IP: The Protocol That IS the Internet
How this deceptively simple addressing and routing protocol became the foundation that literally every other internet technology builds upon
Here’s a mind-bending fact: every single thing you do online - every web page, every email, every video stream, every Zoom call - is carried by IP packets. TCP runs on IP. UDP runs on IP. Every protocol we’ve talked about sits on top of IP.
IP (Internet Protocol) isn’t just part of the internet - it IS the internet. The “internet” is literally a network of networks connected by IP routing. Without IP, there would be no global network, just isolated computer systems talking to their immediate neighbors.
And here’s the beautiful part: the core concept of IP is so elegantly simple that you can explain it in one sentence, yet so powerfully scalable that it connects billions of devices across the planet.
The IPv4 packet header - this 20-byte structure enables global communication between any two computers on Earth
The Revolutionary Idea: Universal Addressing
Before IP, computer networks were like isolated islands. Each network had its own addressing scheme, its own protocols, its own way of doing things. If you wanted to connect different networks, you needed custom gateways that understood both sides.
IP solved this with one brilliant insight: what if every computer on every network had a unique address in the same format?
That’s exactly what IP addresses do. Whether you’re on a university network in Tokyo, a corporate LAN in New York, or a home WiFi in rural Montana, your device gets a 32-bit IP address (in IPv4) that uniquely identifies it on the global internet.
192.168.1.100 might be your laptop. 8.8.8.8 is Google’s DNS server. 172.217.14.174 could be one of Google’s web servers. Every device gets a number, and those numbers can talk to each other regardless of what physical networks they’re connected to.
How IP Works: Best-Effort Packet Delivery
IP’s job is beautifully simple: “Take this chunk of data and try to get it from IP address A to IP address B.”
Notice I said “try” - IP makes no guarantees. Packets might:
- Get lost along the way
- Arrive out of order
- Take different routes to the destination
- Get corrupted during transmission
- Arrive multiple times (duplicates)
This seems like terrible service, but it’s actually brilliant design. By keeping IP simple and unreliable, it can:
- Work over any physical network (Ethernet, WiFi, fiber, satellite, carrier pigeon)
- Scale to billions of devices without complex state management
- Route around failures automatically
- Let higher-level protocols add reliability where needed
TCP adds reliability on top of IP. UDP adds speed on top of IP. ICMP adds error reporting on top of IP. IP is the universal foundation that everything else builds upon.
The IP Header: Every Field Has a Purpose
Look at that diagram above - every single field in the IP header serves a crucial purpose:
Source and Destination Addresses
The most important fields - where this packet came from and where it’s going. These 32-bit addresses enable global routing.
Time to Live (TTL)
Prevents packets from circulating forever if there’s a routing loop. Each router decrements the TTL; when it hits zero, the packet gets discarded.
Protocol Field
Tells the receiving computer what to do with the payload. Protocol 6 means “hand this to TCP.” Protocol 17 means “hand this to UDP.” Protocol 1 means “this is an ICMP message.”
Fragmentation Fields
Large packets can be broken into smaller pieces to fit through networks with smaller maximum transmission units (MTUs). The fragmentation fields let the destination reassemble the original packet.
Header Checksum
Error detection for the header itself. If the checksum doesn’t match, the packet gets dropped because the routing information is probably corrupted.
Routing: How Packets Find Their Way Across the Globe
The magic of IP is how packets automatically find their way from source to destination across a network of 70,000+ autonomous systems worldwide.
Here’s how it works:
Step 1: Local Delivery Check
Your computer first checks if the destination IP is on the same local network. If it is, it delivers the packet directly via Ethernet.
Step 2: Default Gateway
If the destination is remote, your computer sends the packet to its default gateway (usually your router).
Step 3: Hop-by-Hop Routing
Each router along the path:
- Looks at the destination IP address
- Consults its routing table to find the “next hop”
- Forwards the packet to the next router
- Decrements the TTL by 1
Step 4: Final Delivery
When the packet reaches the destination network, the final router delivers it to the actual destination computer.
The brilliant part is that each router only needs to know the next step - not the entire path to every destination. This distributed decision-making is what allows the internet to scale globally.
Subnetting: Organizing the Address Space
IP addresses aren’t just random numbers - they’re hierarchically organized to make routing efficient.
An IP address like 192.168.1.100/24 actually contains two pieces of information:
- Network portion: 192.168.1.0 (the first 24 bits)
- Host portion: 100 (the last 8 bits)
This means 192.168.1.0/24 can have hosts from 192.168.1.1 to 192.168.1.254 (256 addresses total, minus network and broadcast addresses).
This hierarchical addressing allows routers to:
- Aggregate routes (one entry for 192.168.1.0/24 instead of 254 individual entries)
- Make forwarding decisions based on network prefixes
- Scale routing tables to handle millions of networks
The IPv4 Address Exhaustion Problem
IPv4 uses 32-bit addresses, which provides about 4.3 billion unique addresses. In 1981, this seemed like an infinite number.
By 2011, we ran out. The global pool of IPv4 addresses was exhausted.
This led to some creative solutions:
NAT (Network Address Translation)
Most home networks now use private IP addresses (like 192.168.x.x) internally and share a single public IP address through NAT. Your router translates between internal and external addresses.
