Components

Server Components - Upgrade your systems with refurbished hardware and benefit from an extended usage lifespan

A powerful IT infrastructure consisting of servers, network technology, and client PCs or laptops is an essential component in businesses of all sizes. Servers provide computing power and storage capacity to all users. Storage and processing of large data volumes, as well as the execution of applications and services, are central tasks of these systems, where performance is primarily dependent on the configuration and installed components. With advancements in the software sector and growing business areas with an increasing number of users and data, server systems frequently face increasing demands.

If the computing power or storage capacity of your servers is no longer sufficient for changing conditions, this does not necessarily require purchasing a newer server generation. Upgrading with more powerful components can often enhance performance without the need to replace entire servers. However, compatibility of the respective system with the planned upgrade must be considered. The technical documentation of the devices provides detailed information on supported standards and generations of various components.

• Processors - Intel Xeon or AMD EPYC

The processor is responsible for the computing power of a server. Powerful systems often support the use of multiple CPUs (Central Processing Unit) and are therefore capable of processing a large number of tasks simultaneously. The leading manufacturers of server processors are Intel and AMD. Both companies offer models in different performance classes, which differ, among other things, in the number of computing cores and clock frequency. Intel Xeon processors are known for their excellent scalability, while AMD EPYC CPUs combine the advantages of high performance and energy efficiency. When upgrading processors on existing servers, it is particularly important to pay attention to which sockets the systems feature. Both processor manufacturers support their own sockets, which are not compatible with CPUs from the competitor. In addition, the models of the different generations often differ in the required processor socket. Taking compatibility into account, upgrading here can often achieve the greatest increase in performance, since servers are usually not delivered with the strongest models of the supported processor generation. The high availability and price decline of used CPUs enable a cost-effective upgrade, although depending on the processor model, an adaptation of the cooling solution may also be required.

• Cooling - CPU heatsink and chassis fans

Depending on the performance class of the installed components, servers and workstations require appropriate cooling to ensure consistently high performance. Almost all electrical components generate heat that must be reliably dissipated from the case. In particular, processors can reach high temperatures, which are transferred to CPU-heatsinks and released into the surrounding air. Depending on the performance class of the installed CPU, server manufacturers use coolers and fans in standard or high-performance designs. The main differences are the number and size of the heat pipes and fins that transfer the processor's heat to the heatsink. Chassis fans vary in both their speed and size depending on the model to constantly transport the required amount of air out of the case. The airflow is often directed towards the heatsink with air guide elements to ensure the most effective cooling possible. Therefore, when upgrading existing systems, it is important to adjust the cooling solution to the increased heat development following a performance increase.

• Main Memory - RAM modules

The performance of servers is strongly affected by the size of the installed memory in many areas. This is a volatile memory for data of running processes that systems need to access quickly for high operational speed. Therefore, RAM (Random Access Memory) is a crucial component in the field of virtualization or for operating large databases. When servers are used for web applications and need to serve requests from many users, a large memory expansion can also contribute to faster processing and reduced waiting times. Depending on the model, servers usually feature a large amount of memory slots on the motherboard or on riser boards, which can be equipped with RAM modules to expand the memory. With increasing demands of memory-intensive applications, the number of memory modules can be increased, or the memory can be upgraded by replacing the modules with higher-capacity models. In addition to the technical attributes of the memory, such as whether it is unbuffered or registered RAM, the supported memory generation of the server is essential. Memory is available in different generations, which are differentiated by designations such as DDR3, DDR4, and DDR5 (Double Data Rate) and are not compatible with each other. In addition, the RAM modules vary within the generations in their memory capacity and speed. Therefore, when upgrading existing systems, it is essential to check the manufacturer's technical documentation to determine which memory modules can be installed. If, for example, notebooks, compact client PCs or even small NAS systems need to be upgraded, SO-DIMM memory modules often have to be used for expansion. The chassis usually do not offer enough space for the DIMM modules used in servers and workstations.

