"network disruption as a service" and proof of local storage

In an email conversation on 27th March 2015, Wladimir J. van der Laan raised a concern about a proposed exercise to ensure more full nodes on the network. Matt Whitlock had suggested that an entity could pass the query to a random other node and forward its result in response to a challenge. Wladimir pointed out that it would not even be necessary for the entity to run a full node; they could just pass the query to a random other node and forward its result. The challenge required sending SHA256(SHA256(concatenation of N pseudo-randomly selected bytes from the block chain | prover's nonce | verifier's nonce)). The nonces are from the “version” messages exchanged at connection startup. A node can't pass the buck because it can't control the nonce that a random other node chooses. In his first draft, Matt had included the answer to the challenge but removed it as he thought it was unnecessary. He argued that there was no problem with a company hosting hundreds of lightweight servers all backed by a single copy of the blockchain since the goal of the exercise was to ensure more full nodes on the network.