What Is the OSI Model?

The OSI model is a seven-layer conceptual framework that standardizes how network devices communicate by dividing the communication process into distinct functional layers. The International Organization for Standardization (ISO) published it in 1984 as standard ISO 7498 so products from different vendors could interoperate. This guide defines the model, walks through all seven layers and what each one does, maps real devices and protocols to their layers, contrasts the OSI model with the TCP/IP model that runs the actual internet, and shows why the layers still matter for troubleshooting.

What Is the OSI Model?

The OSI model is a seven-layer reference framework that describes how data passes between applications across a network, defined by the International Organization for Standardization. OSI stands for Open Systems Interconnection. It separates network communication into seven layers, and each layer talks only to the layers directly above and below it.

• Conceptual, not code: no single product implements all seven layers as separate pieces; the model names functions rather than running on a wire.
• One function per layer: a job such as routing belongs to one specific layer, so naming the layer names the function.
• A shared vocabulary: it gives engineers, vendors, and documentation one agreed way to describe what a network does.

The OSI model sits alongside the practical TCP/IP protocol suite that actually carries traffic on the modern internet. Best for: learning network concepts and pinpointing where a fault or function belongs.

What Are the Seven Layers of the OSI Model?

The seven layers run from Layer 1 at the bottom to Layer 7 at the top, each adding one defined service on top of the layer below. Read top to bottom, here is the one-line job of each:

• Layer 7, Application. Delivers network services straight to user programs through protocols such as HTTP, SMTP, and DNS.
• Layer 6, Presentation. Translates, compresses, and encrypts data (including TLS) so the application receives it in a usable form.
• Layer 5, Session. Opens, maintains, and closes the dialogue between two applications, and can restore it after an interruption.
• Layer 4, Transport. Segments data and controls reliability, using TCP for ordered, acknowledged delivery or UDP for fast, connectionless delivery.
• Layer 3, Network. Routes packets between separate networks using logical IP addresses; this is where routers and IP live.
• Layer 2, Data Link. Frames data and moves it between devices on the same local network using MAC addresses; this is where switches live.
• Layer 1, Physical. Transmits raw bits as electrical, optical, or radio signals over cables, fibre, or the air.

A common way to remember the order from the bottom up is the phrase “Please Do Not Throw Sausage Pizza Away.” The layers also split into two groups: Layers 1 through 4 move data across the network, and Layers 5 through 7 handle application and user-facing functions. The full table lists each layer with its function and example protocols.

How Does the Physical Layer Work?

The Physical layer, Layer 1, transmits raw bits as electrical, optical, or radio signals over the physical medium. It has no concept of addresses or meaning, because it moves only ones and zeros.

• Defines the hardware rules: cables, connectors, pin layouts, voltage levels, and data rates.
• Standards: Ethernet cabling under IEEE 802.3 and Wi-Fi radio under IEEE 802.11 set the Physical layer rules.
• Devices: a hub repeats signals here without inspecting them, so it broadcasts to every port.
• Common faults: a disconnected cable, a damaged connector, or electrical interference on the line.

How Does the Data Link Layer Work?

The Data Link layer, Layer 2, packages bits into frames and delivers them between devices on the same local network using MAC addresses. It adds a header carrying the source and destination hardware address.

• Hardware addressing: each network interface card carries a unique MAC address that identifies the sending and receiving device.
• Devices: a switch forwards frames at Layer 2 by reading the destination MAC address.
• Two sublayers: the IEEE splits this layer into Logical Link Control (LLC) and Media Access Control (MAC).
• Error detection: a frame check sequence catches corrupted frames at this layer.

How Does the Network Layer Work?

The Network layer, Layer 3, routes packets across separate networks using logical IP addresses. It determines the path a packet takes from the source network to the destination network.

• Devices: a router operates here and forwards packets based on the destination IP address.
• Dominant protocol: the Internet Protocol (IP), defined by the Internet Engineering Task Force (IETF), is the main Network layer protocol.
• Fragmentation: this layer splits a packet to fit the maximum size a given link allows.
• Diagnostics: the Internet Control Message Protocol (ICMP) carries error and diagnostic messages here, and the ping command relies on it.

How Do the Transport, Session, Presentation, and Application Layers Work?

The upper four layers handle reliable delivery and user-facing services, each with one defined role.

