Operations 9 min read

How to Choose the Right Server: Key Brands and Essential Specs Explained

This guide explains what a server is, lists the most common server brands, and walks through the crucial hardware parameters—bandwidth, CPU, chipset, memory, storage, network cards, redundancy, hot‑swap and scalability—helping you make an informed server purchase decision.

Open Source Linux
Open Source Linux
Open Source Linux
How to Choose the Right Server: Key Brands and Essential Specs Explained

1. Server Brands

Common server brands include Dell, HP, IBM, Huawei, Inspur, ZTE, Tsinghua Tongfang, Fujitsu, Hikvision and many others widely used across various fields.

2. Main Server Selection Parameters

Bandwidth

A 5 Mbps (5 M) bandwidth equals 625 KB/s (5 Mbps ÷ 8). Assuming a typical web page size of 50 KB, the server can handle roughly 12.5 simultaneous users (625 KB ÷ 50 KB). In practice, because requests are staggered, a 5 M bandwidth can support many more users if the pages are small.

CPU

The CPU is the core processing unit of a computer. Performance is determined by clock frequency, cache size, instruction set and architecture. Important metrics include:

Clock frequency (MHz/GHz) : higher values mean faster data processing.

Cache : larger cache improves data‑hit rates and overall speed.

Core count and threads : each core usually has one thread; Intel’s Hyper‑Threading can provide two threads per core, e.g., a dual‑core CPU can present four threads.

Intel server CPUs include Xeon, while desktop CPUs are typically Core i series.

Chipset

In X86 systems, the chipset consists of CPU, northbridge and southbridge (now integrated as CPU + PCH). It defines the system’s bus architecture and compatibility, allowing various CPUs to work with the same motherboard.

Memory

Servers use ECC memory for error correction. Recommended minimum capacities are:

Entry‑level: ≥2 GB

Group‑level: ≥4 GB

Department‑level: ≥8 GB

More memory improves performance for databases, proxies, web services, etc.

Storage

Typical storage interfaces:

SATA : Serial ATA, widely used, currently SATA II.

SCSI : Small Computer System Interface, high bandwidth, low CPU usage, suited for mid‑to‑high‑end servers.

SAS : Serial Attached SCSI, serial version of SCSI, still first‑generation.

SSD : Solid‑State Disk, no moving parts, offers low power consumption, high reliability, and better performance.

For higher reliability, servers often use RAID cards (SATA RAID for entry‑level, SAS RAID for department/enterprise level) to create disk arrays.

Network Card

Servers should have at least one gigabit NIC installed in a 64‑bit PCI slot. Specialized servers (FTP, file, video streaming) may require two gigabit NICs.

Redundancy

Reliability is critical. Redundancy measures include:

Disk redundancy : RAID arrays protect data if a drive fails.

Component redundancy : duplicate NICs, power supplies, fans, etc., ensure continued operation after a hardware fault.

Hot‑swap : ability to replace drives or cards without shutting down the server.

Scalability

Servers provide ample drive bays, memory slots, CPU sockets and expansion slots, allowing growth in storage, memory, processing power and additional peripherals. Power modules can also be expanded when many drives are added.

Other parameters exist but are not covered here.

ScalabilityServer Hardwareredundancyserver selectionBandwidth CalculationCPU specs
Open Source Linux
Written by

Open Source Linux

Focused on sharing Linux/Unix content, covering fundamentals, system development, network programming, automation/operations, cloud computing, and related professional knowledge.

0 followers
Reader feedback

How this landed with the community

Sign in to like

Rate this article

Was this worth your time?

Sign in to rate
Discussion

0 Comments

Thoughtful readers leave field notes, pushback, and hard-won operational detail here.