Having Staging and Production environments with different network virtualization

Question

Current Setup

1. We have a staging environment which used NFV (be precious NSX) to configure the network (Software load balancer and software firewall)
2. The Production environment has used the traditional networking techniques (hardware load balancer and hardware firewall)
3. Both environments have the same application server version and database version

When it comes to staging vs production, it's defined as mirroring the actual production environment as closely as possible.

Is above setup a proper staging and production setup? is it possible to ignore the network factor?

Answer 1

The point of staging is for testing. The point of testing is to reduce risk. Any differences between staging and production are areas of risk. The risk may be large or small depending on your situation:

• If it's a brand new production system, you should be able to use it as your staging system before it is opened to the public. So parity is not an issue.

• If it is an existing production system, and you're just pushing software changes, chances are any network issues have already been teased out of it. The risk is small.

• If you are pushing network changes as part of the deployment the risk is high. There should be somewhere you can stage those.

It is certainly "OK" to have a different network setup (it's often unavoidable), but a risk should be recorded using whatever governance process your organization utilizes. QA can evaluate the risk and may be able to recommend additional test cases to be executed during or after your go-live. Executives should get visibility to the risk on any project status reports, and need to decide whether to fund a project to bring the staging environment into parity.