(In this SmartBits, Zulfikar Deen outlines “System deployment architecture and testing“. The video is at the end of this blog)

it’s extremely important to understand system deployment architecture. Let’s say we are delivering the same system and as an organization we decide to deploy it in the cloud. It requires a completely different way of testing and we need to show whatever was appropriate works correctly.

If we decide to do a hybrid for various reasons, the solution remains the same, but as an internal decision-making, we may have to make a decision based on compliance. In India, we could take the data into the cloud without much of an issue, especially when the data centres are hosted here, but the same may not be the case in other countries. 
The same solution deployed differently in a different country implies testing has to be different. If it is a hybrid we are trying to work with the two different system fidelities, where part of the system sits on-premise, some of them on cloud. We need to be sure that the data flows through properly and it is secured.

We may not worry about system security between two systems if the entire system is deployed at one place, whereas if it goes through public and private clouds then testing has to be slightly different. If it is completely on-premise then deployment and  testing differs.

If we decide to use a part of it even in the cloud and using only infrastructure as a service the testing would be different. Again if we use a part of a service, say using a data factory from Azure, the testing has to be different because we are using a service provided by the cloud in a different way. We need to make sure it works, if we decide to use advanced services. If we were to use it as part of a whole system, it can’t be tested on-premise. Definitely we need to understand the deployment architecture, how we are planning to deploy and the testing therefore has to be appropriately done for that.