Balancing Software Architecture and Infrastructure: Avoiding Long-Term Pitfalls in Startups

I’ve been noticing a pattern in startups where the focus on software architecture exists but infrastructure is often overlooked or they quickly decide on serverless and call it a day.

That’s taking on a considerable amount of technical debt that will cause problems later on. A quick side note - software architects are not the same as infrastructure, having worked on both they are different skillsets but need to work in unison to deliver a solution.

What might haunt your decisions in the long term include:

• Security - often times people make short term decisions and when you decide to get SOC2 type 2 compliance it becomes evident how much rework is required.
• Budgets - I’ve worked with clients where we’ve reduce their costs in cloud services by 10x simply because they misunderstood needs. Serverless is a great way to burn through any budget for instance.
• Overengineering - the team insists on microservices even though the business only has 5000 daily users, not a path you want to be going down. The cost of maintaining that infrastructure can be significant, not to mention the resourcing needs.
• Underengineering - the system will not scale appropriately when needed. Often times you get the developers building against assumptions of the system so transfer to another environment becomes cumbersome.
• Overlooked items - think about backups, poor logging practices, disaster recovery that’s never been tested, high availability requirements provided to clients but not evaluated in the architecture.

This doesn’t mean I’m recommending a waterfall approach and figuring out every nuance in the system before starting. It’s worth doing enough upfront architecture and having the right people in the room when significant shifts are needed.