• Hard Disks - Drives, RAID controllers and accessories

When the storage capacity of a server is no longer sufficient due to increasing amounts of data, it can be expanded by installing additional hard disks. The possible upgrade options depend heavily on the circumstances in the respective system. One deciding factor is the supported form factor. Hard drives are available in both 3.5" LFF (Large Form Factor) and 2.5" SFF (Small Form Factor) formats. The server's drive bays determine which drive format can be used. To be able to quickly install and exchange hard drives, they are inserted into systems with specific carriers depending on the form factor and server model. In addition, optional drive cages often allow the installation of further bays in the server chassis. Depending on the application requirements, storage space can be expanded with HDD (Hard Disk Drive) or SSD (Solid State Drive) models. HDD drives are generally more affordable and offer high storage capacity. SSD drives, on the other hand, are usually more expensive to purchase, but enable significantly higher access speeds and are less susceptible to external influences such as shocks and vibrations.

The connection of the drives can be done via various standards such as SATA, SAS or NVMe depending on the hard drive model, whereby both the installed backplane of the server and the installed storage controller must support the required transmission protocol. The RAID controller is used to manage all installed drives and allows hard drives to be combined into an array. This can increase access speeds, or reduce the risk of data loss if individual hard drives fail by mirroring the data on multiple drives.

SATA drives offer additional benefits besides the lower price. Almost all servers already support the standard in the basic configuration without optional upgrades, and for providing simple services and as a data storage for environments with a small user group, the performance is usually sufficient. SAS hard drives, on the other hand, are suitable for applications with higher requirements due to their higher transfer speeds and advanced management capabilities. To achieve even better performance, NVMe SSDs (Non-Volatile Memory Express) can be used, which are connected via the server's PCI Express (PCIe) interface. With extremely high read and write access speeds, these data storage devices are suitable for applications with the highest demands.

• Graphic Cards - Accelerators for servers and workstations

In addition to upgrading the processor and memory, upgrading the graphic card can often also result in a significant performance boost. In general, the tasks of accelerators in different systems are specialised. Workstations, which are used for tasks such as 3D modeling in product development or media editing, are typically equipped with graphics accelerators. These graphic cards offer high performance in displaying and processing visual content at high resolutions and support the connection of multiple monitors. Cooling of the graphics memory (VRAM) and graphics processing unit (GPU) is usually active, so these graphic cards featuring their own fans to dissipate the generated heat. In servers, on the other hand, computing accelerators are mainly used. These graphic cards execute completely different tasks and differ significantly in their design and technical implementation. The development and application of artificial intelligence (AI) is becoming increasingly important, so compute accelerators specialised for this application area provide high computing power. In machine learning (ML) and the analysis of large and complex datasets, such as in science and medical research, server systems are often equipped with multiple high-performance accelerators that can perform this type of computation faster and more efficiently than processors. However, output of the image signals is not done through the accelerator cards, so they typically do not feature video outputs for connecting monitors. Cooling is mostly achieved by utilising the existing airflow in the server case, so the graphic cards being supplied with air from the chassis fans using air guides to transport the heat. NVIDIA and AMD dominate the market for graphics and compute accelerators. Both manufacturers offer an extensive range of models for almost all applications in different performance classes. To select the best possible solution for an upgrade, the technical documentation of the software being used typically provides guidance through the specified system requirements.

• Network Cards - Connectivity for any network environment

For easy integration of servers into the existing IT infrastructure, systems usually feature integrated network adapters that can provide sufficient performance for many use cases. If high transmission speeds are required, onboard solutions quickly reach their limits as they often only operate with a data rate of 1 Gbit/s. Additional network cards can be used to expand the connectivity of servers and adapt them to the application's requirements. In addition to the widely used Ethernet standard with 10 Gbit/s, dedicated network adapters typically also support the Fibre Channel protocol with significantly higher transmission speeds through the installation of transceivers. The interface converters make it possible to process both electrical and optical signals and therefore offer flexible adaptation options for almost any network environment. The network card is installed mostly in the PCIe Riser-Cage (PCI-Express) of the server or in manufacturer-specific expansion slots. For particularly compact systems such as blade servers, there are special variants available with mezzanine cards that can be installed particularly space-saving in the case. Depending on the performance of existing switches and other network equipment, network cards can offer perfect connectivity for servers in a wide range of network environments.

