VIMRO IT Blog

4 Steps to Building a Fault-Tolerant LAN

Start with a clear roadmap for your LAN and be sure to plan for potential failure points along the way.

LAN

Building a network doesn’t need to be complicated if you have a smart plan in place. Network design should always begin with a well thought-out blueprint that takes into consideration any potential weak spots and accounts for them well in advance of problems actually happening. It is only by using the “Be prepared!” motto that your business can avoid costly fixes further down the road.

Besides just mapping out how to build the network, engineers and managers need to provide clear documentation so that other users can follow the intended usage if issues arise. It’s not enough for one person to know everything about the new network – that information should be communicated and accessible to anyone who might need access.

When planning your business’s Local Area Network (LAN), here are the logical steps to follow:

#1 Identify and Create a List of Potential Failure Points in the Network

The only way to reduce the impact of an issue is to be well prepared in advance. When it comes to problems, it’s never a matter of if, but rather when… so being ready is imperative. All the following variables should be considered, and a plan should also be laid out for all the different scenarios. Some questions to answer might include, “If there’s a failure, how will it impact the business? How quickly will they be able to recover?” A typical fault-tolerant system design takes the following into account:

  • Physical connections – This includes how the cabling is laid out, how it’s being run and the location of the room
  • Wiring the switches – Involves how the switches are wired in relation to the cabling
  • Network redundancy – How much uptime the network is expected to provide
  • Network services redundancy – Typically refers to applications such as web servers and business applications

#2 Maintain Physical Connectivity

The connections must always have backup in place in order for the system to continue functioning properly – even in the event of a failure. All switches should have two different wires going to alternate switches, so if one fails, there will still be a path for network traffic. To be considered fully redundant, the switch should go to other switches (this is called a full mesh configuration).

To truly take advantage of redundancy at the switch layer, the system should be connected to multiple switches. If the server network adapter (NIC) is connected to two different network switches , it will stay up and continue working – this is called multi-homing. Physical redundancy is not just switches. Severs and critical systems should also be multi-homed and multiply redundant.

#3 Plan Out Wiring and Cabling

Cabling should always be laid out according to industry standards and should never be left exposed. Cables left loose on the floor pose a tripping hazard and make it more likely that cables get pulled.

Besides just being properly laid out, all cables should also be meticulously labeled. Proper labeling makes troubleshooting faster and easier. Any office should have an accurate cabling diagram that illustrates the physical path of all cables in the building.

#4 Create Network Redundancy

Network redundancy goes beyond the physical component. A layer 2 or 3 redundancy protocol can be used to detect when an individual switch is down and automatically send network traffic over a different switch to minimize the chances of downtime.

  • VRRP/HSRP – transfers IPs and gateways to different routers when connectivity is lost in one
  • Dynamic routing – changes network routes to accommodate for failed device; automatically removes it and redirects the network traffic
  • Redundant default gateways – main network redundancy is accomplished here
  • VLANS – offer flexibility and redundancy when integrated with dynamic routing
  • Network services redundancy – refers to actual applications, such as DNS or WINS

Some additional considerations include cooling, electrical components, vendor support and potential operator errors. It’s vital to keep the equipment at the proper temperature – if it overheats, no amount of redundancy can help the situation. Vendor relationships should include valid contracts that would help to get any situations resolved promptly.

Proper security measures and restrictions need to be in place to keep unauthorized employees from typing incorrect commands and potentially shutting down the entire network. Operator training and strict privilege-based access controls are a common way to combat these types of issues.

Overall, building a functioning network requires a good plan and follow-through. Rather than just hoping everything will go well, having a system in place for issues and a clear understanding of what to do in each situation will help keep outages to a minimum.

Find out what makes cabling so difficult here.

Share this post:
Facebook
Google+
Twitter