We provide a wide variety of servers, built to your requirements by Scan 3XS Systems. This guide takes you through all the major components of a server build and highlights areas to consider. It also covers server accessories and some wider infrastructure points too, as these may all place a part in the performance and suitability of your server.
What makes 3XS Servers Special?
Our in-house team of experts, 3XS Systems, has been building servers for more than 20 years. We know which components are the fastest and most reliable, and how to fine-tune and combine them for the best application performance. Here's what you need to know.
NVIDIA Certified
Scan is an NVIDIA Elite Partner, awarded for our GPU-accelerated systems expertise.
3 Years Warranty
3XS Systems include a three-year warranty offering complete peace of mind.
Our experienced technical support engineers are available seven days a week.
handshake
Trusted By You
Our servers are used by a host of leading organisations.
Server Form Factors
Tower Servers
Tower servers are designed for a small office where there may not be a dedicated rack cabinet or server room.
Rack Servers
Rack servers fit inside 19in rack cabinets, located alongside other servers and network equipment.
Multi Node Servers
Although these are rack mounted chassis too, they differ as the chassis will contain multiple dense servers designed for very specific compute tasks.
Servers differ from desktop PC and workstations in that they are designed to operate 24/7. This ‘always on’ capability is delivered by server-grade or enterprise-class components that are designed to cope with the continuous demands made of servers. This higher-grade specification of server components also delivers better residency and reliability, so it is never wise to build a server using consumer components.
Scan 3XS Systems servers are built using server-grade components from the same market leading component manufacturers used by multinational brands such as Cisco, Dell, Fujitsu and HPE, however 3XS offers greater customisation during the configuration process.
Server Workloads
Before examining each of the hardware, software, networking and infrastructure options in detail, it's important to think about the types of workload your server(s) will be tasked with. This is critical, as workloads have widely different configuration requirements. Below you'll find high level guidance for popular server workloads - just bear in mind that the specific configuration will depend on a number of factors beyond budget, such as the number of users and client devices, and the size of your datasets.
WORKLOADS
FILE STORAGE
VIRTUALISATION
DATABASE
VISUALISATION
AI / COMPUTE
CPU(s)
Low to medium number of cores
High number of cores
High-frequency
Medium to high number of cores (minimum 6 cores per GPU)
Medium to high number of cores (minimum 6 cores per GPU)
System Memory
Low to medium capacity
High capacity
High capacity
High capacity (minimum 2x GPU VRAM)
High capacity (minimum 2x GPU VRAM)
GPU Accelerator(s)
No requirement
Application dependent
No requirement
GPUs are essential for visualisation
GPUs are essential for these workloads
Storage
Lots of storage bays (configured for performance, capacity and redundancy as required)
High performance SSDs
High-endurance SSDs
SSDs for OS - data will be stored in a dedicated storage appliance on the network
SSDs for OS - data will be stored in a dedicated storage appliance on the network
Networking
Two high bandwidth NICs for redundancy (minimum 10 gigabits per second)
Multiple high bandwidth NICs for redundancy (1 port per VM, minimum 25 gigabits per second)
Multiple high bandwidth Smart NICs for redundancy (minimum 25 gigabits per second)
Multiple high bandwidth Smart NICs for redundancy (minimum 25 gigabits per second)
Multiple high bandwidth Smart NICs / SuperNICs / DPUs to keep GPUs fed with data (minimum 100 gigabits per second)
Operating System
Windows / Linux
Windows / Linux
Windows / Linux
Windows / Linux
Linux
Now let's take a look at each of these configuration options in more detail.
Server CPUs
The CPU is one of the most critical components to consider when building a new server. The choice of CPU or CPUs will govern the server’s performance when it comes to compute tasks, multitasking or virtual machine (VM) capability - derived by the number of cores, system memory and amount of cache. In high performance servers aimed at HPC or AI workloads, the CPU should be considered alongside accelerators and network cards that act to offload tasks away from the CPU(s). There are two main server CPU families to choose from - AMD EPYC and Intel Xeon.
