The Golden Path to Platform Engineering

The New Stack · Intermediate ·📄 Research Papers Explained ·2y ago

Key Takeaways

The video discusses the emergence and ascension of platform engineering, highlighting the role of Internal Developer Platforms (IDPs) and the concept of the Golden Path, with tools such as Backstage, Cortex, and Humanitec's orchestrator product.

Full Transcript

[Music] you're watching the new stack makers a podcast for people who develop deploy and manage at scale software for more information and articles about ATS scale Technologies please visit the new stack. now enjoy the show humanite Tech Powers Enterprise grade internal developer platforms their products let platform teams remove bottlenecks by letting them build golden paths for developers Enterprises use humanite Tech to standardize how they build and deliver software enabling true developer self-service this reduces lead time and slashes time to Market hello uh welcome to the latest edition of the new stack makers podcast today we're going to learn about internal developer platforms idps uh from the company that uh kind of started the uh uh the wave of discussion around idps uh first by uh uh introducing a concept uh uh platform engineering which we've been covering for a couple years now but a series of posts for the new stack provided by humanite Tech uh raised some interesting questions about how platform engineering answers questions and concerns around devops practices uh it was uh quite a uh uh quite a a controversial opinion and we'll include some of their early posts in the write up for this podcast but over time uh platform engineering uh kind of was refined into the concept of idps and uh internal developer Platforms in the way that uh the idea behind platform engineering was to make developers lives easier allow them to concentrate on the code itself and not all the tooling around the code and uh idps were a way of doing that and humanite Tech uh was the company uh that certainly helped Usher in the era of idps today we have the uh CEO of humanite tech uh with us today uh Casper Von grber Casper thank you for taking the time to talk oh thank you for having me all right so first we want to divide this interview up into two segments uh first we'll talk about idps what they are how they're different from platform as a service or Ides or any one of other things and then we want to learn about what human attch offers in this area and what makes them uh uh kind of stand out from the pack uh so let's let's jump right into it uh what is an internal developer platform well an internal developer platform is the sum of the Tekken tools that a platform engineering team puts together to build golden paths and help the developers consume things with a High degree of sanitization low cognitive load basically you know Paving the roads if you want um an internal developer platform is so to speak the result of the work of a platform engineering team you know it's the platform that they're building uh it's not a platform for externals it's a platform for internals it has um you several components we call them planes or to be more precise McKenzie calls them planes McKenzie has built a reference archit Ure for modern internal developer platforms um and they differentiate five planes the developer control plane where you would have things like a portal um you would have things like version control system so you know how does the user interact with the platform the second plane would be the integration and delivery plane uh CI registry platform orchestrator you have the resource plane you know those are your actual resources your Cloud resources your on-prem resources databases file storage clusters you have your observability plane and you have your security plane and you know that these this is how we roughly divide the areas of a of a platform how the platform then looks and how the IDP looks at the end is a little bit subject to you know what the platform engineering team builds and that depends on what the organization needs terrific yeah how does yeah can we diive a little bit further into that uh I mean how does look to the developer is this like a uh portal of some sort do I when I'm building an application how do I choose the components that I would uh use it becomes a platform if you have some sort of golden path and abstraction if you want you know you want to make you want to get away from the concept of just throwing everything at the uh developer like hey you know why don't you do everything in uh you know Terra form and um why don't you learn this and why don't you learn this and this you want to give them them a delightful way to consume this and that means it essentially if you want to simplify this the de developer should say hey I need a workload and my workload needs a database it should be a type postgress I need this specific configuration and then I need file search and you should be just sending this through the pipelines and that should then be matched by the defaults provided by the platform engineering team and the what you want should be running okay how you're consuming this is a little bit up to the individual preference of the user and what the platform team has configured for you so we're seeing some teams like to use a user interface and a portal some teams like to use code-based interactions like a you know workload specification score for instance some teams like CLI some teams like apis sometime some teams like chatbots it doesn't doesn't really matter what matters is you have a abstraction you have something where developers can you know order something and that abstract request gets gets fused with the Baseline configurations provided by the platform team and as a result you have the running um application which is following all of the standards observability Etc um right by default so it's the experience of a platform as a service but really Enterprise grade holistically thought through all of the different components that then uh make make a um platform does this limit uh developer Choice as for the tools or is that not as an important issue as some people make it to be no I think it's it's a very very important issue um it is so vital that we do not go back to the platform um age of 2010 basically where we we had platforms that basically hindered the developers and said like hey you need to use these defaults there are no other options and that is a bad bad uh thing to do we don't and it's incredibly unproductive we don't want to get back uh to to to those times which is why we want to provide sensible to Falls I call them golden paths Netflix calls them paved roads developers can stay on those paths if they do so they get certain slas they get certain guarantees but they are free to leave the path they are free to actually um you know go low level and configure the things in detail should the security posture allows and this is a very important point that we're making here good platform engineering is never about covering 100% of the use cases it's never about pleasing 100% of your developer POA um population it's about 8020 cover 80% with delightful experiences for those that want to go off the the golden path let them do that so this