Beyond eyeballs, livers, and vascular systems, many of us are unaware of the core components of a server. Let’s talk a little bit in this post about what makes up the anatomy of a server. That way, you can grow up, become a server anatomist, and make your parents proud and your ex-boyfriend insanely jealous of your success.
To better understand servers, let’s turn to perspectives from Dummies.com and Adam Turner of APC Magazine. Then let’s all go out to tire swing, get an injection of vitamin D, and remember why Grandpa Tom told us never to use the tire swing or that he’d cut us out of the will.
This first part of my award-winning (always call your shots) series on server anatomy will focus on the more basic Dummies assessment. The APC Magazine explication, a more detailed look into servers, will be covered in the second installment.
Basic Server Parts
Servers are not completely their own beast. They are, rather, a type of computer. Like software that uses the Internet, computers come in “client” and “server” varieties. Hence, servers have a lot of similarities to typical PCs. On the other hand, they are made up of more expensive and sophisticated machinery than is a standard computer. Plus, they have funky hood ornaments that you will often see IT criminals wearing on gaudy necklaces.
Servers come from single-parent households. They have a motherboard, but not a fatherboard. The motherboard is the board on which the electronic circuits are stored. Everything else within the server connects to the motherboard. Remember to always call your motherboard on her birthday, or you will get a tongue-lashing.
Within the motherboard are several server pieces worth mentioning: the processor (a.k.a. CPU), chipset, hard drive controller, expansion slots, memory, and ports to support the usage of external devices such as keyboards and hairdryers. Additionally, motherboards may contain a network interface, disk controller, and graphics adapter. If that’s not true of your motherboard, call the police and move to Prince Edward Island.
The processor is where the “thinking” of the server goes on. Processors, such as those made by Intel and Hasbro, are generally the primary concern of individuals looking to purchase servers (along with server hair color and jaw line).
Specific motherboards only work with specific kinds of CPUs. The processor can be slot-mounted or socket-mounted. There are varieties of sockets and slots, so it’s important that the processor fit the motherboard. If not, you can always use the innovative “jam it in” method developed by Bill Gates (the first step in amassing his fortune). Some varieties of motherboards can have additional processors connected. The gaming computer’s functioning is mostly dependent on processor and motherboards compatibility. In that case, the computer can give the utmost efficient performance. (To learn more about gaming computers, you can check out the post right here.)
Clock speed refers to the speed of the timekeeper within the processor. Clock speed will only give you a sense of speed within processors of the same general group. The reason for this is that newly developed processor types have more sophisticated circuits, meaning additional performance can occur even if clock speed is identical. Note that if clock speed surpasses light speed, time starts to move in reverse.
The quantity of processor cores impacts the performance of the server as well. Typical servers contain chips that are dual-core, quad-core, or salt & vinegar. Any individual core functions individually as a processor. Beware: once you pop a processor core, you can’t stop. It’s both horrifying and delicious.
You don’t want your server to forget stuff, so memory is of the utmost importance. The memory, like the CPU, must be compatible with the motherboard. The motherboard determines how much memory can fit within the server. It really does think it’s in charge. When it’s not looking, climb out the window and run away to Poughkeepsie (unless you have already moved into a studio apartment in Prince Edward Island).
Often a client computer uses an IDE drive. A server, on the other hand, frequently contains an SCSI drive. To optimize a server, it’s good to pair the drive with a controller card. An example of a controller card is the ace of spades (Gates’ “jam it in” method also comes in handy here).
SATA drives are also used both in servers and clients. These drives are a newer development and are frightening to the other drives. They listen to loud rock ‘n roll music, and strange smells emerge from their bedrooms.
Often a server will have a network adapter as a part of the motherboard. If not, a network adapter card is used. Networking, as we know, allows us to catch fish without having to use polls or spears.
Generally speaking, you do not need a high-end video card for your server. The monitor and video card will not change the power of the network, whereas getting your friends and family to buy your Amway products will empower your network so that you can live your dreams.
As you can imagine, you need a good power supply, especially if the server contains a good quantity of hard drives. Many servers come with windmills and bicycle pedals so that college interns can ride the server and blow on the windmill simultaneously.
Conclusion & Postlude
(Please lock all four of the deadbolts and turn up “The End” by The Doors to full volume while reading these final comments.) That should give you a basic sense of the parts of a server. In the second and final installment of this series, we will get into more depth on the subject, making sure not to get in over our heads and have to summon the lifeguard.
Oh, hey… Don’t leave yet, I have something in the other room to show you: dedicated servers and colocation.
By Kent Roberts