AMD EPYC CPUs
The AMD EPYC series of processors set the new standard for the modern datacentre. Featuring leadership architecture, performance, and security, AMD EPYC processors help customers turbocharge application performance, transform datacentre operations and help secure critical data. You can learn about the specific models in our AMD EPYC CPU BUYERS GUIDE.
Intel Xeon CPUs
The Intel Xeon 6 processor family introduces an innovative modular architecture that allows datacentre architects to configure and deploy servers that are purpose-built for your unique needs and workloads across private, public and hybrid clouds. Intel Xeon 6 processors have either Performance-cores (P-cores) for maximum performance or Efficient-cores (E-cores) for maximum efficiency in power- and space-constrained environments. You can learn about the specific models in our INTEL XEON CPU BUYERS GUIDE.
Server Memory
When it comes to servers, the best approach is to use fewer, larger capacity DIMMs, as they will use less power and you still have room for expansion in the future. There are a number of DIMM types available - UDIMM, RDIMM, LRDIMM and MRDIMM - the best choice will be dependent on your workloads. Click the tabs below to explore the differences between them.
UDIMM
UDIMM or Unbuffered DIMMs address and control signals directly reach the DRAM chips without going through a buffer and without any timing adjustment. UDIMMs are lower in capacity and frequency and cannot maximise server performance because they only operate in unbuffered mode and thus cannot support the maximum RAM capacity.
RDIMM
RDIMM or Registered DIMMs deploy a register between the CPU and the DRAM chip for data transmission, which reduces the distance of parallel transmission and improves efficiency. RDIMMs are easier to increase in capacity and frequency than UDIMMs and they support buffered and high-performance registered mode, making it more stable than UDIMM. This gives it the highest-capacity server RAM performance and a wide range of applications in the server RAM market.
LRDIMM
LRDIMM or Load Reduced DIMMs do not use registers but are simply buffered. The buffering reduces the power load on the underlying motherboard but has little effect on memory performance. The direct benefit is to reduce the load and power consumption of the memory bus but also provides the maximum supported capacity of server RAM.
MRDIMM
MRDIMM or Multi-Capacity Rank DIMMs improve performance and efficiency by leveraging multiple ranks within a single module. An MRCD (multiplexed registering clock driver) addresses two ranks of DRAMs on alternate clock cycles, as well as MDBs (multiplexed data buffers) to direct the data stream to and from the correct DRAM devices.
All of the above types of DIMM are available with a feature known as Error Correction Code or ECC. ECC memory allows the system to recreate the correct data in real time in case of an error, using a form of parity, which is a method of using a single bit of data to detect errors in larger groups of data. ECC memory looks different to non-ECC memory as it has an extra chip that performs the ECC calculations. ECC memory must be matched otherwise the error correcting capability will be lost. ECC memory is not always registered or buffered, however, all registered memory is ECC memory.
Server GPU Accelerators
Thanks to their highly parallel architecture, GPUs are an order of magnitude faster than CPUs for visual and compute workloads. The leading server GPU accelerator cards are from the NVIDIA Datacentre GPU range, and feature a varying number of CUDA (general purpose), RT (ray tracing) and Tensor (AI) cores for different workloads. You can learn more by reading our NVIDIA DATACENTRE GPU BUYERS GUIDE.
Aside from adding one or more GPU cards to your server, there are also NVIDIA-certified architectures, such as the NVIDIA RTX PRO SERVER range. Powered by NVIDIA RTX PRO GPUs, they provide best-in-class performance across a wide range of workloads including visualisation and rendering, scientific computing and HPC, and small-scale AI. Designed for datacentre environments, 3XS RTX PRO Servers are fine-tuned by our hardware engineers and workload specialists for maximum performance and reliability.
Featuring higher spec than your average GPU-accelerated server, they offer up to eight NVIDIA RTX PRO Blackwell Server GPUs with up to 768GB VRAM, NVIDIA ConnectX Smart NICs, and advanced functions including multi-instance GPU (MIG), where each GPU can be shared with up to four users, fully isolated at the hardware level.