is uh a platform engineer is primarily in operations rather than development I take it they're the people who have to set up the infrastructure for the developers the platform engineering team is put together by a diverse set of groups and it really depends on what a team can afford if you want you like platform engineering and platforming shouldn't be that you know big role like that big dting thing that takes ages of to start you should be able to start very Scrappy with one two people for one or two teams now um if you have the the funds a platform engineering team should have a product manager because you should treat your platform as a product your platform engineering team should have people from Ops and infra your people your team should have developers Advocates you know those can be very large they can be one or two people site roles absolutely Ely possible but I'm a big fan of um you know making sure that there is a divice diverse set of voices and opinions when these platforms get built all right terrific terrific uh let's move on to humanite Tech how what was the origin story of the company how did you guys get started well we are platform Engineers by heart we've been doing it for for ages and you know we've just observed that people are building the same things over and over again if we look at platforms right we're basically where CI was 10 years ago everybody building their pipelines and Reinventing the wheel and then later you see okay well other specialized teams build certain components and humanite Tech is really about empowering the builders empowering platform Engineers to build fantastic platforms that really serve the need of the organization and we do that by offering different products different components that you can use to build these platforms we are own and we are you can use these things to then actually glue them together with what you have right now we are allowing you to integrate everything very rapidly in a modular way we have the um processes to help you do that we've just seen hundreds of different platform setups we have very battl tested ways to do that to guide you um and then you know most importantly we have uh the products to actually uh enable that for you uh terrific terrific uh yeah talk about the products a bit now uh so one one one uh uh thing we've been hearing quite a bit is backstage a platform that originated of Spotify uh is uh a popular choice do you interact with with backstage at all or or what is your relationship with backstage yeah so um backstage is definitely one of the user interfaces that you can put on top of a um internal developer platform um I would you know to stay neutral name a few others um cortex for instance Port is a option that's often looked at Backstage being the open source equivalent has benefits has downsides um the again a a something like this like a portal is a UI into a platform it and it's something yes we do integrate with um we also offer uh an own option in in uh from from from that perspective in that and uh it integrates very well with our other two core products the platform orchestrator and the orchestrator is basically about bringing together the abstract request of the developer whether that abstract comes from the portal CLI API or UI or code based uh doesn't really matter but it's about bringing that abstract request from the developer um then actually enriching this with the Baseline configurations provided by the platform team those Baseline configurations could be written in Terror form crossplane whatever and from the combination of those two abstract request and Baseline configuration the orchestrator then orchestrates everything the tools creates the configs prepares the deployment does the deployment itself hands over to um Aro CD for instance all of that and that allows you to get Dynamic configuration management um Resource Management environment management um cost control uh outbreak and governance so all of those features that you want to have in your platform that are usually super hard to get the orchestrator will get you that and will help you function as the glue of everything you're building uh around this so the orchestrator is the core uh core product uh is a core engine for human attex deployments yeah so the orchestrator absolutely is is is sitting at the center of a of a platform post CI uh and registry and it's taking the you know whatever comes from the portal whatever comes from our code based interaction method we have a product called score an open source product where developers can describe things as code it will take those abstract requests it will take the you know the um the image that flies in from the CI and it will look at this analyze it say ah okay this is the context this is what's requested looks like this is a workload I need a database file seage Etc what does the platform engineering team want me to do takes this creates app and infrastructure configurations on the Fly and executes this and it's doing this with every single deployment and that makes allows us to reach a very high degree of sanitization if you know baselines change you now want to have um grafana as an integration whatever the orchestrator can get that done and actually create everything nice nice so you're kind of the uh the kubernetes of the app development world I guess yeah let me ask you this what is I mean we've seen a lot of entrance in the space we've seen a lot of passes uh platform is service on the space rebranding themselves is IDP some from some large companies like VMware what is what is humanex uh Cutting Edge what's your secret sauce uh what is it that you do better in anybody else in this field yeah absolutely so human Tech is not about replacing anything human Tech is not your One Stop Shop platform as a service that wants to knock out everything else humanite Tech is really and I want to say this again empowering the platform Engineers to build great platforms we give you uh the building blocks you have to put them together and we are used by some of the biggest brands you're using every day to some of the more smaller teams we have a large scale and we are giving you battl tested components to build these platforms that are reliable that developers love and that fit the exact need of your organization and we do that better than anybody else all right fantastic fantastic uh we want to thank you guys not only for appearing on the podcast but you might notice some manitech uh sponsor notes in some of our stories uh they have provided us with the uh the funds to research both platform engineering and uh idps which are which turn into stories that you guys are reading so uh we thank uh humanite tech for uh uh allowing us to uh explore this area more uh and uh thank you for appearing on the show and thank you uh listeners for uh tuning in well thank you for having me thank you for the very fruitful collaboration and looking forward to the you know next months and years if you like this video please give us a thumbs up and if you'd like to see more videos like this you can always subscribe to our YouTube channel we're on all the major social media plat forms you can always find us at the new stack. we hope to see you [Music] soon

