Many networks grow organically over a period of time and at some point work is required on them. The question is when is the right time to take stock and decide that a proper design is required.

The problem with organic growth is that often switches are added and simply daisy chained together, leading to large flat networks. These can quickly be overcome with large amounts of broadcast traffic and IP address ranges run out of space.

The lifecycle approach examines the network on a periodic basis to check on the requirements for the network and the ability of the network to fulfil these. Based on the requirements a network design is put in place and the network built to this.

Organic growth will only work for a period – at some stage a hierarchical network design will be required and when it is you need to understand what’s connected to your network and how.

So what are the rules?

1. A properly designed network is built with a hierarchy: Cisco’s typical design is for a core, distribution and access layer network. In practice this is often overkill, leading to the ‘collapsed core’ network – which just has two layers – core/distribution and access layer.

2. The network should be able to expand and grow: it will not outgrow itself in any area as other areas start to grow. A non-hierarchical network (typically a small network that has grown larger simply by tacking a little extra on each time it is required) is not scalable – it will reach a point where some aspect of the network cannot grow larger either due to capacity restraints or because of impact on network performance.

3. Scalability is important: a hierarchical network is scalable because it is built in a modular fashion. As a modular reaches maximum practical capacity, another module is added.

4. Security: any network must be secure, but its not just a case of a firewall at the edge and forget about the rest. Think about secure protocols running through the network – SSH and SNMPv3 being ideal for managing your equipment. Use FTP and not TFTP to upload images.

5. Reliability: as soon as you put a network in place, people start relying on it. A network which is not built sufficiently robustly – or conversely is so complicated that it can never be fixed if it fails are a recipe for severe pain down the track. Consider duplication of core equipment – possibly between different communications rooms – and don’t forget air conditioning, UPS and diverse fibre runs between them.

A modular network typically employs less single points of failure and faster failover / convergence times. The result is a network that causes you a lot less stress to look after. The other key point to note is network management. If you have a highly resilient network, but you don’t know the baseline, its going to be difficult to spot when an issue has occurred.

In summary: design your network properly and get to know it when its running and you’ll be able to get the maximum uptimes.

IPTel Solutions are an Australian based company
providing the design and delivery of Enterprise computer networks. Specialising
in Wireless, we also work in Routing & Switching and Voice. Visit us at
www.iptel.com.au

IPTel Solutions are an Australian based company providing the design and delivery of Enterprise computer networks. Specialising in Wireless, we also work in Routing & Switching and Voice. Visit us at http://www.iptel.com.au

Author Bio: IPTel Solutions are an Australian based company
providing the design and delivery of Enterprise computer networks. Specialising
in Wireless, we also work in Routing & Switching and Voice. Visit us at
www.iptel.com.au

Category: Computers and Technology
Keywords: cisco, network design lifecycle, network design