Infrastructure choices in projects are always a tough call. Some swear by "reliability and speed," while others push for "flexibility and scalability." What’s your take? Is it better to stick with on-prem servers or go all-in on cloud services? Let’s debate—what’s actually more advantageous in your experience? 🤔
Cloud or local? Which should take the lead in projects?
👁️ 9 views💬 3 replies❤️ 0 likes
3 Replies
I think the answer to this question varies from project to project, but in the long run, my preference is for cloud services. When we launched our first startup in 2018, we hosted our services on local servers, but as customer traffic grew, we faced serious delays in IP queries and disk I/O operations. Apart from the time loss, we also had to deal with server maintenance and technicians.
When we switched to AWS, all we had to do was pay a monthly bill, and everything else became automated.
Especially for startups, cloud architectures make more sense to minimize investment costs and quickly launch an MVP. For example, one project started with 10,000 requests per day but received 2 million requests in just 3 months. We couldn’t have kept up with this growth using local servers. The only challenge we faced with the cloud was unexpected bills—which we now manage using cost management tools and reservations.
Of course, local servers have their advantages: they can be preferred for certain data compliance and highly sensitive private operations. But 90% of the time, the cloud is more advantageous for me. For instance, in an e-commerce project in 2022, we were using local servers, and our servers crashed during Black Friday. The cloud allowed us to handle the sudden demand through automatic scaling.
What if your project is an app for managing medical appointments in real-time where latency can be critical? In that case, even though the cost and scalability benefits of the cloud are great, you might end up choosing on-premises or bare metal solutions because a millisecond delay in data transmission could change the outcome for a patient. Don’t you think that in scenarios like this, reliability and local control outweigh any theoretical advantages of the cloud? At least until edge computing technology matures fully.
I always thought local servers gave more control in prototyping and game projects, but doesn’t the flexibility of the cloud really take a load off at the end of the day? No storage headaches, instant scaling—what’s been less stressful in your experience?