Network Architecture

Network architecture refers to how computers are organized in a system, and how tasks are allocated between these computers. It is the organization of an enterprise’s computer network infrastructure designed to achieve the goals of the organization.

Network Architecture Primer

Network Architecture Definition

Network architecture refers to how network elements are organized in a system, and how tasks are allocated between and across those elements. It is the complete physical and logical design of an organization’s network infrastructure often represented as a map or schematic diagram.

Designed by network architects, managers, administrators, engineers, and design engineers, the network architecture is a network’s functional organization and configuration, from a signal’s generation to its termination. It details the interconnected physical components of a network (topology), the operational principles, procedures, and protocols governing how the network functions, and the media used for data transmission. 

The design of the architecture itself determines the services it can deliver. The more adaptable the network architecture is to the evolving needs of an organization, the more useful (and indispensable) it becomes to its users. For example, bandwidth capacity is a significant consideration to achieve reliability, scalability, and flexibility. 

Understanding the rudiments of network architecture aids in grasping the scope of network designs, as well as easing the day-to-day management of networks. These rudiments will be discussed in details, below.

Principles of Network Architecture

The design of networks should leverage the following common set of architectural principles:

  1. Hierarchy
  2. Modularity
  3. Resiliency

Hierarchical Design Principles

A hierarchical network design tries to divide complex flat networks into multiple smaller and more manageable networks. Each level in the hierarchy is focused on a specific set of roles. This design approach offers network architects the flexibility to optimize and select the right network hardware, software, and features to perform specific roles for the different network layers. 

Common network layers include the Core layer, Distribution layer, and the Access Layer. The Core layer focuses on the transport of communications between sites and high-performance routing. The distribution layer provides policy-based connectivity. The Access layer provides workgroup/user access to the network. 

Modular Design Principles

Modular design makes the network more scalable and manageable. As a result, network changes and upgrades can be performed in a controlled and staged manner, allowing greater stability and flexibility in the maintenance and operation of the network. 

The extra cost and challenges of operating a modular network are mitigated by the design’s ability to adapt to future business expansion plans. Consequently, focusing only on the capex without considering opex will likely cost more in the long run if the network was not architected and designed to meet current and future needs. Business-driven network design should always follow the principle of “build today with tomorrow in mind.” 

Resilient Design Principles

By utilizing the design principles of a modular and hierarchical design the network by definition has already gained a measure of resiliency simply by being able to identify and manage any network issues within a particular layer or module. Adding additional components for redundancy and avoiding the creation of single points of failure will also add to the network’s resiliency. Finally, on-going monitoring and optimization by the network operations team will help identify issues earlier before they bring the entire network to a halt.

The OSI Model

For management and maintenance purposes, understanding where a network issue is occurring is critical. The Open Systems Interconnection (OSI) Reference Model or OSI Network Model provides a conceptual model of how and where network communications are occurring. 

Layer 1 – Physical:

The physical layer aggregates all the tangible components in a network system. This includes the electrical, mechanical, and physical parts of the network such as cables, bus, radiofrequency, voltages, and other physical requirements. 

All the components classified under Layer 1 could be referred to as the transmission media, as they are all involved with the transmission of raw bits over a communication channel. Additionally, this layer defines the network topology and the mode(s) through which media is to be distributed in the network.

Layer 2 – Data Link: 

Here, different network interface hardware on the network system relate to encode, decode, re-encode and transfer data between network nodes. As well, this layer enables devices in the identification of each other while establishing communication with another node. A network card’s MAC address is a good example of a data link layer address, with which a computer can be identified on a network, during interaction with other devices. 

In summary, layer 2 allows for the transfer of data from one node to another. Some of the protocols applied in the LAN data link layer by different vendors include Ethernet, ARCnet, AppleTalk, Token Ring, etc. In cases where a network communication extends to the internet from the LAN, data link protocols like Serial Line Internet Protocol (SLIP) or Point-to-Point Protocol (PPP) might be used instead. 

Layer 3 – Network: 

Communication across different physical networks (such as routers) is enabled in this layer, through the use of a secondary identification layer like an IP address, as with TCP/IP networks. Internet Protocol addresses mostly serve as unique identifiers on IP-based networks. 

Layer 3 is the most distinct layer during interactions with TCP/IP networks. It is essentially the layer where data packet forwarding occurs through routers. Both connection-oriented and connectionless services are provided on this layer, allowing for interconnection across heterogeneous networks. 

In a nutshell, some of the essential functions allowed in this layer include:

  • Routing of data via available pathways
  • Decongestion of bottlenecks in the sub-net
  • Accounting; ensuring the number of bits sent and received during a transmission tally.
  • Layer 4 – Transport: 

The transport layer is where data transfer is coordinated between end systems and hosts. The best-known example is TCP/IP – Transmission Control Protocol, built on top of the Internet Protocol.

Layer 4 ensures consistency during data transmission by

  • Retransmitting lost segments 
  • Correcting bit-level errors
  • Fragmenting traffic
  • Sequencing and reordering packets for easy assemble

Alternatively, there is the User Datagram Protocol (UDP) which is schematically lighter, when compared to TCP/IP Protocol. UDP is used for multimedia transmission, as well as for other low over-head transmissions.

Layer 5 – Session:

In this layer, one device communicates with another in a coordinated “session” within defined response times. Essentially, the session layer has the primary function of allowing for a connection to be established. The connection then gets shut down when its purpose has been accomplished or an allotted time has been exhausted.

In some cases, this layer may allow for the authentication of the users in a connection process. Other functions include synchronization and the establishment of one way or two-way communication.

Layer 6 – Presentation:

The presentation layer is where data is “presented” to an application or the network. It is also where an application format is prepared for or translated to a network format, and vice-versa. Additional functions include encryption, data compression, and reformatting

Layer 7 – Application:

The application layer is the layer that users see, it’s the one with which they interact with other users, servers, the computer, etc. The interactions could be via software applications such as Microsoft Office or Google G Suite. The application layer allows applications to access network services by providing a set of interfaces that supports the operation. 

Maintaining 100% Visibility across Enterprise Network Architecture

Regardless of the network architecture that your business requires, constant performance monitoring is necessary to ensure that all critical components are at their optimum performance 24/7. Having that performance visibility across all the domains of a given network architecture increases the enterprise’s to maintain network uptime and productivity. 

LiveAction is the leader in network performance management providing network monitoring and diagnostics across all layers of the network architecture in different capacities. Used and trusted by NetOps in the largest organizations in the world, LiveAction makes it easy to identify and quickly resolve issues anywhere in the network proactively.