A Brief History of the OSI Model and its origins

History of the OSI model
7 min

A brief history of the OSI Model

In a world of diverse networking standards, implementations, protocols, and applications, the Open Systems Interconnection (OSI) model offers a handy common-ground representation. It unifies different communication systems into an abstract for a hierarchy that makes it easy to understand, teach, and learn how to do networking effectively.

Of course, that’s just scratching the surface.  

Because our network monitoring software leverages automated network topology mapping and diagrams, we wanted to cover more about the history of the OSI model. Currently, our topology maps display elements from; Layer 1 (links and physical devices), Layer 2 (MAC addresses), and Layer 3 (IP addresses). 

Here’s how the OSI model began and how it stacks up against the alternatives. Read on for the history of the OSI model.

History of the OSI model

In terms of the history of the OSI model, there are a few discrepancies. There is disagreement as to how the OSI model actually began. By the late 1970s, the world would recognize that there was an increasing need for standards in terms of how connected things communicate. 

Researchers from France, the UK, and the USA, began two projects to develop just that. These groups would attempt to develop a set of standards for the computer networking space. These projects would aim to become the standard for computer networking and interoperability amongst vendors and device manufacturers. 

One team working on this was the International Organization for Standardization (ISO). The International Telegraph and Telephone Consultative Committee, (CCITT) undertook the other. Each of these bodies produced a document attempting to standardize the way computer networking protocols would take shape in the future. 

Merging of the OSI models

1983 saw the merging of the two documents and the formation of The Basic Reference Model for Open Systems Interconnection. The OSI model was meant to be the industry standard for the basis of computer networking and the Internet. But, not everything went according to plan. The idea behind OSI was to get everyone in the industry to agree on standards for interoperability across vendors. At the time, many devices and networks in general were leveraging and supporting many different protocols. Basically, there were a lot of devices cropping up and many were speaking different languages. 

In the next years, in the 1980s and 1990s, the TCP IP model began to make headway. There was a lot of division as to what was best for the purpose, the OSI model or the TCP/IP model. 

The OSI model never got the traction amongst vendors. While these papers were being written the TCP/IP model was making major headway amongst vendors. More and more vendors started to implement the TCP/IP model as the means for interoperability. 

The OSI model is still in use for teaching and documentation, but not in real networking implementation by vendors. 

About the OSI model

The OSI model characterizes and standardizes the communication functions of a telecommunication or computing system.

This model is layered. In other words, it’s composed of several tiers of abstraction that describe networked communication at various scales and levels of detail. For instance, whereas one layer might look at the higher-level flows in an application that sends data between the nodes in a cluster, another might zoom in on the byte-structured contents that make up the packets.

The lower you go in OSI layers, the closer you get to the hardware itself. The higher layers go in the opposite direction, towards the application.

The original version of the model defined seven layers. And no, it’s not a reference to some famous vision of hell – although the humor might fit depending on the state of your network.

The International Organization for Standardization (ISO) first published the confusingly similarly abbreviated OSI model in 1984. At the time, experts simply thought seven layers would be sufficient to encompass all aspects of any conceivable networked system.

So did their predictions turn out to be correct? Well, if it’s any clue, the OSI model is still used today. It’s commonly used as a reference for describing network protocols, training IT professionals, and interfacing multiple architectures. Since its inception, however, IT professionals have gone back and forth over its merits compared to alternatives.

History of the OSI model: Core Ideas and Definitions 

OSI is a service definition that gives an abstracted meaning to the way entities interact across layers. This differs from communication protocols that offer a concrete technical definition of how messages should propagate within a single layer.

Imagine you have two servers that need to share information. The message doesn’t just magically teleport from an app on the first machine to the app on the other. Instead, it transits down the layers and eventually reaches the transmission line. Once it makes the jump across the gap to the other device, it has to repeat the process in reverse by ascending layers until it reaches the receiving application.

For any starting number N representing a layer that transmits a message, the OSI model can be used to explain the journey in terms of a few key concepts:

Protocol Data Units (PDUs) are abstracted messages that include payloads, headers, and footers.

Service Data Units (SDUs) are equivalent to the payloads.

At each subsequent transition from some layer N to some layer N-1, a layer-N PDU becomes a new N-1 SDU. This payload then gets wrapped up in a layer N-1 PDU with the relevant headers and footers. On the opposite end, the data passes up the chain, unwrapping at each relevant stage until it’s just a payload that can be consumed by the corresponding layer-N device.

Standards of the OSI model

Layer 1: Physical Layer

One nice thing about the OSI model’s longevity is that there aren’t a lot of different standards to keep up with. You’ll find everything you need to know in ISO 7498, which explains the model, its security aspects, naming and addressing standards, and management practices.

The physical layer defines the electrical and physical specifications for devices. This includes the media, interface, and transmission mode.

This is as close as you’ll get to the bare metal in OSI terms – At this level, it’s all about the bits.  