• Motherboards - The platform of the server

The base of every server is the motherboard. Here, all components are installed and connected. Manufacturers such as HPE, DELL, Lenovo, and Fujitsu design their mainboards in form, size, and mounting points that space in the 19" rack or tower of the respective server model can be used optimally. Depending on the generation, the motherboards differ, for example, in their CPU sockets and chipsets, which decisively affects the selection of compatible components such as processors and memory. When motherboards are needed as spare parts for the fastest possible replacement after a defect, it is therefore essential to pay attention to the device generation. While a board may fit into a server chassis of the previous generation, it will not be able to accommodate older components due to new slot standards. Furthermore, special care must be taken when installing the mainboard, as the fragile contacts and electrical components can be easily damaged if handled improperly. However, replacing a defective board can be a cost-efficient investment compared to purchasing a new server if the system's performance still meets all application requirements. Used and professionally tested motherboards are available at affordable prices and allow the continued use of an older server after replacement, which both saves budget and is environmentally friendly.

• Power Supply Units - Reliability through redundancy

A reliable continuous operation of servers for the provision of services, applications, and data is one of the most important factors. The basis for this is a stable power supply. Depending on the model and configuration, servers are equipped with multiple redundant power supply units (PSU). On the one hand, this distributes the load across all available power supplies, which can reduce the wear and tear of the individual units and therefore increase their lifespan. On the other hand, in the event of a sudden power supply failure, the remaining units guarantee the power supply of the system and ensure availability. To increase the reliability in the event of disturbances in the power supply network, power supplies can also be coupled with an UPS (uninterruptible power supply) to safeguard against power outages. In addition to a high efficiency, which is becoming increasingly important due to rising energy prices, appropriate dimensioning is crucial when selecting power supplies. Depending on the performance of the installed components, the power consumption of servers varies significantly, so it is important for stable continuous operation to pay attention to sufficient output power of the installed power supply units. The replacement of a defective power supply is usually relatively easy, as they often support hot-swap and can be replaced during operation.

• 19" Rack chassis - Barebones for individual configurations

Unlike complete systems offered by server manufacturers, a barebone server is essentially an incomplete system. 19" rack chassis usually only contain the motherboard and power supplies. Components such as processors, memory, storage controllers, as well as hard drives and network cards must be installed separately. This solution is often ideal when existing components are to be reused. For example, with a barebone server from SuperMicro, cost-effective systems can be realised to meet individual requirements.

• 19” Rack rails and other accessories

Servers are typically mounted in server cabinets or racks using 19" rack rails. This secures the systems in place, but also allows them to be easily removed from the rack for maintenance purposes. For this, the compatible rack rails from the server manufacturers are normally used. However, flexible and adjustable heavy load rails can often be attached as a versatile alternative. The requirements for an IT infrastructure vary depending on the size and industry of a company. To meet these requirements, server systems can be specifically adapted to almost any environment through optional accessories and suitable spare parts. Optical drives, for example, facilitate on-site administration, and additional risers enable the integration of additional PCIe cards. If the internal connections of a server are insufficient for specific configurations, they can often still be realised through the use of different cables and adapters.

Refurbished server components available at low prices and delivered quickly - ServerShop24

To adapt your systems to increasing demands, used and carefully tested server components are mostly an ideal alternative to purchasing a new and usually expensive server generation. By upgrading with more powerful hardware, a performance increase can often be achieved, which can lead to a cost-effective extension of the lifespan of existing servers. This benefits not only your budget, but also the environment. Because reused devices do not have to be newly produced and contribute to a sustainable usage of raw material resources. In our online shop, you will find a large selection of used components and accessories at affordable prices, which you can receive with short delivery times through fast shipping from our extensive stock.
Since 2010, we have been your reliable partner for professionally refurbished used servers, storage systems, and network equipment. In addition to pre-configured and ready-to-ship devices from well-known manufacturers such as HPE, DELL, Lenovo, and Fujitsu, we also offer compatible upgrades and individual solutions. If you require any further assistance regarding products and configuration options, our friendly and experienced support team can help you by phone, email, live chat, and through our ticket system. Contact us - we will take care of your request!