A business that requires a server is very dependent upon it and if it goes down then that business can be halted until it is brought back up. The hardware, software, operating system and backups all can be configured to give you a server that is not only stable, but one that can allow the administrator to know if there is a problem.
The components of your server should be chosen based upon what you are using the server for. How much memory,processor size, hard drive size, etc. all depends on the demands that you will be placing on the server. There can be some redundancy in the hardware to help prevent problems no matter what the needs on the system would be.
While the power supply isn't considered the most important piece to a computer, if it doesn't work the system is shut down completely. There are many options for power supplies that range from high efficiency to redundancy. A hot swappable dual power supply will keep the system up and running if there is a power supply failure. The system will start to beep to let you know that there is a problem, but it will continue to run without a hiccup. When the new component arrives you take out the failed supply and plug in the new one without shutting down the system.
RAID systems are a very common way to build redundancy in data storage. One RAID setup will mirror the data exactly between two hard drives. If one hard drive fails the system keeps running as if nothing has happened. Another RAID setup allows for data to be spread across multiple hard drives. As the data is split across drives it allows for simultaneous writing on the multiple drives increasing the speed of writing large files. A third configuration will not only write across the multiple drives, but it saves the data in a way that if a drive fails no data is lost and the system continues to run smoothly allowing the failed drive to be replaced. I have setup one office that uses the hot swappable drives as a backup system. At the end of the day they pull out a drive, put in a new one and tell the system to use the new drive is to be used to mirror the still present drive.
As you can see there are many things to take into account when designing a server for a business and I have only touched the surface. Even though there is a lot we can do to help keep a server running smoothly there are times when a problem can occur. For this reason we offer a monitoring program that we can install on y our server that informs us if there is a problem. Once that notice comes we can quickly remote in to see if it is a false alarm or if there is a problem. If a real problem exists we can inform you what we are seeing and make a plan to fix the problem. This service runs $30/month for monitoring a single system.