Quote from: nullius on December 31, 2017, 10:33:10 AM

Quote from: Spendulus on December 31, 2017, 05:22:57 AM

Quote from: cellard on December 29, 2017, 04:04:44 PM

...
We already have studies that show at what rate nodes would get wiped out at current block size and in increases of 2MB up to 8MB. This study was developed by Bitfury:

So if BCash was getting used AND spammed as BTC does, blocks would get filled and 95% of current nodes would get wiped out of the system.

What does it mean "wiped out of the system?"

Is that computed as a side effect of the fact that people don't have, say for example, another 200 gb to dedicate to block chain for each year that passes?

Look at the numbers cellard posted! Do you have a home PC with 32GB spare RAM to dedicate to a Bitcoin node? Can your home connection pass 99.2GB of daily traffic? Thats all with full 8MB blocks; and according to Bitfury, all that hardware buys you a whopping 28tps. Its still 2 orders of magnitude under the throughput of Paypal, and 34 orders of magnitude under that of Visa.

I dont see any figures here on iops; but I can guess qualitatively. Got RAIDed enterprise-class SSDs?

Disk space is the smallest problem with big blocks. Nodes of modest means can prune. But the above table shows that there would be nothing to prune: They wouldnt be able to keep up with the network, or even run without getting hit by the OOM-killer.

(For comparison, the Steem documentation specifies a minimum requirement of 32GB RAM for a Steem node. Most users simply use steemit.com. So decentralized. But it has the magic word, blockchain.)

To compete on a practical basis with Visa cannot mean that every node has the entire transactional volume passing through it.

To compete with 1/100 the volume of Visa cannot mean that every node has the entire transactional volume passing through it.

Block sizes are not relevant to this issue.