
Decommissioning-as-a-Service
Identify unused servers and save IT costs
Unused servers can be expensive – identifying them is a challenge
Many companies, regardless of industry, face a similar challenge: the server infrastructure has historically grown and grown. This leads to unnoticed and unused servers and the system landscape is sometimes difficult to keep track of.
Benefits
€ 60 Mio.
Resources
60,000,000,000 € are wasted by companies worldwide per year on cloud services they do not need.
360 kg
CO2 emissions
A single server generates 360 kg CO2 emissions each year.
30 %
unused servers
An estimated 30 percent of servers in data centres are unused.
Our solution for identifying and shutting down unused servers
Decommissioning-as-a-Service is PwC's end-to-end tool that identifies the unused servers and then comprehensively guides the decommissioning process together with our experts. We make sure you get the perfect solution. That's why our services are always tailored to your business.
01Discovery and Assessment Phase
Our team checks the monitoring data and prepares it for the process in a comprehensible and transparent way. Subsequently, the so-called training of artificial intelligence takes place. As a result, the AI creates a candidate list for unused servers that can be decommissioned.
In the Spotlight
Pricing
Time and material
Only pay for the effort involved.
- You pay our advisors according to a fixed hourly rate
- You agree on this hourly rate in advance with our team
- Financial flexibility for you by adjusting the project scope at any time
for an initial analysis
Risk and reward
You only pay if we find unused servers.
After completion of the project, you pay a one-off percentage of the savings negotiated in advance with PwC.
- Pay only for the realised savings through the use of our tool
- Financial advantages through payment only in case of success
for an initial analysis
Fixed project based
- The fixed price fully reflects our performance in your company
- Financial predictability through remuneration without fluctuations
- Choice of three sizes, depending on the number of servers
Depending on the number of servers.
FAQ
How can unused servers be identified?
Unused servers are often characterised by divergent load patterns. Used systems often have seasonal and recurring load patterns. One such recurring pattern is, for example, a non-existent load at the weekend. Unused systems often have a permanently very low load level with few fluctuations.
Is the tool also suitable for tidy and clear server landscapes?
Yes! The tool is suitable for every landscape. In such cases, the risk-and-reward model is recommended. This is because additional costs are only incurred if unused servers are found and decommissioned. Our consultants provide transparent and professional support in the complete optimisation of server performance.
Who shuts down the unused servers?
PwC supports your company and all employees in the orchestration process. The partial decommissioning of your IT is then carried out by your own IT department following detailed preparatory planning.
Can the tool be linked with other monitoring tools?
Data protection plays a prominent role at PwC and this is of course also uncompromisingly guaranteed in connection with Decommissioning-as-a-Service.
Therefore, direct linking with a monitoring tool is not allowed. However, we support a number of data formats of common monitoring tools. The data is uploaded to our tool anonymously and we guarantee that no personal data is stored on PwC servers.
How does PwC ensure that only servers that are actually unused are switched off?
After identifying a potentially unused system, a visual inspection of the system often takes place to substantiate the assumption. If the impression gained by a system administrator is confirmed, we carry out a so-called scream test. This involves a coordinated and temporary shutdown of the system to check whether there are messages from users via central message channels. These include, for example, the IT hotline or other outage services. If there are no messages within a two to four week time window, then we can determine with a very high degree of certainty that the server is indeed unused. This is a common and much tested procedure.





