Last Updated on November 21, 2017 by Ruchir Shastri
You’ll require a dedicated server, either when your shared or VPS hosting account is no longer capable to meet current hosting needs, or you’re planning to launch a brand new project which demands higher server resources. In either cases, you don’t want to compromise with the server performance, right?
Choosing a dedicated server with the best performance is actually difficult, when there are significant options available in the market. It’s indeed an important decision so you just can’t head to any other web host and order a generic dedicated server. So this guide will explain some basic principles how you can choose a dedicated server with the best performance without spending a fortune.
The desktop-grade processors like Intel’s Core i5 or i7 are adequate for many small businesses with just basic requirements. However, to deploy a huge CPU processing, definitely you’ll need business-grade processors like Intel’s Xeon series CPUs.
These processors produce great processing power due to their Hyper-Threading architecture, multiple cores, bigger L3 caches and ability to handle intensive loads, which make them more suitable for server environment.
Also, Xeon series CPUs support ECC (Error-Correcting Code) memory, which can detect and correct most common data-corruption even before it takes place, ensuring the data integrity and stability of the system.
To get the best out of your high performance dedicated server, we’d suggest you to go with the Xeon E7 v6 family – the latest generation of high-end processors with improved performance and efficiency. Though you may still find v5 and v4 Xeon servers available in the market, and they are quite capable to deliver decent performance till date. You can also get the latest generation of low-end Xeon chips like E3 and E5, typically used in entry-level servers.
If you are an AMD fan, then you can go with its business-grade processors launched under
Opteron family, ranging from the entry-level 3000 series to powerful 6000 series.
For the entry-level servers, RAM typically ranges from 8GB to 64GB, which can be further expanded from 128 to 512 GB for high-end servers. These memory modules come in different types and speeds, which are dependent upon the processor and motherboard. For example, 2,400MHz ECC DDR4 memory is the standard for most entry-level Xeon-based servers.
However, RAM speed has far less impact on overall performance of a dedicated server, so it’s okay if you go with lower MHz memory modules. Also, your server must have capability to upgrade the RAM for the future expansion, so you don’t have to migrate to another server.
While configuring the server, you can choose between solid state drives (SSD) or a traditional option of hard disk drives (HDD). HDDs feature a spinning platter and a read/write arm that uses magnetism to store data. SSDs don’t have moving parts and keep information in flash microchips. For these characteristics, SSDs are much faster and more expensive.
For the best performance, we’d suggest mounting all SSDs on the server or go at least with the hybrid option for better performance. With the hybrid option, operating system is installed on SSDs, and standard HDDs are used to store files.
The hybrid option does not deliver the performance like all SSD setup, but it can be considered as a good compromise between performance and price.
RAID (Redundant Array of Independent Disks) is vital for any multi-disk dedicated server as it increases the performance and reliability of data storage. The most common types of RAID are:
- RAID 0: It stripes data across multiple drives to improve the performance, though one of the drives fail all the data will be lost.
- RAID 1: It mirrors/copies the same data on 2 drives, so if one of the drives fail there’s still a complete and usable copy of the data on the other drive.
- RAID 5: It requires at least 3 drives. It stripes the data and adds “parity data” so that any one of the drives fails, it can be replaced with the new drive promptly without losing the data.
- RAID 10: It’s the combination of the striping and mirroring.
- RAID 50: It’s the combination of the striping and parity.
Many entry-level servers come with RAID 0 or 1 with limited options, but it’s better to go at least with the RAID 5 or RAID 10. For the Windows-based dedicated servers, hardware RAID is more common, wherein its advantages are better realized. In contrast, software RAID is more popular in open-source systems, wherein its low-cost makes it an attractive option.
Traffic & Bandwidth
Traffic or “data transfer”, is the volume of data transmitted between your server and the web over a certain period of time (monthly in most cases). It is measured usually in gigabytes (GB) and terabytes (TB). So if you buy a server with 100TB traffic limit, that means there’s a limit of 100TB of data that can be passed between your server and web during the month.
Bandwidth is described as a data transfer rate of a dedicated server, or the amount of data network is capable to transmit over a given period of time. It is measured in Megabits per second (Mbps) and sometimes Gigabits per second (Gbps). Dedicated server bandwidth typically ranges between 100 Mbps and 10Gbps, with 1 Gbps a common bandwidth for high performance dedicated servers.
Some providers offer dedicated server with unmetered traffic, but unmetered does not mean unlimited. In fact unmetered plans often restrict the amount of data you can send-receive because data is transferred at a very low speed.
We hope going through this post you’d have a better idea as how to choose the best dedicated server without compromising on performance. Since, there are plenty of good options available in the market, we recommend to do your own research before purchase. In case, you still have questions or would like to get a professional advice, you can always contact our expert sales team.