Refurbished Servers

Many organisations know they should be doing more to make their IT estates more sustainable by extending product life, cascading assets and buying refurbished equipment. However, if you are used to buying new equipment every three years, it is difficult to know how to start.   

Understanding where refurbished product can fit into your IT estate as well as how to cascade assets, replace and upgrade components, allows you to maximise your overall budget and reduce the environmental impact of your IT infrastructure. All it takes is a little understanding of the system, the environment and overlying software. Whilst we cannot give all the answers to individual situations in this blog, we can signpost the first steps towards this.  

Step 1: Think about your workload 

Whilst some workloads will need the latest generation processors, unless you are running high performance compute or Artificial Intelligence, you can often make do with older equipment as long as it is configured correctly and still supported by the overlying software.  

It is worth noting that an immediate past generation of server can be upgraded to outperform the latest basic generation of server (ref IEEE Moore’s Law paper). This is achieved by upgrading the grade of the CPU and increasing the Random Access Memory, which allows the CPU to run faster. As computers operate similar components to servers, this configuration message holds true for them too. 

Step 2: Consider the effect of software and firmware 

The picture becomes more complicated when you start adding in software upgrades on certain types of equipment. Later generations of Windows, for example, do not run on older machines. Therefore, software upgrades can dictate the generation of hardware you are using.  

There is also an issue around security patches, firmware upgrades and software support. IT managers running mission critical applications with high exposure to threat vectors will often only be comfortable with new equipment because of the wrap around support packages they offer. 

Step 3: Consider the destination for the IT hardware 

For Data Centre equipment, it is much easier to install refurbished equipment in back up environments, failover systems or test and development environments. These are less vulnerable to cyber attack and also not mission critical. If they fail, engineers can order spares and upgrades from the secondary market and have the work carried out by third party maintenance contractors or their own staff.  

In production and primary environments, the same principle applies. If the user in question runs mission critical operations, holds very sensitive information or accesses proprietary systems, IT managers will have to ensure that firewalls and security measures protect the core system rather than the devices. They will also want to allocate devices based on use cases - older machines for lower-level users, newer machines for users pushing the envelope.  

If you’re unsure whether refurbished can, or is recommended to, run your chosen applications, our team can offer free advice and recommendations based on their many years of experience within the industry. 

Step 4: Consider time, expertise and budget  

For many of our customers, budgetary constraints make the choice of refurbished equipment very attractive. Buying refurbished equipment supported by a three-year warranty on the hardware and third-party maintenance contracts often has to be supported by the skills and experience of the IT manager. It requires working with the supplier to correctly size IT estate, designing systems and maintaining the appropriate software support. Whilst this can be more time intensive than the wrap around support of new equipment, the cost savings are significant.  

Refurbished and remanufactured IT is different from refurbished and remanufactured mechanical equipment. Its lack of moving parts means that it either works or does not work. Peer reviewed research in the IEEE Transactions on Sustainable Computing demonstrates that refurbished servers have equivalent performance to new on a like for like basis.  

There are caveats to this:  

  1. You want to source from a clean and efficient refurbishment facility. Research shows that dust on servers, for example, can reduce their effectiveness by up to 10%. Good indicators of high quality providers are certifications such as ISO 9001 and other industry certifications.
  2. Think about how you configure machines to the user workload. Increasing RAM in a server, for example, can increase the efficiency and performance of older machines. However, this RAM needs to be inserted into the correct dimm slots in a server for this to work most efficiently. More than that, overprovisioning of RAM will create less performance for more energy draw.  
  3. Look for thorough in-coming and out-going testing from your refurbished supplier. The majority of processors fail in the first few months. In theory, this means that those that make it to a second life have a higher probability of being robust. However, this needs to be carefully checked before they are deployed, and the supplier needs to guarantee this with a manufacturer comparable warranty for peace of mind.

Step 5: Assess upgrade opportunities 

In amongst all this, it’s worth noting that percentage increases in refurbished equipment do not need to only take place at machine level. You can also extend the life of your assets for lower-level users by replacing failed components, adding more capacity in one area (e.g. memory) or upgrading technology (e.g. replacing hard disk drives with solid state drives).

With data sanitisation software and processes well developed, many of these spares and upgrades will be available as refurbished products, perform as well as new and cause no conflict in the core system.  

Step 6: Understand the metrics of success 

Most organisations have measurable sustainability targets in the form of carbon footprint analysis, waste reduction, resource consumption and circular economy indicators. The importance of each metric will depend on the size of the business and the type of environmental impact it needs to control. For organisations that run a lot of IT workloads, energy efficiency and scope 2 carbon emissions will be a large area of focus.

For those that have a high refresh rate on IT equipment, embodied carbon and waste will be of concern. Being able to balance one over the other, ensuring that IT estates are run with the minimum of energy waste during use phase and material waste at purchase and disposition will be very important in all cases.  

At Techbuyer, we make it our business to engage in cutting edge research on energy efficiency for IT estates and information about carbon savings through the purchase of refurbished equipment and sale for reuse.  With products like our Carbon Savings Calculator and Interact DC, our team are on hand to guide you through the choice and to offer advice where needed.

To contact our team head to our refurbished servers page here for more information.