Skip to main content Farhan Javed

Posts on Devops

  1. Optimizing Costs and Performance with Bare Metal and Cloud.

    For over a decade, the “cloud-first” approach has dominated IT strategy, promising unparalleled scalability and ease of use, and proving to be a revolutionary force for innovation, particularly for startups. However, this strategy is now “breaking budgets”, with organizations often spending “2-5x more than necessary on infrastructure”. Research indicates that a significant “20-30% of all cloud spending is wasted” due to factors like over-provisioning, idle resources, suboptimal service selection, and punitive data egress fees.
  2. Why Dukaan Ditched the Cloud for Bare Metal

    The Problem: Dukaan, a fast-growing e-commerce platform, was a model cloud user. They leveraged the scalability of their cloud provider to grow rapidly. However, this growth came with an astronomical price tag. Their cloud bills were soaring to unsustainable levels with their peak at $90,000 per month, eating directly into their margins. Performance, particularly for their databases, was also a concern due to the inherent limitations of virtualized disk I/O.
  3. My First GSoC Attempt

    Even though my GSoC proposal didn’t get selected, the process itself was an amazing learning experience. I worked on Sugar, a lightweight tool that simplifies container management using a declarative .sugar.yaml config. Sugar basically abstracts away complex Docker CLI or Compose commands and wraps them neatly into a user-friendly interface. It’s CLI-first but powerful—and that’s what made the idea of building a TUI (Terminal User Interface) on top of it so exciting to me.
  4. How Hotstar Application Scaled 25 Million Concurrent Users

    Hotstar and others often don’t solely rely on traditional auto-scaling methods for critical events because traditional auto-scaling can be too slow. For instance, it can take 200-300 seconds for 200 servers to spin up, which is too long for “tsunami traffic” that requires rapid scaling. Instead, Hotstar uses proactive automated scaling algorithms based on historical data and anticipated demand to provision machines before the rush. They also employ strategies like “preparing for three times of peak” and obsessive load testing with “homegrown” load generators to ensure readiness.