I work with virtualized computer systems; mostly VMware. Our storage solutions include HPE. On the storage side of things, HPE and other big name enterprise storage solution providers offer very flexible/scalable solutions to meet the storage challenges the EOS block producers would need to have. Within VMware, the maximum supported VMDK size on an VMFS-5 datastore is increased to 62 TB. ... So the logical disk limitations presented to a client VM would be a bit of a problem. But the VM itself could simply take advantage of ZFS to overcome this, if a single large volume is required at more than 62TB. But I'm guessing EOS has , or will soon have a method of breaking up the block chain data across multiple block producer node storage volumes. Also, there's linux file linking tools baked in to address problems like this; where individual volume sizes are an issue and you need to reference files physical localed on a separate location or volume...

On the bandwidth side of the conversation... I'm not as familiar with the exact problem because the manufacturing facility I work at doesn't use high bandwidth on a WAN (Wide Area Network) basis. But I have little doubt that in large cities, a EOS block producer could purchase multi-GB/s WAN links, perhaps even with hot failover connections.

Posted using Partiko Android