Cloud & Edge Computing
Where code actually runs — and why that location increasingly matters as much as the code itself.
1.6.1Definition
Cloud computing means running applications on infrastructure owned and operated by a third-party provider (AWS, Azure, Google Cloud), rented on demand rather than bought and physically maintained. Edge computing pushes computation and content physically closer to the end user — running in many geographically distributed locations rather than one or a few central data centers — to cut the network distance data has to travel.
1.6.2Why It Exists
Before the cloud, running a digital product meant buying, housing, and maintaining physical servers — a large upfront cost and a fixed capacity that couldn't easily flex with demand. Cloud computing exists to turn infrastructure into a metered, on-demand utility: pay for what's used, scale up during traffic spikes, scale down when idle. Edge computing exists to address a different constraint entirely — physical distance and the speed of light. Even a perfectly efficient cloud server in one location is still slow to reach for a user on the other side of the planet; the edge model exists specifically to shrink that distance.
1.6.3Core Cloud Service Categories
- IaaS (Infrastructure as a Service) — raw virtual servers, storage, and networking; maximum control, maximum responsibility for setup and maintenance.
- PaaS (Platform as a Service) — a managed environment for deploying code without managing the underlying servers directly.
- SaaS (Software as a Service) — fully finished applications delivered over the internet, requiring no infrastructure management at all from the user.
- Serverless — code that runs in response to events without a persistently running server to manage, billed by actual execution time.
1.6.4Common Mistakes
- Choosing a cloud region arbitrarily. Deploying infrastructure in a region far from the actual user base, adding unnecessary latency that no amount of application-level optimization can fix.
- Ignoring cost structure until the bill arrives. Cloud costs scale with usage in ways that are easy to underestimate — an inefficient query or unbounded storage growth translates directly into cost, unlike fixed-cost physical hardware.
- Treating "the cloud" as inherently more reliable without designing for the redundancy (multiple availability zones, failover) that actually delivers that reliability — a single-region, single-instance cloud deployment can fail exactly like a single physical server.
1.6.5Best Practices
- Choose deployment regions based on where actual users are, not on habit or where a team happens to be located.
- Use edge caching and CDNs (content delivery networks) for static assets regardless of where the core application runs.
- Set concrete cost alerts and budgets early — cloud spend that scales invisibly with usage is a common source of surprise.