CIDR (Classless Inter-Domain Routing)
Instead of fixed address classes, networks can be any size using subnet masks. This allows more efficient allocation of address space.
IPv6: The Long-Term Solution
IPv6 uses 128-bit addresses, providing approximately 340 undecillion (3.4×10³⁸) unique addresses. That’s enough to give every atom on Earth its own IP address.
IPv6: The Future (That’s Taking Forever)
IPv6 addresses look completely different: 2001:0db8:85a3:0000:0000:8a2e:0370:7334
IPv6 brings major improvements:
- Massive address space (340 undecillion addresses)
- Built-in security with IPSec
- Better auto-configuration
- Simplified header format for faster processing
- No need for NAT - every device can have a global address
But adoption has been painfully slow. As of 2025, IPv6 traffic is still only about 35% globally. Why?
IPv4 keeps working thanks to NAT and careful address management. Organizations don’t want to upgrade working infrastructure unless they absolutely have to. “If it ain’t broke, don’t fix it” mentality.
The transition is happening, but it’s taking decades instead of years.
Why IP is Genius: Simplicity Enables Complexity
IP’s brilliance lies in what it doesn’t try to do:
No Reliability Guarantees
IP doesn’t promise packets will arrive. This simplicity means IP can work over any underlying network without complex state management.
No Flow Control
IP doesn’t try to manage network congestion. This allows higher-level protocols to implement the flow control that makes sense for their specific use case.
No Security
Original IP had no encryption or authentication. This kept it simple and allowed security to be added as a separate layer (like TLS).
Stateless Operation
Routers don’t need to remember previous packets. Each packet is independently routed based only on its destination address.
This minimalist approach enabled IP to:
- Scale to billions of devices
- Work over any physical medium
- Evolve without breaking existing implementations
- Support protocols that didn’t exist when IP was created
The Network Effects: IP Connects Everything
Once IP became the universal standard, it created powerful network effects:
More devices using IP → More valuable to connect to the IP network → Even more devices adopt IP
This virtuous cycle is why proprietary networking protocols (like NetBIOS, IPX/SPX, AppleTalk) all eventually lost to IP. Nobody wanted to be on an island when everyone else was on the global IP network.
Today, everything speaks IP:
- Your smart TV gets Netflix via IP
- Your car downloads software updates via IP
- Industrial sensors report data via IP
- Security cameras stream video via IP
- Your refrigerator (unfortunately) might tweet via IP
IP in Practice: The Foundation Layer
Every protocol we’ve discussed depends entirely on IP:
TCP creates reliable connections over IP packets
UDP provides fast messaging over IP packets
DNS resolves names to IP addresses
HTTP sends web pages via TCP over IP
Email (SMTP) travels via TCP over IP
BGP tells routers how to forward IP packets
Without IP, none of these would work. There would be no global connectivity, no universal addressing, no way for different networks to interconnect.
IP is like the electrical grid - you don’t think about it until it’s not there, but everything else depends on it.
The Human Impact: Connecting the World
The social and economic impact of IP is staggering:
Global Communication: Anyone can communicate with anyone else, regardless of location or network provider
Economic Opportunity: A developer in Bangladesh can serve customers in Silicon Valley
Information Access: Knowledge and education are available to anyone with an internet connection
Social Networks: Billions of people can connect and share experiences instantly
Remote Work: Physical location becomes irrelevant for many jobs
All of this is possible because IP provides universal connectivity. Any device with an IP address can potentially communicate with any other device with an IP address.
The Engineering Marvel: 50+ Years and Still Growing
IP was first defined in RFC 760 in 1980, building on earlier ARPANET work from the 1970s. The basic packet format has remained remarkably stable for over 40 years.
The fact that a protocol designed when computers were room-sized mainframes still works perfectly for smartphones, IoT devices, and cloud computing is a testament to truly great engineering.
IP has scaled from connecting a few dozen research computers to connecting billions of devices without fundamental changes to the core protocol.
A Personal Note
I’ve been working with IP since the early 1990s, and I still find it elegant and powerful. The beauty of IP is that it solves exactly one problem really well: getting packets from point A to point B across interconnected networks.
By not trying to solve every problem, IP created a stable foundation that other protocols could build upon. This separation of concerns is what allowed the internet to evolve so rapidly.
Every time I type ping 8.8.8.8 and see packets flowing to Google’s DNS server and back, I’m reminded of the remarkable achievement that IP represents. Those packets might travel through a dozen different networks, across multiple continents, managed by different organizations, using different technologies - but they just work.
That’s the mark of a protocol that truly changed the world.
Looking Forward: IP’s Next 50 Years
IP continues to evolve:
- IPv6 adoption is accelerating as IPv4 addresses become scarce
- Mobile IP enables seamless connectivity as devices move between networks
- Software-Defined Networking gives administrators more control over IP routing
- IoT is connecting billions of new devices to the IP network
The fundamental principles of IP - universal addressing, best-effort delivery, and stateless routing - remain as relevant today as they were 50 years ago.
IP didn’t just enable the internet. IP IS the internet.
Want to see what runs on top of IP? Check out our deep dives into TCP (reliable delivery), UDP (fast delivery), and DNS (address resolution). Or explore our complete protocol library.
