Problem with content addressing and its immutability
Since IPFS, using the hash value of a content as the immutable address for decentral location becomes popular. In our development, we realize that once the content is generated with its immutable hash, then who is going to serve it with some level of quality of service. Consumers hope some nodes will keep these immutable data for the community, but hosting data online with immutability for free is kind of imaginary. The best you get for free is someone host immutable content for a while. Then the “while” service costs immutability, since the data itself can not warrant its existence.
A new idea from TAU dev is to forgetting the immutable data concept. Accepting all data communication is mutable in nature with various changing content at the same time in a probability distribution. From A to B communication, B will receive various data from A at the same time. There is no guarantee of data delivery or integrity. The community blockchain consensus is the only place to link up those data and maintain integrity. Everything else is mutable with probability to exist or disappear.
A potential mutable data structure is that a node will use “bloom filter” with false positive risk to summary a certain period message history, then publish it along with messages on the same mutable data channel. Let receiver to get those pieces of data and take risks to reorganize them.
The community accumulative history of bloom filters can help the community from secrete chain attack. The attacker’s secrete chain messages will not be popularly recorded in the bloom filters from community history.
On these data collection, POT concensus protocol can help link up the transactions and form up a blockchain for the community, where we might be able to find certain level of immutability as long as community peers are sufficient.