Foundational Focus: OSI Model – Breaking Down the 7 Layers
The OSI model is a conceptual tool used to discuss and describe network functions. The use of a standard reference model is essential to communicate ideas as well as create new technologies. It is a good idea to be familiar with the OSI model, the features assigned to each layer, and examples of common protocols or technologies associated with the OSI layers.
The OSI model has been the de facto reference model for networking protocols since the mid-1990s. The OSI model is formally known as Open Systems Interconnection (OSI) model ISO/IEC 7498-1. International Organization for Standardization (ISO), is a global standards-settings group comprised of members from various national standards groups. International Electrotechnical Commission (IEC) is another global standards-settings group; however, it focuses on electrical, electronic, and related technologies. The IEC works closely with ISO and other groups to establish standards for computers, networking, and communications. The OSI model is a conceptual model rather than a technical specification. This means it is used to discuss, describe, compare, and contrast actual technologies rather than directly mandating elements of technology.
The OSI model is comprised of seven layers, with layer one positioned at the bottom of the layer stack, and layer seven at the top. The layers have assigned names as well as number references.
The OSI model is used to describe the function and purpose of the various elements in network communications. In theory, data is received by the protocol stack into layer 7, the application layer, from software. The received data is labeled as a service data unit (SDU). Each layer adds its own layer-specific header to the SDU, thus creating a payload data unit (PDU). The PDU is then passed down to the next layer below, where it becomes the SDU of that layer. This process of traversing down the layer stack is known as encapsulation. Once layer 1, the physical layer, receives the PDU from layer 2, the data link layer, the data is transmitted over the network medium (i.e., twisted pair cable, fiber optic cable, or wireless).
When a network interface receives a signal of data from the network medium, it processes the PDU in reverse. This reverse unpacking process is known as de-encapsulation. Each layer reads its corresponding header of the PDU, processes and removes the header from the PDU, creating an SDU, then passes the SDU to the layer above. This is repeated until layer 7, the application layer, receives its PDU and passes the actual data to software.
While the generic term for a header and payload data for a given layer is known as a PDU, some layers and/or protocols have unique names for this structure. These include:
4 - Transport - TCP - segment
4 - Transport - UDP - datagram
3 - Network - packet or datagram
2 - Data link - Frame
There are two additional oddities to the process of encapsulation. The first is that the data link layer often adds a header and footer to the SDU to create its PDU. For example, the most common data link layer technology is Ethernet. Ethernet adds a header containing the destination MAC address, source MAC address, and EtherType designation (i.e., identification of the type of payload), while the footer includes a checksum value to perform integrity checks. The second is that most layer 1 physical layer technologies do not add any headers (or footers) to the PDU from layer 2. Instead, start and stop delimiter bits might be used on asynchronous communication (i.e., not time-synched), but these are not considered to be part of the PDU, just part of the transmission technology.
The header added by a layer is configured to include information relevant to the same layer on the receiving system. This layer-to-layer communication via header (and footer) is known as peer layer communications. It is essential that headers (and payload) arrive uncorrupted and are unpacked (i.e., de-encapsulated) in the correct order. Fortunately, communication errors are rare, intentional corruption is detectable, and most systems use standardized protocol stacks, such as TCP/IP, thus peer-layer communication is generally flawless.
Layer 7 - Application
Layer 7, the application layer, is the interface between the protocol stack and application software. The software might be client utilities or server services. It is the ability of software to communicate with the standardized interface of application layer protocols that makes network communications possible. In fact, the use of common application layer protocols allows for fully interoperable computer communications. The application layer is assigned the responsibility to check whether a remote communication partner is available, confirm communications with that partner are possible, and evaluate whether or not there are sufficient resources to maintain a communication.