I received a message from a friend admitting he’s unclear about what I do professionally. I’ve written this article to simplify and explain the daily responsibilities of an SRE, System Architect, and Infrastructure Engineer.
- Lead arch for CDN security for several popular websites managing infrastructure for 200M+ users per month.
Imagine a busy highway with millions of cars (users) traveling on it every month. I am in charge of making sure that the highway (for several popular websites like People, Shape, Health, Investopedia, Allrecipes etc) is safe, smooth, and can handle all those cars. - Integrated Fastly’s Signal Sciences to defend against malicious traffic & DDoS attacks – fending off 60+ million monthly requests.
I set up a security system to keep out unwanted visitors or “bad cars.” Each month, this security system kept over 60 million of these bad cars from causing trouble. - Reduced AWS costs by migrating workloads to SPOT & Graviton, saving $1.2+ million p.a.
Think of AWS/GCP as a place where the highway’s tools and equipment are stored, but it costs money. I found a more cost-effective way to store and manage these tools, saving over a million dollars every year. - Architected a serverless tool to check IP origins during attacks, clear web caches, and a single pane of glass for load balancers. (Rapid CDN Switching & a break-glass tool to be used in outages)
I created a tool (like a security checkpoint) that quickly identifies and deals with unwanted visitors or issues. It also has an emergency switch that helps redirect traffic and deal with any major problems quickly. - Architected Prometheus utilizing Thanos to make it highly available & ES clusters in EKS with fault tolerance.
I designed a system (like a traffic control center) that makes sure everything runs smoothly on the highway and can handle any issues without breaking down. Also used some specific tools and technologies to do this, but the main idea is that they made it very reliable. - Revamped Kafka with Envoy to ensure seamless connectivity during node IP changes.
Imagine cars (data or messages) are constantly traveling to specific destinations (computers or servers). The addresses of these destinations are like the exits on the highway. Now, the system directing cars to the right exits is called “Kafka”.Sometimes, the exit numbers (addresses) change. This can cause confusion, making cars miss their exits or even get lost.
To handle this, I introduced a skilled navigator named “Envoy”. Whenever an exit number changes, Envoy quickly updates the route for the cars, ensuring they always reach the correct exit smoothly, without any hiccups. In essence, I improved the highway’s navigation system to make sure cars always find their way, even when the road conditions or exit numbers change. - Designed and implemented infrastructure for enhanced monitoring and observability using Elasticsearch on Graviton, optimizing for cost-efficiency and performance.
Imagine now to keep traffic flowing smoothly and to prevent any accidents, you’d want a really advanced traffic monitoring system that watches the cars, understands where traffic jams might occur, and identifies any potential issues on the road.
To build this top-notch traffic monitoring system, I used a tool called “Elasticsearch”. Think of Elasticsearch as a set of high-powered binoculars that can see every car, its speed, and its destination. Now, to ensure that these binoculars work efficiently without spending too much money, I mounted them on a tower called “Graviton”.
This tower is special because it’s designed to offer a great view of the highway, while also being less expensive to build and maintain. So, in essence, I built a state-of-the-art traffic monitoring system on our highway, making sure it’s both cost-effective and high-performing, ensuring smooth and safe travels for all cars.