How Network Design Works
The specifics of your entire network design stage will vary based on size and complexity.
However, the approach of solid network design is to take in every relevant input, like budget, objectives, and future scope, and output a best-fit network diagram.
#1: Identify the Requirements
Before you begin any network design project, start by gathering information and developing clear business and technical requirements. Without clear targets, the rest of the design process will be trickier than it needs to be.
Some examples of high-level business objectives are the following:
- Opening a new office.
- Improving end-user experience.
- Cutting network infrastructure costs.
- Complying with a new industry regulation.
- Improving uptime.
Ideally, by the time you’ve started designing the network, the work should have already been done to identify precisely which business requirements you’re looking to achieve. Following on from this, it’s vital to have an understanding of the technical requirements. This includes:
- Security requirements
- Bandwidth options
- Uptime SLAs
- Specific Protocols
The chief of these constraints is budget, but you should also consider any technical constraints, such as legacy systems that will need continued support.
#2: Assess the Current State of the Network
In almost all cases, network design is an iterative process. While this can add to a feeling of complexity, it often presents a silver lining in the form of preexisting elements—successfully retaining and reimplementing systems into the new design is an opportunity for budget and skill retention.
With new requirements in hand, perform a close inspection of the current network’s status. This phase’s goal is to develop a clear understanding of:
- Network performance
- Key applications
- Both the physical and logical layouts
Automated network mapping tools and security scanners are often essential for pinpointing the current network’s data flows and security considerations.
#3: Choose Your Network Topology
You can begin sketching out the important aspects of the new network layout. As you’ll see in the best practices section below, physical design is just as important as logical design.
Physical Design
Draft a graphical, top-down view of your new network. This topology helps visualize the different paths that data can take across a network, and helps determine where network hardware will be needed most. For example, a star topology diagram can indicate how the data flow is centralized and therefore find the most strategic positions for physical network devices.
This extends to how much the copper and fiber cabling it will need. From this, draw up a figure of switch ports required, Wi-Fi access positioning locations, and cooling requirements for any in-house server stacks.
Throughout this process, keep network security best practices in mind.
Logical Design
Logical design refers to the way in which data flows to and from devices. A key consideration within logical design is IP addressing: each address needs to uniquely identify a device within the network. Well-structured IP addressing allows for ranges of devices to be logically and neatly assigned to suitable groups.
As an organization grows, this quickly becomes a jumbled mess – making it a great focal point when designing a new network.
#4: Select Hardware and Software
With the sketch of your network topology and an idea of the devices it will need to run, select the network devices that will achieve this. Trace the flow of data packets from endpoint to server, establishing:
- Where will need a router
- What access points will need to be connected
- What switches to rely on
Software must be a final key consideration – the right operating system and applications need to align not only with the new network’s capabilities, but also ideally match your IT team’s current skill sets. Network Function Virtualization (NFV) now allows for services such as network optimization and firewalls to be offered as on-demand software, so the options are greater than ever.
Following this, it will be a case of implementing this new design, and assessing its in-production performance. After making any necessary tweaks and changes, it will slowly be rolled out in full.
The 3 Network Design Best Practices
By sticking to a few key best practices, it’s possible to build security, performance, and high-scalability into the very design of your new network.
#1: Modularity
Simplicity is at the core of network design: sometimes, however, enterprise networks need more bespoke or complex setups. To retain a degree of modularity while still providing the right services for the right devices, it’s best practice to split things into their simplest network components.
The most common application of this is the three-layer model:
- Core. a way to quickly and efficiently sort how packets are sent to distribution switches.
- Distribution. To keep each LAN connected, you need a distribution layer – made up of distribution network devices like switches. A lot of smaller networks stop at this, but as networks grow in scale, the demands made on these distribution switches can start to reach breaking point.
- Access. The access layer is the one that users directly interact with – it connects devices such as PCs, printers, and IP phones, to a network, such as a LAN.
This basic segmentation allows for a solid foundation that sets a base level of reliability and performance. Potential issues can be isolated and resolved more easily without impacting the entire infrastructure.
Beyond this, it’s possible to introduce more logical segmentation – like a low-bandwidth access module that can provide connections for low-bandwidth workstations and printers used in the specific departments, like accounting, as well as the necessary switches in the distribution layer.
Greater segmentation allows for more granular access controls, and therefore security, as long as you keep it as simple as possible.
#2: Fault Tolerance
Fault tolerance ensures network availability and reliability even when components fail. By building in redundancy, the network can continue operating without service disruption, which is critical for business continuity.
At the design phase, then, it’s good to incorporate failsafe technologies for IP networks like Hot Standby Routing Protocol (HSRP) – this allows user traffic to immediately recover from failures in edge devices or access circuits. It works by giving two or more routers a single IP and MAC address, letting multiple routers act as a single virtual one.
This group continually exchanges status messages, and when one goes down, hosts are simply able to forward IP packets to another within the same IP and MAC address group.
Other ways to build resilience into the network’s design is:
- Virtual Router Redundancy Protocol
- Link aggregation
#3: Security & Management
Robust security measures are vital from the outset – firewalls, access control lists, and network segmentation are all key components of this.
If you’re reliant primarily on firewall policies, consider how best to micro segment the network. If your security team would benefit from a single security platform, consider how application logs and deep packet inspection can be safely connected up to a central security tool.
Running parallel to this is a plan for maintenance: detail the monitoring and troubleshooting processes this new network demands, alongside a schedule for ongoing performance and vulnerability checks.
Feedback