There's some calculations of that in the Lightning white paper which gives an estimate of about 138mb to manage channels for every person on the planet.
I have recently discovered it is very easy to extend the storage of a BTRFS filesystem and turn it into any kind of RAID array. With two you can do JBOD or RAID0 for more space or RAID1 for failure tolerance, or 3 and also maybe Raid5, 4 to get RAID10.
I think we are coming to a time where the cost of storage and extending storage drop quite dramatically and if it's cheap enaugh to put a full node on literally every router then I guess we can increase the block size :D
I know that throughput-wise transaction handling is at least 2000/s on the slowest implementation, BTCD - bitcoin core is a lot faster. So, I don't think there is any problem looming. When the block size is needed to grow we are in a position to grow it without a lot of loss of decentralisation. I think at least doubling can happen within the next 10 years. Very likely will.
Yes, I agree over time the technical requirement for running a Bitcoin node have gotten so much cheaper. Nowadays I see you can get a 2TB ssd on Ebay for just $50, pretty crazy!
I still believe that block size should be kept as small as possible and that it is important that the blocks of the future are full in order to provide a sufficient security budget to miners.
reply