Internal Storage
The type and capacity of any internal storage required in a server will very much depend on what the intended use of the server. There are two types of drives - HDDs and SSDs, the former features spinning disk platters on which the data is stored, whereas the latter has no moving parts and stores data on NAND flash memory. Each type has its advantages and disadvantages depending on the cost effectiveness or performance required, as summarised in the table below.
Solid State Drives
SSDs have no moving parts and store data on NAND flash memory, providing much higher performance and lower access times for regularly accessed data.
Hard Disk Drives
HDDs feature spinning disk platters on which data is stored, making them cost-effective for less used or archived data.
HDD
SSD
Comparison
Performance
Hundreds of MB/sec
Thousands of MB/sec
SSDs are much faster
Access Times
5-8ms
0.1ms
SSDs have almost no latency
Reliability
2-5% failure rate
0.5% failure rate
SSDs much more reliable
Resilience
Susceptible to vibrations
No moving parts
SSDs are much more resilient
Energy Use
6-15W
2-5W
SSDs much more energy efficient
Form factors
2.5” / 3.5”
2.5” / 3.5” / U.2 / U.3 / M.2 / PCIe card
SSDs offer more flexibility including compact M.2 drives for the OS
Capacity
Up to 26TB
Up to 30TB
Similar in maximum capacities
Use case
Less used or archived data
Regularly accessed data
Use both for tiered storage environments
Cost
£-££
££ - ££££
SSD more expensive, especially at high capacities
Whichever type of internal storage drive you choose, you should ensure they are datacentre- or enterprise-class drives designed for 24/7 use. You can learn more about both types of drives and RAID data security by reading our DATACENTRE SSD AND HDD BUYERS GUIDE.
Server Network Cards
Network connectivity in a server can range from a single 1Gb/s (Gigabits per second) network interface card (NIC) to multiple high-throughput, low-latency 800Gbps Ethernet and InfiniBand Smart NICs, SuperNICs and DPUs, that offload many tasks from the CPU. Which type of network card to choose depends on what your server will be used for, as summarised in the tabs below.
NICs
Although only a small component in an overall system build, the NIC can contribute to a huge uplift in performance. Basic server NICs start with throughput speeds of 1Gb/s through a single port, scaling to 40Gb/s at the top end, featuring two or four ports. All processing of data is performed either by the CPU(s) and GPU(s) installed in the server, and thus introduces latency as data is transferred around the server. To reduce this latency and increase throughput speeds a Smart NIC is required.
Smart NICs
A Smart NIC performs all the tasks of a regular NIC but in order to deliver higher throughput, the Smart NIC offloads tasks from the CPU, system memory and OS. This offloading is referred to as Remote Memory Direct Access (RDMA) for InfiniBand cards and RDMA over converged Ethernet (RoCE) for Ethernet cards. Additionally, SmartNICs provide innovative hardware engines that offload and accelerate security with in-line encryption/decryption
SuperNICs
While a DPU and a SuperNIC share features and capabilities, SuperNICs are designed for network-intensive computing, providing RoCE network connectivity between GPU servers, optimising peak AI workload efficiency. As the sole purpose of the SuperNIC is to accelerate networking, it consumes less power than a DPU, which requires significant resources to offload applications from the CPU(s).
DPUs
DPUs offer an uplift over Smart NICs and Super NICs by offloading, accelerating and isolating a broad range of advanced, storage, networking and security services. They provide a secure and accelerated infrastructure for HPC or AI workloads in largely containerised environments. Featuring speeds up to 800Gb/s, they also include NVMe over Fabric (NVMe-oF), GPUDirect Storage, encryption, elastic storage, data integrity, decompression and deduplication for reduce storage management overhead within the server.
You can learn more about networking protocols, types of card and other networking considerations by reading our NETWORK CARDS BUYERS GUIDE.
