Conference | DevOps Days - Zurich

Webinar Transcript:

Introduction: A Journey Through Platform Engineering

So, thanks. I'd like to thank everyone for the invitation. I'm going to talk about something novel because there aren't many applications like it. I hope it works out. Also, I'm planning to do a live demo, which is always a risky idea because it often fails. We'll see.

I don't expect anyone to know me, so I'll do a quick introduction. Since I'm the only speaker, there's no one to introduce me, so I can introduce myself. I'm from Switzerland. But not everyone here is from Switzerland. I'm not from the bad part of Switzerland, which is French-speaking. On top of that, my parents are from the Italian part of Switzerland, which is even worse. Just kidding, I love it.

From Research to Cybersecurity Startups with Platform Engineering

I studied here, but I've lived most of my professional life outside Switzerland. I worked for IBM in their research division in New York and Tokyo. That was after I finished my PhD at EPFL in Switzerland. Working in research makes you see everything as a problem to solve, and you always want to put your personal touch on it. I think research is great, but it gets old. So, eventually, I moved on and created my first startup, Strong Codes, in 2013. I love cybersecurity, which was what Strong Codes was about.

Creating and Growing Startups

I don't expect anyone to know what code obfuscation is, but that's what we did. We were eventually acquired by Snapchat. I have two kids, so I know Snapchat. I left Snapchat because I like creating products, so I started Strong Network, the company I'm part of now. We have a platform that I'll demo briefly to show you what it does.

I also like writing. Writing helps make ideas more concrete. I've written a lot of things, including articles on our website. Someone told me yesterday that I don't say everything in the articles, and that's true because if I did, everyone would copy us. There's always a balance between what to share and what to hold back.

Startup life moves quickly. I like writing about information security, DevOps acceleration, open innovation, design thinking, and job theory. I was happy when the keynote speaker yesterday mentioned Christensen because it meant I was on the right track.

Applying Platform Engineering to the Coding Stage

Applying platform engineering to the coding stage is rare. You can use containers and container registries to prepare developer environments. They can download the container and work locally. But coding involves many things beyond containers. We have files, keys, DevOps tools, IDEs, and plugins, and containers are just a small part of the picture.

Why not apply platform engineering to create a full onboarding experience for developers? This is important because onboarding developers in many industries takes a long time. At Snapchat, it took two to three weeks to onboard a developer. It's not just technical but also administrative. You get your laptop, install things, download repos, and clone projects. The Snap app was many gigabytes, and it took forever over VPN.

My goal is to use platform engineering to accelerate onboarding for an entire team while ensuring compliance with organizational policies. The onboarding should handle risk controls, security, laptop configuration, access control, repositories, data sources, and more. The idea is that when you get access to your machine, you should authenticate once, and that's it.

Creating Projects with Platform Engineering

I depicted it this way: the platform engineer creates a project, and users join the project with specific roles. We want to have Predictive Analytics to forecast costs after six months and Prescriptive Analytics to decide the kind of machine needed. Each machine should be fully installed, secured, and have access to diverse resources, like GitHub, SSH machines, containers, and more. I wrote an article on this and will publish it. If you want a preprint, just ask.

Episode One: The Online Coding Environment

Development can now happen in the cloud. I was skeptical at first, thinking it was just SSH-ing to a machine, but it's friendlier and transparent. You can use Docker locally or use cloud-based containers from platforms like AWS Cloud9, Azure CodeSpaces, Google Workstations, and others. These are called Cloud Development Environments (CDEs), which became a Gartner category last year. I wrote about this extensively, tracing the history of moving data online, from CRM data to development data.

Episode Two: Security in Platform Engineering

Security. Organizations need security policies for laptops and code, including protecting against vulnerabilities and securing infrastructure. Secure CDEs involve an IDE and a secure browser. The IDE accesses the container, while the browser provides secure application access. Access is proxied, keeping credentials off developer environments. Credential theft is a major issue, as seen in attacks on Slack, CircleCI, and Okta. Proxying solves this.

Episode Three: Securing the Cloud with Platform Engineering

Securing the Cloud. With infrastructure control, we can monitor all interactions, including clipboard and network traffic. It's like what Citrix does, but this is local code execution in a browser. The idea initially was to replace Citrix, but Citrix became our biggest client. The surprise was that infrastructure security becomes code security when you control the infrastructure and understand the business process. You can stop credentials from being copied into the clipboard or flagged open-source licenses when code is copied from Stack Overflow.

Episode Four: DevOps and Platform Engineering

Secure CDEs bring new possibilities, like Predictive Analytics and improving DevOps. I wrote about this in The New Stack. I'll speak about it in Bern, focusing on DevOps novelty.

Episode Five: Abstraction with Platform Engineering

Abstraction. Now, we can create virtual secure laptops, allowing for BYOD with client attestation. Secure laptops enable developer experience, productivity, and access control while minimizing costs.

We provide an abstraction of the secure laptop, with parameters for developer experience, dev productivity, access control, and security. We create this abstraction with a Jira task that specifies the machine template, security policy, and invited team members. The task triggers the API, creating the project, and everyone is onboarded.

Platform Engineering for Predictive and Prescriptive Analytics

The goal is to have predictive and prescriptive analytics for cloud costs and machine needs. The article covers all this in detail, along with mitigations for tool sprawl, productivity, and compliance. I also discuss cloud-native application protection strategies. If you like t-shirts, I have some with skulls or other designs. Feel free to connect with me. We have time for questions. Thank you!

---

All material in this text can be shared and cited with appropriate credits. For more information about our platform, please contact us at hello@strong.network