What is Internet Protocol (IP)? (System Design for Beginners – Episode 2)

Modern systems designed for scale—such as social media platforms, video streaming services, B2B SaaS, and e-commerce applications—depend entirely on the internet to function. To understand how these large-scale systems operate, we must first examine the foundational mechanics of the internet, network transmission, and the Internet Protocol (IP).

1. Why Do We Need the Internet Protocol?

At a high level, modern distributed systems involve communicating machines that need to exchange information reliably over a shared network. However, two machines cannot simply transmit massive blobs of data directly to one another across arbitrary distances.

Instead, data transit requires an organized infrastructure of intermediate devices and a standardized set of rules governing how communication occurs.

• The Scale Problem: Millions of distinct machines ranging from client browsers to massive cloud data centers need to talk to each other simultaneously.
• The Need for Standardization: Without a universal language and rulebook, hardware from different vendors could not interpret data sent across shared communication links.
• The Role of Rules: The internet requires a rigorous framework to define how connections are established, how messages are structured, and how data is successfully transferred.

2. Core Mental Model: The Postal Mail Analogy

To visualize how data moves across the internet, consider the traditional postal mail service.

Source Machine (Sender)
   |
   | splits message into letters (Packets)
   v
Local Router
   |
   | inspects destination address on envelope
   v
Intermediate Routers (Transit Network)
   |
   | forwards packet independently hop-by-hop
   v
Destination Machine (Receiver)
   |
   | rearranges packets in correct sequence
   v
Application Display (User Screen)

• The Message as Letters: Just as a long letter is split across multiple pages and envelopes if it exceeds weight limits, network data is broken down into small, manageable chunks called packets.
• Envelopes as Headers: Every postal letter contains a "To" and "From" address on the outside. Similarly, every network packet embeds source and destination IP addresses inside its header.
• Postal Sorters as Routers: Postal sorting facilities look only at the destination address on each individual letter and forward it closer to its destination, without needing to know the entire history of the postal truck's journey. Routers behave identically with network packets.

3. How Data Moves: Packets, Routers, and Addressing

When a machine sends data across the internet, the raw payload is never transmitted as a single continuous stream.

Packetization and Routing Mechanics

1. Segmentation: The source machine breaks down large files or messages into smaller, uniform chunks known as packets.
2. Metadata Attachment: Each packet is stamped with a header containing critical routing metadata, most notably the source IP address and destination IP address.
3. Hop-by-Hop Transit: Because machines cannot connect directly across wide-area networks, packets are dispatched to local routers.
4. Network Interconnection: Routers communicate with neighboring routers, passing packets forward until they reach the target destination network.

Server-Client Interaction Model

In typical web architectures (e.g., streaming a movie on Netflix or browsing a social media feed):

• The user's device acts as the client making an outbound request.
• The remote cloud server receives the request and streams back the requested data split into thousands of individual packets.
• The client device's network stack receives these packets, rearranges them into their original sequence, and renders the content smoothly on the screen.

4. The 4-Layer Network Architecture Stack

Networking functionality is organized into distinct conceptual layers. Understanding where the Internet Protocol sits helps clarify the separation of concerns in system architecture.

Layer Breakdown Table

5. Key Characteristics of the Internet Protocol (IP)

The Internet Protocol is engineered with specific design choices that prioritize speed, scale, and simplicity over strict delivery guarantees.

• Connectionless Nature: Each packet is entirely independent. The network does not set up a dedicated, permanent circuit or session in advance between sender and receiver.
• Dynamic Routing: Packets do not need to follow the exact same path or hit the exact same sequence of routers. Different packets belonging to the same file can take entirely different physical routes across the globe.
• Out-of-Order Arrival: Because packets take independent paths with varying network congestion, they may arrive at the destination out of sequence.
• Best-Effort Delivery & Packet Loss: IP does not guarantee delivery, sequential ordering, or duplicate suppression. Packets can occasionally be dropped by congested routers.
• Reliability Hand-off: Reliability, retransmission of lost packets, and reordering are intentionally offloaded to higher layers in the stack—most notably the Transport Layer (TCP).

6. Stateless Forwarding and Hierarchical Addressing

Two architectural pillars allow the internet to scale to billions of devices without collapsing under state management overhead:

Router A
   |
   | reads packet header (Stateless Forwarding)
   | consults local Routing Table
   v
Router B (Next Hop)

1. Stateless Forwarding:

   • Routers do not maintain memory or pre-flow state regarding previous packets.
   • When a packet arrives, the router examines its header, consults its internal routing table, forwards the packet to the next hop, and immediately forgets about it. This keeps routers lightweight and lightning-fast.

2. Hierarchical Addressing:

   • IP addresses are structured hierarchically (much like country codes, postal codes, and street numbers).
   • This hierarchy allows routers to aggregate routes. Instead of storing an individual rule for every single device on earth, a router only needs to know which general direction (gateway) leads to broad network blocks, making global routing scalable.

7. Summary and Key Takeaways

• Internet Dependency: Modern large-scale software systems (social media, streaming, SaaS) require a robust understanding of internet mechanics to architect effectively.
• Packet-Based Transmission: Data is split into smaller, manageable chunks called packets containing source and destination addresses.
• Router Interconnection: Machines do not connect directly; packets hop across intermediate routers to traverse the globe.
• The Postal Analogy: IP addresses act like mailing addresses on envelopes, allowing routers to forward packets independently without knowing the entire journey.
• Four-Layer Stack: Networking functionality is divided into Application, Transport, Internet, and Network layers.
• IP is Connectionless: Packets travel independently, can take different routes, can arrive out of order, and may occasionally be dropped.
• Reliability Belongs to TCP: The Internet Protocol provides "best-effort" delivery; absolute reliability and ordering are handled by upper layers like TCP.
• Stateless & Hierarchical: Routers use stateless forwarding by inspecting packet headers against routing tables, and IP addresses use a hierarchical structure to make global routing scalable.