Moving from a small network of interconnected PC’s to a server environment can bring many benefits. If can also be very expensive. This is often the reason that home users and small office environments often continue to struggle along with a workgroup environment when they could benefit from having a server. When people think of servers, they generally think of very expensive machines that reside in a locked room somewhere and are managed by professionals. In large scale environments with dedicated servers, this is often the case, but in a small office, or even home environment, someone with moderate technical ability can easily build and maintain a small server.
When building a small server, instead of thinking about server hardware, explore your options in standard PC hardware. Even in larger environments, due to the rapid growth in the technology field, it is not uncommon for workstations to actually be more powerful than servers. Since this is the case, why not build a server from standard PC parts? Think about what you are looking for in a server. In most home or small office environments, the main reason for having a server is data storage and sharing. This task can easily be managed by PC class hardware.
Start by selecting a case. You will want to choose a slightly larger case with adequate ventilation. This machine will most likely run 24 hours a day, and will create more heat than most PC’s so you want to make sure it is able to cool properly. You also need to ensure that your case has at lease a 500w power supply. Next you will want to choose a motherboard and CPU combo. I recommend Intel motherboards and CPU’s. Intel manufactures many boards to natively support RAID which is important for a server. I would recommend a mid to upper range CPU to extend the life of this server. For a simple file server, 4 GB RAM will be plenty, and this will allow you to install a 32 bit operating system. Since tasks preformed on a server will not be graphics intensive, choose a motherboard with integrated graphics to save the expense of a video card.
Hard drive selection is mainly what will differentiate this from a standard PC. You will want to choose hard drives with enough space to allow for data growth. File sizes continue to grow so always plan for more space than you think you will need. Make sure to choose enterprise class drives that are designed to operate 24/7. Depending on the amount of storage space you anticipate needing will help you decide how to configure your drives. If a single drive will provide enough storage space for your needs, including anticipated growth, you will want to purchase two drives and configure as RAID 1, or mirrored. This will make the two drives appear as one drive, and will actually make two copies of every file in case one drive fails. If it is not possible to purchase one drive large enough to meet your needs, you will want to purchase additional drives, and configure them as a RAID 5 array instead. This configuration uses parity to protect your data, and allows your data to be rebuilt in the event one drive fails. To configure RAID 5, you will need to purchase one more drive than you anticipate needing to act as the parity device. For instance, if you anticipate needing 2 TB storage space and choose 1 TB drives, you will need to buy three hard drives to configure a RAID 5 array.
The next decision you need to make is how you intend to backup your data. A RAID array helps maintain the integrity of your data in the event of a hardware failure, but should not be considered backup. RAID will not protect your data from the devastating effects of a virus, or a user accidentally deleting or overwriting important files. Depending on your environment and needs, you should consider an external hard drive, or tape backup. Tape drives can be very expensive when compared to external hard drives, but are better if your needs dictate taking data off-site for more secure storage. If you do choose an external hard drive, purchase two drives and rotate them so you always have a copy of your data that is not connected to the server. Your backup is no good if for example a lightning strike takes out your server, and the external hard drive that is attached to it.
You also need to choose an operating system. I recommend Microsoft Windows Server 2008. This will be one of the more costly components of your server build, but it has many advantages over installing a workstation OS. The main advantage is it allows you to configure your server with Active Directory which allows for central management of accounts. Rather than have to duplicate a user’s account on every workstation they may need to use, you create the account on the server, and then can then login to any workstation that is a part of the Active Directory domain. The domain model makes it much easier to manage data security as well. You can create groups and add user accounts to these groups to simplify management of data. You then create shares and assign permissions to groups of people to allow them access to the data they need while baring them from access to data they should not see. For example, you could only allow the accounting department access to payroll data.
Choosing to move to a server environment can be a hard decision due to the considerable expense and configuration required. By building a small server using PC class hardware, you can gain all of the advantages while minimizing expense. After the initial configuration is complete, which may require outside help depending on the ability level of your users, day to day management tasks are not drastically more complicated than maintaining a standard PC. Since users do not work directly on a server, they tend to have drastically fewer problems than user workstations are are generally easy to maintain. Do some reading, evaluate your needs, and consider setting up a small PC class server to simplify your computing environment.
