Shelby fun opinions

@shelbyserves interest is growing very fast and there are fantastic questions about the data lifecycle. My opinions below:

1) High-performance decentralized hot storage is a new category. The most important one that Shelby serves. It’s necessary for any real-time serving use case.

2) Data has a lifecycle. Working on data infrastructure at Meta for 7 years - we saw the server data would be highly accessed for the first 1-3 days, counting DAU for example. Then it would drop to less than 1/10 of the accesses and then after a retention period, move to archival storage or just be deleted. In the supercomputing space it was parallel file systems like Lustre, Panasys, GPFS, etc. were used for “scratch” storage. NFS might be the longer term storage. Shelby, in the fullness of time, should support the full data lifecycle.

3) Hot data in @shelbyserves should be cached and there should be incentives to supporting ultra low-latency reads and paying for in-memory costs.

4) There will be compute layers built on Shelby. The white paper describes several different use cases and approaches.

5) Shelby is for all devs. Large enterprises and small startups. All blockchains. Shelby is the answer for building real-time decentralized apps, for everyone.