I've been thinking about upgrading my personal-use, home-network hardware - most of my equipment was acquired used about ten years ago. The capabilities of current hardware is a bit staggering. I'm wondering about the pros & cons of two different architectures:
- A few small (desktop-class hardware) machines (a workstation, a NAS, a network services machine)
- One large (server-class hardware) machine that does it all.
Often, large system performance metrics are based on a single application under heavy load. In my situation, there would be many [hundreds(?) of] applications/daemons, each under low load most of the time. Some of these might be large footprint (e.g., PostgreSQL), while others would be relatively low resource.
How well could OpenBSD utilize a six-core Xeon machine with 32GB of ECC RAM for my use-case? Would the GENERIC kernel need to be tweaked?