Original Description

In this latest edition of The New Stack podcast, we speak with Kaspar Von Grünberg, CEO of platform services provider Humanitec, about the sudden rise of platform engineering. Along with discussing the emergence and ascension of platform engineering in this episode, we also discuss the role that Humanitec plays in helping organizations establish platforms for developers, as well as Backstage, a popular open source internal developer platform that was developed by Spotify for its own developers. An IDP, our guest Kaspar Von Grünberg explained, is a standardized interface for developers to build applications using a golden path of vetted tools and libraries, allowing for a high degree of efficiency for both the developers themselves as well as the engineers who are supporting the developers. They can include an integration and delivery plane, a continuous integration registry, a platform orchestrator, observability tools and a resource plane. "How you're consuming this is a little bit up to the individual preference of the user, and what the platform team has configured for you. So we're seeing some teams like to use a user interface and some teams like to use code based interactions," Von Grünberg explained. In some ways, a IDP is reminiscent of the platform-as-a-service packages of a decade ago. They also were designed to help developer efficiency, though devs chafed at the limited number of tools they were allowed to use in these walled gardens. That was a mistake, Von Grünberg said. Those platforms required developers to use a small set of pre-defined times. "We don't want to get back to those times, which is why we want to provide sensible defaults," Von Grünberg said. A good IDP will provide developers with "golden paths" or "paved roads" as Netflix calls them. "Developers can stay on those paths if they want," Von Grünberg said. They can enjoy the security default and service-level agreements (SLAs) from the engineers. But developers are also free to l
Watch on YouTube ↗ (saves to browser)
Sign in to unlock AI tutor explanation · ⚡30

