The hardware requirements of running a Node are pretty low. You need a system that can run 24/7 and that has at least 600 GB of storage. The storage is needed to save a local copy of the Bitcoin blockchain which is growing over time and hence it's wise to account for future growth too when setting up your system. It's also possible to run a pruned version of the Bitcoin blockchain which would take up a lot less storage space and is probably what most will end up doing eventually.
The following is sufficient to run a Lightning Node such as Umbrel.
- 4GB of RAM - 600+ GB of storage, SSD highly preferable.
You should pick your hardware according to your reasons for running a node. How critical is 24/7 uptime and how robust do you want your setup to be? Do you want to be able to account for disk failure? Here is a basic run through of popular options.
Rusty old laptop
If you have an old laptop or computer lying around somewhere, running a node on it can be a feasible endeavor as long as it has enough storage space and RAM. However, laptops are generally not designed for 24/7 operation over long periods of time and if the laptop is old, be prepared for hardware failure at some point. I'd generally advise against using such a system for a Lightning node you plan to put any serious liquidity into.
Single board computer (Raspberry Pi)
A popular hardware platform for node operators. The Raspberry Pi is a single board computer that can be expanded with external storage to make up a perfectly suitable system. It's cheap to buy and the form factor makes it very appealing for something that needs to run 24/7. It's still not a particularly robust platform however. The SD cards tend to fail quite frequently, especially on systems with any sort of heavy IO, so make sure you get a high quality card.
Enterprise Server Grade Setup
If you're serious about creating a more robust hardware setup, you might want to consider enterprise grade hardware. The cost is significant but you can use the hardware for a large selection of services, not just a Lightning Node.
As an example, consider the following setup.
- A 1U SuperMicro server with 4 drive bays, a single CPU and numerous RAM slots for easy expansion. - 2-4 large capacity SSD's
For the host operating system, Proxmox is a great open source virtualization platform based on QEMU/KVM. You can then set up your disks as a ZFS storage pool with mirroring which instantly gets you local storage redundancy. If one disk fails, you won't lose any data and when you replace the drive the mirroring brings the new disk up to speed automatically. Couple that with ZFS snapshots that can be created at no cost (due to CoW) and sent off site and you have a pretty robust and reliable setup with hardware that is designed to take a beating 24/7 for years.
Your Lightning node software of choice can then be run in a VM.