Operating Systems
The operating system governs how the server is managed and how applications are deployed. Servers most commonly run either Microsoft Windows Server or a Linux distribution.
Windows Server 2025
Windows Server 2025 is the latest operating system for servers from Microsoft. To cover small office servers right up to datacentre deployments, it is available in three versions - Windows Server Essentials, Windows Server Standard and Windows Server Datacenter.
There are many distributions of Linux available, however the most frequently used is Ubuntu. It is based on Debian and is utilised by many organisations as it is renowned for its dependability, smooth operation, and efficiency. Alternatively, Red Hat offers a variety of open source software that DevOps engineers and businesses prefer.
External Storage
As previously mentioned it is not uncommon for storage capacity to expand outside the confines of the server chassis. This may be in addition to the internal storage the server has to offer, or if significant fast flash storage is required, then a network storage system may be the best approach. 3XS servers can be configured as dedicated storage devices or to compliment a wide variety of storage systems from a number of leading brands.
NAS
Network Attached Storage (NAS) systems are the simplest way to add extra storage capacity. Usually connected over Ethernet, we offer various NAS options from both QNAP and Synology in 4- to 16-drive bay versions, allowing you to scale efficiently and cost-effectively. You can learn more about NAS solutions in our dedicated NAS BUYERS GUIDE.
Tiered Storage
Tiered storage is designed to act as fast access, regular storage and archive in a single solution. Different drives types are used to achieve the parameters needed - NVMe SSDs for hot data, SATA SSDs for the bulk storage, and cost-effective HDDs for the archive capacity. Inbuilt software handles the data management features like de-duplication and compression. You can learn more by reading our EXTERNAL STORAGE BUYERS GUIDE.
AI Optimised Storage
PEAK:AIO has developed a software platform from the ground up for AI workloads to complement any AI server. Delivering ultra-low latency and tremendous bandwidth at a price which allows more investment to be made on GPU resource and less on storage - ensuring your project’s funds are better spent and your team is more productive. Learn more about PEAK:AIO storage.
Server Networking and Infrastructure
In addition to the server(s) and storage, any business or organisation will need a wide infrastructure including network switches, UPS and racking to house everything.
Network Switches
Network switches are devices that connect multiple PCs, workstations, servers or storage devices to enable them to communicate within an organisation and share and access the Internet connection to the wider world. Switches are available in either Ethernet for most common uses or InfiniBand for high-end visualisation, HPC and AI workloads. With various throughput capabilities and feature sets, it is essential to choose the right switches and best support your hardware devices across the network. To learn more, read our NETWORK SWITCH BUYERS GUIDE or learn more about high-end networking in our dedicated AI NETWORKING BUYERS GUIDE.
UPS
An Uninterruptible Power Supply (UPS) sits between the mains power source and servers. It ensures they receive a consistent and clean power supply, whilst also protecting them from power surges and failures. A surge could damage components within the server and a failure could interrupt data being saved on the device, resulting in errors. The battery runtime of most UPSs is relatively short - 5 to 15 minutes - but sufficient to bring an auxiliary power source online, or to properly shut down servers and other devices. To learn more about correctly sizing a UPS, extended runtime options and connections to your servers please read our dedicated UPS BUYER GUIDE.
Rack Cabinets
Even in a smaller office a rack cabinet provides a secure environment for your servers, storage, network switches and UPS. They also work to keep cabling simple and clear and to stop unauthorised users from accessing mission critical infrastructure. Cabinets are usually an industry standard 19” wide and can be configured to a range of heights - typically 14U - 47U, and with a variety of sides, shelves and mounts for power distribution units (PDUs).
Professional Services
Although all our 3XS Systems servers come with a comprehensive 3-year warranty, more complex server configurations can be further supported with a number of PROFESSIONAL SERVICES, including installation, network health checks, infrastructure consultancy, and cyber security and data sciences services.
Ready to Buy?
We hope you’ve found our server buyers guide, giving you the additional knowledge to get the most out of our intuitive server configurators. Click below to see our great ranges of servers designed for many popular workloads.
