What Makes a Server Tick? A Complete Guide to Server Fundamentals
This article provides a comprehensive overview of server fundamentals, covering the definition of a server, its classification by form factor, CPU architecture, and usage, as well as detailed explanations of hardware components like CPU, memory, storage, I/O, and networking devices.
What Is a Server?
A server is a high‑performance computer that provides services to other machines over a network. It features fast CPUs, reliable long‑running operation, strong I/O throughput, and scalability.
Server Classification
By Form Factor
Tower servers – resemble desktop PCs.
Rack (chassis) servers – standardized size, mounted in racks.
Blade servers – high‑density modules inserted into a chassis.
By CPU Architecture
CISC (x86, x86‑64) – AMD, Intel.
RISC – ARM‑based processors (e.g., Huawei Kunpeng, Ampere, AWS Graviton).
Other architectures – MPIS, PowerPC, SPARC, VLIW (IA‑64, AMD Athlon 64).
By Scale and Use
Enterprise vs. entry‑level.
General‑purpose vs. specialized servers.
X86 servers vs. non‑X86 (mainframes, Unix servers).
Server Components
Hardware
CPU, memory, storage, power supply, fans.
I/O adapters (RAID, NIC, HBA, etc.).
Management module (in‑band or out‑of‑band).
Software
Firmware (BIOS/UEFI).
System software (OS, DBMS, runtime libraries).
Application software (general or custom).
CPU Details
Key CPU parameters include:
Clock speed (frequency) – higher values mean faster processing.
Core count – more cores increase parallelism.
Threads – logical execution units per core (e.g., Hyper‑Threading).
Cache hierarchy (L1, L2, L3) – larger caches reduce memory bottlenecks.
Front‑side bus (FSB) speed.
TDP – thermal design power, indicating cooling requirements.
Virtualization extensions (VT‑x, AMD‑V).
Memory Overview
Memory works closely with the CPU and is classified as:
RAM – volatile main memory.
ROM – non‑volatile read‑only storage.
Cache – high‑speed buffer between CPU and RAM.
Memory technologies include SRAM (fast, low‑power) and DRAM (common main memory). DRAM types: SDRAM, RDRAM, EDO‑DRAM, FPM‑DRAM, and DDR generations (DDR, DDR2, DDR3, DDR4, DDR5). Important memory parameters:
Frequency (memory clock).
Capacity (e.g., 4 GB, 8 GB).
CAS latency (CL).
ECC – error‑checking (parity) used in enterprise modules.
Access time (ns) – lower values mean faster access.
Storage Subsystems
Servers use both front‑mounted (often for hyper‑converged storage) and rear‑mounted (local OS/app storage) drives.
HDD – mechanical disks.
SSD – solid‑state drives (higher speed, higher cost).
Hybrid HDD (HHD) – combines magnetic and NAND flash.
Interface protocols: SCSI, FC, SAS, SATA, PCIe, M.2.
Network and Expansion Cards
Typical PCI‑based cards include:
NIC – Ethernet adapters (RJ45, optical SFP/SFP+ modules).
HCA – InfiniBand adapters for high‑performance computing.
HBA – Host Bus Adapters for Fibre Channel or iSCSI storage.
RAID controller – hardware RAID configuration.
Optical modules (SFP, SFP+, QSFP, etc.) vary by speed (1 Gb, 10 Gb, 25 Gb, 40 Gb, 100 Gb) and mode (single‑mode vs. multimode). Fiber connectors include LC, SC, FC, ST.
Illustrations
Signed-in readers can open the original source through BestHub's protected redirect.
This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactand we will review it promptly.
Architects' Tech Alliance
Sharing project experiences, insights into cutting-edge architectures, focusing on cloud computing, microservices, big data, hyper-convergence, storage, data protection, artificial intelligence, industry practices and solutions.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.