Layer 2: Data Link Layer

The data link layer is responsible for the error-free transfer of data frames from one node to another. Networks achieve Layer-2 transmission by using medium access control (MAC) addresses to determine how devices should handle and access data. They also depend on logical link control (LLC) frameworks to encapsulate different protocols.

Data link layers also provide flow control and error control. Some examples include the common Wi-Fi, ZigBee, and Ethernet standards.

At this level, it’s all about frames.

Layer 3: Network Layer

The network layer is responsible for routing packets between nodes. It typically corresponds to the use of IP addresses and other logical addressing schemes to make routing decisions. Network protocols like multicast group management, address assignment, and routing are all associated with the network layer.

At this level, it’s all about packets.

Layer 4: Transport Layer

The transport layer is responsible for the end-to-end delivery of packets. It provides mechanisms for error-free delivery, flow control, and congestion control.

The OSI model draws distinctions between different transport layer protocols based on their fundamental features, such as their connection modes, reliability, timeout retransmission capabilities, and multiplexing support. According to most experts, the TCP and UDP standards fall under Layer 4 even though their original design weren’t to conform to the OSI model.

Layer 5: Session Layer

The session layer is responsible for the setup, maintenance, and termination of communication sessions between applications. It can handle multiple connection directionality modes, and in many applications, it’s implemented on a case-by-case basis to handle specific data flows, like streaming media.

Layer 6: Presentation Layer

The presentation layer is responsible for translating between different application formats. This is where jobs like encryption, decompression, compression, decoding, and encoding take place. The presentation layer, a.k.a syntax layer has high-level semantic workloads.

Layer 7: Application Layer

The application layer is the highest tier in the OSI model. It provides an interface between the application and the network and includes the high-level protocols you know and love, such as HTTPS, NTP, IMAP, DNS, SNMP, and your preferred assortment of OS file-sharing standards.

When building on the OSI model, it’s important to remember that Layer 7 also distinguishes between application logic and the layer itself. As such, you can further divide Layer 7 into two sublayers:

The common application service element sublayer (CASE) supports widely used services like reliable transfers and remote operation.

The specific application service element sublayer (SASE) includes a range of protocols particular to each application, like Remote Database Access, Distributed Transaction Processing, and Virtual Terminal.

Cross-layer Functions

As with most abstract models, OSI isn’t quite perfect, but it tries to make amends for its oversights. One of the ways it does so is by recognizing cross-layer functions. This means services that can impact multiple layers without explicit restriction to any of them in particular.

Cross-layer optimization eliminates the restrictions of the classic OSI model by accounting for situations where you might need to communicate between levels. For instance, you can use it to make the quality of service tweaks in one layer based on the feedback you get from another. Or your network monitoring system might modify data link layer behavior to minimize congestion based on what’s happening at the application layer.

 One caveat to remember about cross-layer functions is that they can make your life as an admin way more complicated. The whole point of these deviations is that they have useful side effects on other aspects of the model, but this is a double-edged sword. It pays to exercise caution – and trust in your dependency graphing tools – when designing around cross-layer functions.

OSI vs. the TCP/IP Model

The TCP/IP model is an alternative set of protocols that overlaps with much of what the OSI model does. Historically, however, it’s gone through a lot more changes and tweaks.

 Some major differences include that:

  • The OSI model uses seven layers compared to the four levels of TCP/IP,
  •  The OSI model is more abstract, while TCP/IP is more concrete,
  • Whereas TCP/IP was initially a defense agency initiative, OSI evolved in the industry world, and it never quite gained the same popularity, and
  • The OSI model is more comprehensive, covering all aspects of networking, while TCP/IP focuses on the specific task of connecting computers.

How does the TCP/IP model cover everything in fewer layers? For one thing, it groups layer 5-7 together in a single application layer. Experts haven’t quite agreed on how to interpret the distinctions between the two protocols at the lower levels, like the link layer, but that hasn’t stopped them from using TCP/IP to build robust enterprise architectures.

Learn more about the OSI model vs. TCP/IP model.

Doing More With the OSI Model

This sums up the history of the OSI model. The OSI model can help you rethink your approach to network architecting and management. It couches the defining characteristics of networks in common terms that anyone can understand. In the process, it serves an important role not only in academic training. Additionally, it serves a role in bringing new admins up to speed in corporate environments.

Of course, it also helps to contextualize the feedback you get from your network. This is why so many MSPs, IT professionals, and sysadmins turn to Domotz network monitoring software.

With dashboards that tie abstract data to real-world signals, Domotz empowers you to monitor networks as events unfold – no matter how you prefer to conceptualize them. 

Share via Social Networks

You might also like…

Read more top posts in this category

Want more tips on Network Monitoring?

Ready to get started with Domotz?

  • Powerful
  • Automated
  • Simple
  • Affordable
Start Your Free Trial Contact Sales