Both on-premise servers and cloud have their own strength. On-premise servers have lower latency, great privacy, and can be made for various kinds of customization. Even though servers have been getting better improvements with lesser and lesser price, sometimes it is not cheap for some occasion.

Cloud providers offer their common modules or services with their level of security and quality which means by using AWS your company already uses the same level of infrastructure architecture as Amazon, and that is also the case for Azure and GCP.

For example, you can rely on AWS high-available services to build an e-commerce website without having to worry about any downtime issue caused by power outage or natural disaster. This also can be achieved by having multiple servers in different places but it will consume more time and cost by doing all of the configuration and setup ourselves while it is already provided with reasonable charges.

Cloud has been selected as the best option for early stage companies or startups. Not only does it reduce the cost to own an on premise server and all the engineering cost to maintain it, cloud services have done a good job abstracting all the small details to provide the best infrastructure for your application.

High Flexibility

For us, the most intriguing aspect of cloud services that we are really interested in is definitely their flexibility. In the old days, if you wanted to create a tech company then you are probably need to order your server first. Engineers need to wait for the server to arrive at your office, plug all the cables, and configure it beside the server. When things go wrong, enginners are required to be physically presence to fix the problems. With cloud, engineers can create a server within minutes and turn it off a couple hours laters just from their desk. This capability seems simple but it brings a huge benefit in many other cases.

Pay on Demand

The second most interesting aspect of cloud is its pay on demand feature which can greatly reduce the cost of running the infrastructure when used properly. In simple terms, when a server is just running for five hours then all you need is to pay for that five hours. If it is one second then you pay for one second. This feature is extremely beneficial for automating simple straightforward tasks in your company that occasionally used in certain period. For example, an attendance machine server only requires you to boot up when your employee takes a fingerprint to save it into your database. Or, one of the real world cases is to handle website hot traffic hours by increasing your memory during those times which will give better performance for your end users with a bit expensive cost for a short time.