If You Can Identify the Target, You Can Hack the Target
One can imagine the glee shared by various nation state adversaries to the US in learning that a large amount of highly classified DoD data will soon be stored in two special built Microsoft Azure data centers.
Now PLA Unit 61398 (https://en.wikipedia.org/wiki/PLA_Unit_61398) can get up early, enjoy some green tea and dim sum, stretch out a bit (https://www.youtube.com/watch?v=HXmls3bApMU) and then focus their efforts on hacking these centralized repositories of highly classified data. The target is clear, the attack surfaces can be measured and the attack vectors can be determined. It is go time!
Now imagine an alternate scenario, whereby this data was stored on Storj’s network (https://www.youtube.com/watch?v=vl3bUzfn2lg). Each file would be broken into 20 pieces (shards), made redundant 2x using Reed Solomon erasure encoding, then each of the 40 shards would be encrypted client-side and distributed across 40 geographically dispersed nodes in Storj’s decentralized network.
The net effect would be like blowing a handful of encrypted sand across a large, wind swept encrypted beach (https://www.youtube.com/watch?v=BaaqYB83H6E), good luck finding the right grains of sand, much less putting them back together again.
In terms of data security, Storj’s network presents an unclear target, an unidentifiable attack surface, and unknown attack vectors. If you can not identify the target, then you can not hack the target.
John Quinn, Co-founder Storj Labs