Space reclaiming subprotocol

[diogohmcruz]

The algorithm for managing the disk space reserved for the backup service is not specified. Each implementation can use its own. However, when a peer deletes a copy of a chunk it has backed up, it shall send to the MC channel the following message:
REMOVED <Version> <SenderId> <FileId> <ChunkNo> <CRLF><CRLF>
Upon receiving this message, a peer that has a local copy of the chunk shall update its local count of this chunk. If this count drops below the desired replication degree of that chunk, it shall initiate the chunk backup subprotocol after a random delay uniformly distributed between 0 and 400 ms. If during this delay, a peer receives a PUTCHUNK message for the same file chunk, it should back off and restrain from starting yet another backup subprotocol for that file chunk.