So you’ve read our previous article in the server buying guide series that compares the pros and cons of on-premise vs. cloud servers and you’ve decided that owning a on-premise server is the right choice for your business. Congratulations – we think that’s a great choice, but now you have some real decisions to make:
- What hardware components are most important to consider when building-out my server?
- How do I choose the right CPU, RAM and storage?
- How should I configure my server so that it serves our business needs now and well into the future?
This article will cover the 3 primary configuration options you’ll want to consider when purchase a server.
While you’ll certainly need to consider additional items like server form-factor, manufacturer and whether to buy new or refurbished – all of which we will be covering in a future article – there are 3 primary configurations options that you will need to be aware of that will directly effect the performance of your server. These server configuration options are as follows:
- CPU / Clock speed
- Random access memory (RAM)
- Storage (HDD & SSD)
Short for central processing unit, the CPU is the brain of the server and will have a direct impact on performance and speed of the applications that you run. When choosing a CPU there 3 main factors to consider:
- Clock speed: Measured in gigahertz (GHz), the clock speed determines how many continuous calculations your CPU can process at any given time. Clock speed, especially in CPU’s with a limited number of cores, is one of the key components that will determine the speed and responsiveness of a server.
- Cores: In the early days of computing when many computers and servers came with only a single core, if the current process or application freezes then the entire computer or server freezes. Today, the ability to have multiple cores on a single CPU means that workload can be spread across various cores and operate independently of one another.
- Threads: Threading is the process by which a physical CPU is broken up into multiple virtual cores allowing it to simultaneously handle multiple processes. In short, two software threads can be running on a single CPU core simultaneously, which increases the computing potential of the server.
So you may be asking yourself, “So what’s more important – higher clock speed or more cores when choosing a CPU for my server”
and here’s our answer: If the most important factor is speed in which your server can load and interact with a given application then you’ll want to focus on having higher clock speeds. Conversely, if the most important factor is your servers ability to run multiple applications at once – albeit potentially running a little slower – then core count is the most important factor when selecting a CPU. For an in-depth look at cores vs. clock speed, here’s a great article
from Winner VPS
Often abbreviated as RAM, random access memory is a critical component of your server’s hardware configuration. The operating system, applications and processes that run on your server need to access (read) information and data and the random access memory (RAM) acts as a temporary repository for reading and writing information that is much quicker for the CPU to access than it would otherwise be trying to access that same information directly from the server’s hard disk.
So you may be asking yourself, “How much RAM do I need?” and the answer here is a little more variable, but here are some questions that can help lead you in the right direction:
- What is the primary use of the server going to be?
- How many users will be accessing the server?
- What operating system and applications will I be running on the server once it’s in operation?
- Is the server going to be single-use (dedicated) or multi-use (shared)
- How do we expect the use of the server to grow over its lifetime?
While it’s true that there’s no such thing as too much RAM when configuring a server, it is an added expensive you’ll have to pay for so we like to start by recommending anywhere from 16GB – 64GB of RAM, but that may change depending on how you plan on using your server and how long you expect it to be in operation.
Storage (HDD & SSD)
You’ll often hear the data storage component of a server simply referred to as it’s hard drive, but technically speaking, a hard (disk) drive is just one of the two storage component options you’ll have for your server:
HDD: Short for hard disk drive, this storage option relies on magnets to spin disk heads along a metal disk to read and write data. Differing from the technology found in newer solid state drives (SSD), HDD’s contain physically moving parts, which makes them more prone to potential failures. That said, the modern HDD’s built for use in servers, which often run 24×7/365, are extremely reliable and have minimal latency – especially when compared to the HDD counterparts that are built for personal computing devices. When choosing an HDD, you’ll want to consider the following:
- Interface: In lay terms, the hard drive interface is how the severs motherboard connects to the hard drive. The two primary types you’ll have to choose from are SATA (Serial Advanced Technology Attachment) and SAS (Serial Attached SCSI). SATA drives are slower, but come with a massive storage capacity of up to 8TB. Conversely, SAS drives are extremely reliable and much faster than SATA drives, but have a much lower maximum storage capacity that’s closer to 1TB.
- Speed: Determines how quickly the disk can read or write data and is based on the rotational speed of the disk. Most SATA drives will be 5,400 or 7,200 RPM, whereas SAS drives can clock up to 15,000 RPM.
- Throughput: Otherwise known as the data transfer rate, throughput measures how quickly the retrieved data is returned to the CPU. For SATA drives the max throughput is 6GB/sec and in SAS drives it’s 12GB/sec.
- Size: There are 2 HDD sizes – 3.5″ and 2.5″. The 3.5″ variety will let you store more data, but that comes at the costs of higher power consumption. Currently, the maximum storage capacity of a 3.5″ is around 8TB, whereas as 2.5″ is around a max of 4TB.
SSD: Short for solid state drive, this newer technology utilizes flash storage, which relies on memory chips to store data – the same type that are used in random access memory (RAM). Unlike hard disk drives, solid state drives contain no moving parts, have a significantly higher read/write speed, are resistant to shock/vibrations and consume less power than their HDD counterpart. Sounds like SSD’s are a no-brainer when selecting a storage option for a server in 2020 then, right? Well, yes and no – there are a couple of drawbacks you’ll want to consider:
- Solid state drives are significantly more expensive than hard disk drives
- Solid state drives can sometimes have issues with long-term data storage, but modern-day firmware has significantly reduced that problem
If you’d like a more detailed comparison on on HDD vs SSD, we recommend reading this article
from the Enterprise Storage Forum.