• Transport layer (Layer 4) controls reliability and flow. It segments data and uses TCP for ordered, acknowledged delivery or UDP for fast, connectionless delivery.
• Session layer (Layer 5) manages conversations. It opens, maintains, and closes the dialogue between two applications and can restore a session after an interruption.
• Presentation layer (Layer 6) translates data formats. It encrypts, compresses, and converts data, including TLS encryption, so the Application layer receives it in a usable form.
• Application layer (Layer 7) delivers network services. It provides protocols such as HTTP, SMTP, and DNS that user programs call directly.

The Transport layer is where the TCP and UDP split matters most. TCP guarantees delivery through acknowledgments and retransmission, while UDP omits those checks to cut delay for traffic such as video and voice. These rely on network protocols, the agreed rules each layer follows.

Which Devices and Protocols Map to Each Layer?

The layer a device works at is set by the deepest part of the packet it reads, which is why a hub, switch, and router sit on three different layers.

How Does Data Move Through the OSI Layers?

Data moves down the seven layers on the sending device and up the seven layers on the receiving device through a process called encapsulation. Each layer adds its own header to the data from the layer above before passing it down.

1. The Application layer generates the data from a user program such as a web browser.
2. The Presentation and Session layers format the data and establish the session.
3. The Transport layer splits the data into segments and adds port numbers for TCP or UDP.
4. The Network layer wraps each segment in a packet and adds source and destination IP addresses.
5. The Data Link layer wraps each packet in a frame and adds source and destination MAC addresses.
6. The Physical layer converts the frame into bits and transmits the signal over the medium.

The receiving device reverses the process. Each layer reads and removes its matching header in a step called decapsulation, then passes the remaining data up to the next layer until the Application layer delivers it to the program. The data unit even has a different name at each stage: bits at Layer 1, a frame at Layer 2, a packet at Layer 3, and a segment at Layer 4.

How Does the OSI Model Compare to the TCP/IP Model?

The OSI model uses seven layers as a teaching reference, while the TCP/IP model uses four layers that describe how the internet actually operates. The TCP/IP model combines several OSI layers into single layers.

The full structure of the practical stack appears in the overview of the TCP/IP model. Best for: use OSI to diagnose and describe, and TCP/IP to understand what the wire is really doing.

Why Does the OSI Model Matter for Troubleshooting?

The OSI model matters for troubleshooting because it lets an engineer isolate a fault to one specific layer and rule out the others. A structured approach moves from Layer 1 upward, or from Layer 7 downward, treating the layers as a map of where data can break.

• Check Layer 1 first for physical faults. A dead link, unplugged cable, or disabled port puts the problem at the Physical layer.
• Check Layer 2 for local delivery faults. A wrong MAC table entry or a VLAN misconfiguration points to the Data Link layer.
• Check Layer 3 for routing faults. A failed ping or wrong gateway points to an IP addressing or routing fault at the Network layer.
• Check Layers 4 to 7 for service faults. A blocked port, an expired certificate, or a DNS failure points to the Transport or Application layers.

This layer-by-layer method is why the OSI model is still taught in 2026 even though no device runs it directly. It turns a vague “the network is down” into a sequence of specific, checkable questions.

Last Thoughts on the OSI Model

The OSI model divides network communication into seven layers that move from raw bits at the Physical layer to user services at the Application layer. The International Organization for Standardization defined the framework in 1984 so devices from different vendors could interoperate and so engineers would share one vocabulary for network functions. Each layer adds a header during encapsulation, and the receiving device removes each header in turn through decapsulation.

The model differs from the four-layer TCP/IP model that runs the public internet, yet its layered structure still guides fault isolation during troubleshooting, with each device, from a Layer 1 hub to a Layer 3 router, sitting at a defined place. The OSI model is one part of how data travels end to end, and the wider set of networking topics sits on the hub explaining what a computer network is.

Key Takeaways:

• The OSI model defines seven layers: Physical, Data Link, Network, Transport, Session, Presentation, and Application, each handling one function.
• The International Organization for Standardization published it in 1984 as ISO 7498; the current revision is ISO/IEC 7498-1:1994.
• It is a conceptual reference, not running code; the internet itself runs on the four-layer TCP/IP model.
• Devices map to layers: a hub is Layer 1, a switch is Layer 2, and a router is Layer 3.
• Encapsulation adds a header at each layer on the way down, and decapsulation strips it on the way up.
• The layers give engineers a troubleshooting map that isolates a fault to one specific layer.

Frequently Asked Questions (FAQs)