Playlist

Uploads from The New Stack · The New Stack · 0 of 60

← Previous Next →
1 What's Next for the Cloud Foundry Foundation in 2017 with Executive Director Abby Kearns
What's Next for the Cloud Foundry Foundation in 2017 with Executive Director Abby Kearns
The New Stack
2 How Unikernels Can Better Defend against DDoS Attacks
How Unikernels Can Better Defend against DDoS Attacks
The New Stack
3 Weaveworks is Bringing Horizontal Scaling to Prometheus
Weaveworks is Bringing Horizontal Scaling to Prometheus
The New Stack
4 TNS Analysts Thanksgiving Special: The Evolution of Kubernetes and the Container Ecosystem
TNS Analysts Thanksgiving Special: The Evolution of Kubernetes and the Container Ecosystem
The New Stack
5 How Rancher Labs is Seeing Kubernetes Put to Work in Production
How Rancher Labs is Seeing Kubernetes Put to Work in Production
The New Stack
6 SAP Tests Kubernetes for Cloud-Native Enterprise Software Deployments
SAP Tests Kubernetes for Cloud-Native Enterprise Software Deployments
The New Stack
7 Event Marketing for Today's Developer Evangelists and Community Managers
Event Marketing for Today's Developer Evangelists and Community Managers
The New Stack
8 NodeSource Introduces Certified Modules to Improve Node.js Security
NodeSource Introduces Certified Modules to Improve Node.js Security
The New Stack
9 How Lightstep is Illuminating the Case for Distributed Tracing
How Lightstep is Illuminating the Case for Distributed Tracing
The New Stack
10 How OpenStack Aims to be More Inclusive without being Exclusive
How OpenStack Aims to be More Inclusive without being Exclusive
The New Stack
11 How Shuttlecloud Saves Time and Money by Monitoring with Prometheus
How Shuttlecloud Saves Time and Money by Monitoring with Prometheus
The New Stack
12 Creating Analytics-Driven Solutions for Operational Visibility
Creating Analytics-Driven Solutions for Operational Visibility
The New Stack
13 Understanding the Application Pattern for Effective Monitoring
Understanding the Application Pattern for Effective Monitoring
The New Stack
14 Building On Docker's Native Monitoring Functionality
Building On Docker's Native Monitoring Functionality
The New Stack
15 The Importance of Having Visibility Into Containers
The Importance of Having Visibility Into Containers
The New Stack
16 How Getting Your Project in the CNCF Just Got Easier
How Getting Your Project in the CNCF Just Got Easier
The New Stack
17 Tectonic Summit Pancake Breakfast: How to Sell Kubernetes to the Hypervisor-Minded
Tectonic Summit Pancake Breakfast: How to Sell Kubernetes to the Hypervisor-Minded
The New Stack
18 The Buzz at Tectonic Summit 2016 in New York City
The Buzz at Tectonic Summit 2016 in New York City
The New Stack
19 Bringing Clarity to the Future of Node.js Modules
Bringing Clarity to the Future of Node.js Modules
The New Stack
20 How FluentD Can Help Monitor Microservice Architectures Through Unified Logging
How FluentD Can Help Monitor Microservice Architectures Through Unified Logging
The New Stack
21 Reshaping Front End Development with Warehouse.ai
Reshaping Front End Development with Warehouse.ai
The New Stack
22 2016 Year End Wrap-Up: Discussing Docker, OpenStack, and Open Source
2016 Year End Wrap-Up: Discussing Docker, OpenStack, and Open Source
The New Stack
23 Here's Why You Should Build a Robot Using Node.JS: Because You Can
Here's Why You Should Build a Robot Using Node.JS: Because You Can
The New Stack
24 How the Node.js Foundation is Utilizing Participatory Governance Models
How the Node.js Foundation is Utilizing Participatory Governance Models
The New Stack
25 Set Up an MongoDB Replica Set in Less Than an Hour Using Bitnami Packages
Set Up an MongoDB Replica Set in Less Than an Hour Using Bitnami Packages
The New Stack
26 Determining Who Bears the Burden of Ensuring NPM Module Security
Determining Who Bears the Burden of Ensuring NPM Module Security
The New Stack
27 How Intel Snap uses Telemetry and Kubernetes to Drive Enterprise Efficiency
How Intel Snap uses Telemetry and Kubernetes to Drive Enterprise Efficiency
The New Stack
28 How the NFL Scored a Touchdown with its Open Source React Framework Wildcat
How the NFL Scored a Touchdown with its Open Source React Framework Wildcat
The New Stack
29 Aporeto CEO Dimitri Stiliadis: When it Comes to Security, Context is King
Aporeto CEO Dimitri Stiliadis: When it Comes to Security, Context is King
The New Stack
30 The Buzz at Node.JS Interactive
The Buzz at Node.JS Interactive
The New Stack
31 Why Going Serverless Doesn't Mean 'No Ops'
Why Going Serverless Doesn't Mean 'No Ops'
The New Stack
