Introduction to Physical Installation of Servers: Server Form Factors

What is the difference between a server and a desktop computer? They have similar components. They run the same operating systems and software. Yet, a server is much more expensive than a typical tower. 

The truth is, there isn’t much of a difference. High-performance desktop computers often use the same level of hardware found in servers. Likewise, many servers have “server components” that offer the same level of performance found in a cheap Costco computer.

One difference is the business case. Servers are built for reliability and maintainability. For example, servers often come with two power supplies, two network ports, etc. If one fails, the server can fail over to the backup. 

Servers often include additional features designed to make administration easier. For example, many servers have a management module that lets techs manage BIOS settings remotely without restarting the server. 

Likewise, servers are built to utilize less space than a typical tower PC. More than a dozen servers can fit in an area the size of a kitchen cabinet (the two-door kind that hangs on the wall). Centralizing compute resources into such a small footprint makes running power and managing environmental controls for the IT hardware easier and cheaper. However, all of this comes at the cost of price and additional noise pollution.

Lastly, servers are built with better-quality parts, so they won’t fail as quickly. 

An Overview of Installing and Managing Server Hardware Components

In this video, CBT Nuggets trainer Knox Hutchinson covers server hardware components and the process of installing and managing them efficiently. 

What Types of Servers Are There?

Servers come in all sorts of flavors. They can be broken down into two basic types, though:

• Rackmount servers

• Tower servers

Do Servers Come in Towers?

We said that server computers are designed to use as little space as possible. So, why are there servers built in a tower form factor?

Not all organizations have racks for their servers. Likewise, many small companies only need one computer to power all the services they need. In these cases, hosting a server in a tower form factor makes sense. 

This isn’t to say that a tower doesn’t still adhere to the principles mentioned above. A tower server can fit an awful lot of power into it. While the tower may use a larger footprint than a typical 1U rackmount server, it can easily house more than double the horsepower. 

Rackmount Servers

Rackmount servers are computers designed to be mounted in a rack with one server sitting above another. It’s like stacking computers on top of each other in a closet but in a professional way. 

Rackmounts offer many benefits. For instance, cable management is far easier in server racks since most typically include special channels for running cables. 

Likewise, maintaining servers is easier since most rack-mounted servers can slide out of the rack for maintenance. Most rack-mounted servers are mounted on rails, much like a kitchen cabinet. The PC doesn’t need to be lugged to the closest workbench for repairs. Instead, techs slide the server out from the rack like a kitchen drawer or oven rack. 

How are Rackmount Servers Measured?

Servers are measured with a generic metric called a ‘unit.’ Units are a measurement of height. Each unit is approximately 1.75 inches. 

So if you go shopping for a new server, and it’s listed as a 1U server, the height of that server is about 1.75 inches. A typical rackmount server will come in 1U, 2U, and 4U sizes. 4U rack-mounted servers typically have more drive bays for storage or GPU compute units for machine learning. 

What is a Blade Server?

A typical server is a self-contained PC. They have their own CPU, RAM, power supply, networking hardware, etc. Unfortunately, this also means that a server closet needs more space for additional power cables and outlets. 

Blade servers are a bit different. They remove the reusable components from the server and host them from a chassis. The chassis typically includes power supplies, networking hardware, and specialized communication interconnects to hook blades together. 

Blade servers mount into the chassis, and the chassis mounts into the server rack. Blades can be swapped as needed. This allows for more customizability in the server configuration since blades can come configured specifically for storage, specialized processing power, etc. 

This all sounds great, so why aren’t blades more common? It boils down to cost and repairability. Blade systems are more expensive than a typical rackmount server. Likewise, they tend to use specialized components instead of off-the-shelf hardware. Repairing an end-of-life blade server can be much more costly.

Learn How to Install and Manage Server Hardware

Hopefully, this article helped you understand the difference between a desktop PC and a server. However, it barely scratched the surface. Servers come in all sorts of configurations to address different needs. Frankly, understanding the difference between the hardware used to build a SAN device and a machine learning compute module hooked me on hardware. 

Are you interested in learning more? Then watch Knox Hutchinson’s Installing and Managing Server Hardware Components training at CBT Nuggets. He explains why servers are built the way they are and what specific types of servers exist. He’ll also demonstrate how servers are installed and how to buy cheap, second-hand hardware to play with.