|
|
|
|
|
by paulmd
3948 days ago
|
|
Anything with SD card storage will perform very poorly. Anything that uses USB 2.0 as a system bus will perform like dogshit. Using real SSDs on a real SATA channel makes an immense difference, not just in performance but also in system stability. SD cards are meant to be linearly loaded up with files, dumped to disk, then erased. They do not handle highly random write loads with sectors of extreme write intensity very well, their wear levelling is very minimal and sooner or later you will burn them out if you're not careful. The typical way around this is using a write-protected SD card for bootstrapping (if necessary). You load a minimal driver stub, and then you either boot from a USB stick or you do a PXE boot from an image on the network. USB 2.0 is way too limited to handle your entire system disk+network traffic+data disk load. You want USB 3.0 at a minimum if your computer is set up like that. If you just want something real minimal maybe the SD card options are OK, but I highly, highly recommend a SSD-based system if you can swing it at all. If you're not going to see a ton of traffic, maybe you could build it out once and run it on your corporate network? That could be more appealing than a hosted or colo'd solution, depending on the task. |
|