What is networking for DevOps?
Networking is how your services find and reach each other and the outside world. For DevOps it means the practical layer: IP addresses and subnets, DNS, ports and firewalls, and the load balancers and proxies that sit in front of everything. Most "it can't connect" incidents live here.
Why it matters
Distributed systems are mostly networking, and connectivity bugs are some of the hardest to diagnose without the fundamentals. Knowing how a packet gets from a load balancer to a pod, and where a firewall or security group can block it, turns a multi-hour outage into a quick fix.
What to learn
- IP addresses, subnets, and CIDR notation
- DNS resolution and record types
- Ports, and the difference between listening and reachable
- Firewalls and cloud security groups
- NAT and private vs public networks
- Load balancers and reverse proxies
- Tools:
dig,curl,ss,traceroute
Common pitfall
Assuming a service is unreachable because it is down, when a firewall or security group is silently dropping the traffic. The process is listening fine; the packets never arrive. Check connectivity layer by layer — DNS, then port reachability, then the app — instead of restarting the service blindly.
Resources
Primary (free):
- Cloudflare — Learning: networking · docs
- Julia Evans — Networking zines · article
- AWS — VPC fundamentals · docs
Practice
On a server, find what is listening on which ports with ss, then test whether a
port is reachable from another machine with curl or nc. Trace the path to a
public host. Deliberately block a port with a firewall rule and confirm the
service is up but unreachable. Done when you can isolate where traffic stops.
Outcomes
- Read CIDR notation and reason about subnets.
- Distinguish a down service from blocked connectivity.
- Explain how a load balancer routes to backends.
- Debug connectivity layer by layer with the right tools.