32 How Node.js is Transforming Today's Enterprises
How Node.js is Transforming Today's Enterprises
The New Stack
33 JJ Asghar Interview
JJ Asghar Interview
The New Stack
34 How Capital One is Using APIs to Streamline Auto Financing
How Capital One is Using APIs to Streamline Auto Financing
The New Stack
35 SXSW 2017: How Machine Learning Differs From Regular Programming
SXSW 2017: How Machine Learning Differs From Regular Programming
The New Stack
36 SXSW 2017: Data-Driven Applications with Capital One DevExchange's Hydrograph
SXSW 2017: Data-Driven Applications with Capital One DevExchange's Hydrograph
The New Stack
37 SXSW 2017: How Good Engineers Make Bad Business Decisions
SXSW 2017: How Good Engineers Make Bad Business Decisions
The New Stack
38 CloudNativeCon & KubeCon EU Pancake Breakfast 2017: Kubernetes and the Multi-Cloud
CloudNativeCon & KubeCon EU Pancake Breakfast 2017: Kubernetes and the Multi-Cloud
The New Stack
39 CNCF Executive Director Dan Kohn: What's Next for CNCF in 2017
CNCF Executive Director Dan Kohn: What's Next for CNCF in 2017
The New Stack
40 Exploring the Latest Container Runtime Projects in the CNCF
Exploring the Latest Container Runtime Projects in the CNCF
The New Stack
41 Exploring the Future of the Kubernetes Ecosystem
Exploring the Future of the Kubernetes Ecosystem
The New Stack
42 Kubernetes and Continuous Deployment
Kubernetes and Continuous Deployment
The New Stack
43 Kris Nova of Deis at CouldNativecon/Kubecon in Berlin
Kris Nova of Deis at CouldNativecon/Kubecon in Berlin
The New Stack
44 Docker's Quest for Simplicity with the Evolution of Containerd
Docker's Quest for Simplicity with the Evolution of Containerd
The New Stack
45 Developers First: The Cloud Foundry Service Broker API and Kubernetes
Developers First: The Cloud Foundry Service Broker API and Kubernetes
The New Stack
46 Mapping the Future of CoreOS's rkt in the CNCF
Mapping the Future of CoreOS's rkt in the CNCF
The New Stack
47 Red Hat and Dell EMC: Two Perspectives from DockerCon
Red Hat and Dell EMC: Two Perspectives from DockerCon
The New Stack
48 Capital One Opened its APIs to Third-Party Developers — Here’s What They Learned
Capital One Opened its APIs to Third-Party Developers — Here’s What They Learned
The New Stack
49 SUSE Joins the CNCF, Brings Kubernetes to OpenStack Cloud 7
SUSE Joins the CNCF, Brings Kubernetes to OpenStack Cloud 7
The New Stack
50 How Capital One Brings Open Source To The  Banking Industry
How Capital One Brings Open Source To The Banking Industry
The New Stack
51 OSCON Is Coming Back To Portland, A Show Wrapup With Co-Chair Kelsey Hightower
OSCON Is Coming Back To Portland, A Show Wrapup With Co-Chair Kelsey Hightower
The New Stack
52 Dev Or Ops Doesn’t Matter, You Need Observability
Dev Or Ops Doesn’t Matter, You Need Observability
The New Stack
53 Taking The Next Steps In Developing An Open Source Culture
Taking The Next Steps In Developing An Open Source Culture
The New Stack
54 SXSW 2017: How Capital One Became Technology-First With Open Source
SXSW 2017: How Capital One Became Technology-First With Open Source
The New Stack
55 Apcera   Old Apps Spanning New Clouds
Apcera Old Apps Spanning New Clouds
The New Stack
56 Provenance: The Peace of Mind Chef Habitat Seeks to Deliver
Provenance: The Peace of Mind Chef Habitat Seeks to Deliver
The New Stack
57 InSpec: Human Readable, Automated Compliance
InSpec: Human Readable, Automated Compliance
The New Stack
58 The Evolution of SAP HANA Express
The Evolution of SAP HANA Express
The New Stack
59 Women Engineers Who Inspire And Never Give Up
Women Engineers Who Inspire And Never Give Up
The New Stack
60 Three Perspectives on the Evolution of Container Security
Three Perspectives on the Evolution of Container Security
The New Stack

The video teaches the importance of platform engineering and Internal Developer Platforms (IDPs) in simplifying development and providing delightful experiences for developers. It highlights the concept of the Golden Path and the role of tools such as Backstage, Cortex, and Humanitec's orchestrator product in platform engineering.

Key Takeaways
  1. Build an Internal Developer Platform (IDP)
  2. Implement the Golden Path
  3. Design Abstraction Layers
  4. Create a Developer Control Plane
  5. Implement Integration and Delivery Planes
  6. Use tools such as Backstage, Cortex, and Humanitec's orchestrator product
💡 The Golden Path simplifies development and provides delightful experiences for developers by covering 80% of use cases with abstraction and defaults.

Related Reads

Up next
Welcome to the Next Temperamental Era
Charles Schwab
Watch →