To protect against node failures and downtime, the Storj network leverages Reed-Solomon erasure coding. Erasure coding takes an encrypted file and breaks it down into small pieces. The most significant benefit is that you only need 29 of 80 pieces to recover the file. In the previous network (V2), mirrors were spread across the network and replication would initiate if a storage node would go offline for a certain amount of time. This meant that a significant amount of upload and download bandwidth overhead was necessary to sustain the network. In the new network (V3), all pieces will be parity shards and replication will occur by specific network links, thus removing the overhead present before. This insures that your files will always remain available whenever you need them.
Articles in this section
- What happens if a node hosting one of my chunks goes offline ?
- Can I set conditions about the purpose which my space I offer for rent can be used for ?
- How important is download speed and bandwidth ?
- How much storage do I need to share to be able to run a storage node on Storj?
- What is the throughput limitation of the Storj network ?
- What is Storj's policy on illegal content ?
- If the space I'm sharing goes offline, what happens to the data I'm storing?
- How would one not allow their node to host certain content ?
- How can an average person keep their storage space constantly online in order to maximize the service provided and the profit?