Weather building a VoIP system, or a data network that requires minimal down time it is always important to build in redundancy into the design and implementation of your network.
I recently had an encounter with a multi-national company, which shall remain nameless, who had a system failure and was unable to assist me because of this failure. In the business that this company is in, I found this to be extremely disturbing that a single point of failure could bring this company to its knees until the trouble was isolated and resolved.
When building a VoIP network, if you are using a hosted PBX provider the only redundancy that you need to take care of locally is internet connectivity. A multi-WAN router with multiple internet connections, and battery back up power should be all of the on-site redundancy that you need to keep a single point of failure from bringing all of your phones down.
For data networks, redundancy gets a bit more complicated, and expensive. Multiple servers, constantly mirroring each other, as well as redundant internet connectivity, and redundant power sources (battery back up, diesel generators) should be implemented. For national, or multi-national companies redundancy should not stop at a single data center location - multiple data center locations should mirror each other, to prevent a natural disaster or other catastrophic loss of one or more data center from losing any data.
What needs to be considered is not the up front cost of creating or maintaining this redundancy, but what the cost of failure could be upon the business, or its clients. The loss of data, or the ability to retrieve, or input data into a system, even for short periods of time can result in not only immediate financial losses, but also lead clients to consider going elsewhere to conduct business.
A failure in the evening, weekends, or holidays may not immediately be discovered by the company until the beginning of the following business day, and then may even take some time to have the system back and running to full capacity. If that system is accessible, or used by clients or customers during these off hours they will have no way to access the information stored on that system until the system is fully restored, which in turn can cause them to have financial losses, or lost business/customers as well. It trickles down from one customer to another, and the customers who have a financial stake in your network always being available to them will look elsewhere for the services that you provide, and these are likely the larger clients that you want to retain as customers.
So remember:
Redundant internet connections
Redundant power sources for all network/servers
Redundant servers
And for larger companies:
Redundant data centers
This is not the place to cut corners and attempt to shave a few dollars off of your operating expenses, a failure could end up in loss of revenue and customers, and even potentially bankrupt or put you out of business.
