrying it all the way through, we don’t see as much. We do see a lot of Terraform integrations and we’ll integrate with Terraform or Ansible and we’ll integrate with Ansible. It’s this new idea of connecting all of the stuff together that is its own automation layer that we haven’t seen as a real connected piece. Sometimes what you will see is something like a Azure DevOps, where in Azure, they can connect all the pieces of their cloud together. Those are emergent tools and for a single vendor or a single platform that are starting to connect these together. But without the focus on reuse and infrastructure as code portability, you don’t get the benefit of this cross industry or internal inside of cross team opportunity for collaboration. And that is one of the important parts of how infrastructure pipelines are different.

Swapnil Bhartiya: How much discussion are you seeing already going on around in future plant, or is a concept which you related with new.

Rob Hirschfeld: We see a lot of it from a theoretical perspective. So we see, ThoughtWorks has been doing some leadership here. Gartner has been doing some leadership here, where they’re talking about value stream mapping or how a continuous automation process, or continuous infrastructure automation process needs to be done. The challenge that we see is that those pictures and the way they describe it is, “I have a whole bunch of silos and I need to connect them together to build a pipeline.” And so it’s been described very aspirationally in the industry, this is why it’s not a new concept to have an infrastructure pipeline. The thing that’s been missing is being able to have platforms and tools that actually do the work for you to encourage the behavior, to break down those silos, and connect an end to end automation process together.

Swapnil Bhartiya: If somebody wants to get started with infrastructure pipeline, first of all, what resources are there? What are the best place to look at? Also, there are always pitfalls, mistakes. And also if you can call, best practices.

Rob Hirschfeld: With RackN, we are actively helping customers build infrastructure pipelines. So we’re always happy to talk to people and help them start that journey. But you don’t need to talk to us to start the journey. This is really about connecting systems together and I think one of the common mistakes people make in looking at an infrastructure pipeline, just like we used to do with CICD pipelines is assume that you have to have an end to end process to get started. Really what we’re talking about doing is connecting together more and more things so that infrastructure pipeline could start in the middle of your process. It could start by connecting Terraform and Ansible together really well in your processes so you’re not dealing with two tools, you’re dealing with one connected system. Or managing your provisioning automation in a more seamless way.

Those are really ways that we can get started in building an infrastructure pipeline and then expand it. So people really shouldn’t think that they’re not succeeding at an infrastructure pipeline if they only have two components automated, they should look at it as, “How do I improve it?” And that act in itself is a big lift because you might find that adjacent teams aren’t talking the same language or using the same tools, or if they’re not, they don’t have to use the same tools, but they might not be handing off information to each other in ways that are smooth or designed to be adaptable. And there’s a lot of things that you can do to just get a rolling start with that perspective. With digital rebar, specifically, we encourage individual teams to get started and get their processes running and then look for how do I win to the second tee? And that’s actually a really important thing to do is just to think through and choose tools that are designed to build into pipelines and not basically create more silos.

Swapnil Bhartiya: Rob, once again, thank you so much for taking time out today and talk about infrastructure pipelines. And as usual I look forward to our next recording. Thank you.

Rob Hirschfeld: Thank you.

[/expander_maker]