A recent assessment found that most of our physical hosts are end of life. The issue was raised before but hasn't been addressed, and I'm now looking at a potentially large expense.
I'm considering moving the main data center to Azure. How would you weigh that against replacing the hardware, especially from a security standpoint?
Separate the hardware lifecycle question from the software one. Check whether the operating systems, hypervisor, firmware and remote management interfaces still receive security updates. Old hardware that runs fine can still leave you with an unsupported management interface or firmware vulnerability.
Also check what happens when a component fails. Can you get parts and maintenance quickly enough for the workloads on that host? A server doesn't become unusable on its end-of-life date, but repairability and support matter when uptime is critical.
Don't pick Azure just because the servers are old. Start with an inventory of the hosts, hypervisor and applications, then measure actual CPU and memory usage. Check which applications can move and whether you already have an Azure environment and landing zone ready.
Use that information to price a refresh against a cloud move. Include the work of migration and ongoing management, not just the initial hardware quote or the price of virtual machines. A straight lift and shift can cost more than expected, although some workloads may make financial sense in the cloud.
8w!turnipcase_519:
That gives me a useful starting point. The utilization data should help with both options.
If you need breathing room, check whether extended support or third-party hardware maintenance is available for these models. Get clear terms on parts and response times, then use that as a temporary option while you build a replacement plan. It won't necessarily solve the firmware support problem.
Give leadership a comparison they can act on: replacement hardware, cloud, and extended maintenance if it's available. Show the upfront expense and recurring costs separately, along with the risks of waiting.
Cloud shifts more of the spend into ongoing operating costs. That can be useful for budgeting, but it doesn't automatically make the total lower. On-premises equipment with a recurring payment model may be another option worth pricing.
Answers are reviewed before publication.
2x!clampcaddy_731:
Regulatory or contractual requirements may also affect how long unsupported equipment can stay in service.