Networking Fundamentals
Networking essentials from Hello Interview
🧩 What is Networking?
At its core, networking is about connecting devices (computers, phones, servers, etc.) so they can share data and resources (like files, printers, or internet access).
A network can be as small as:
Two computers linked with a cable,
or as large as:
The Internet — billions of interconnected devices.
🍰 The OSI model
The OSI (Open Systems Interconnection) model is a conceptual framework that standardizes how different networking systems communicate. Think of it as a stack of responsibilities—each layer handles specific tasks and talks only to the layers directly above and below it. It provides one way of thinking about networking. It splits the network communication between 2 devices on a network into 7 abstraction layers.
Conceptually, you can think of Network, Data Link, and Physical layers as Media layers, and think about the Transport, Session, Presentation, and Application layers as Host layers.
Media layers is dealing with how data moves between points A and B. These could be in the same local network or on the opposite sides of the planet.
Host layers is dealing with how data is chopped up and reassembled for transport, and how it's formatted so that it's understandable by both sides of a network connection.
At a high-level, you might have a web browser and a web server communicating with each other at the level 7 of the OSI model and at the level 1 you might have some kind of physical network cards or interfaces. And conceptually, the data flows from the web browser request all the way down to the physical layer and then flows back to the web server (layer 7) across the layers of OSI.
This video additionally explains what happens when you make an HTTP GET request and how the data flows through OSI layers.
Key Concepts: Encapsulation and Decapsulation
As data moves down the stack (sending), each layer adds its own header:
When receiving, each layer strips its header and passes data upward.
What are headers?
Headers are control information added at each layer of the OSI model to help with data processing, routing, and delivery. For example, the Network Layer adds a header with source and destination IP addresses, while the Data Link Layer adds a header with source and destination MAC addresses for hop-to-hop delivery. Each header contains specific instructions and addresses that the corresponding layer uses to handle its part of the communication process.
Decapsulation: When data is received, it moves up the stack. Each layer strips off its corresponding header (and trailer) to process the information and pass the remaining data to the layer above.
OSI with some data engineering context:
Mapping networking concepts to OSI layers
7
Application
HTTP, DNS, FTP, S3 API
6
Presentation
Encryption, encoding, TLS
5
Session
Sessions, sockets, keep-alive
4
Transport
TCP, UDP, ports
3
Network
IP, subnets, routers, NAT, routing protocols
2
Data Link
MAC addresses, ARP, Ethernet frames
1
Physical
Cables, fiber, radio, voltage
Why is networking important for Data Engineers?
Data engineering = moving data between systems
Every data pipeline involves moving data:
From apps → message brokers (Kafka, Kinesis)
From databases → storage (S3, GCS, ADLS)
From on-prem systems → cloud
Between ETL tools → warehouses (Snowflake, Redshift, BigQuery)
All of this is network traffic.
Data engineers must design secure architectures
Modern data platforms must protect sensitive data (PII, finance, logs, events).
You need networking knowledge to secure:
Private subnets for databases and brokers
NACLs / Security Groups
VPC Peering / PrivateLink
VPC Endpoints (to access S3 or DynamoDB privately)
Transit Gateways
Examples:
Kafka brokers must often be in private subnets with controlled inbound rules.
Snowflake or Redshift often require VPC endpoints for secure internal traffic.
Understanding costs and performance
Data movement is expensive and sometimes slow depending on:
Cross-AZ traffic
Cross-region traffic
Data transfer over the public internet
NAT gateway egress costs
Without networking awareness, pipelines become expensive and slow.
Last updated