Accelerated learning in adaptive rebuilding by applying observations to other samples

1. A method for execution by a computing device of a dispersed storage network (DSN) comprises:

obtaining, by the computing device, scoring information for a rebuilding for one or more storage units of a set of storage units of the DSN, wherein the scoring information includes a plurality of rebuilding rates, a plurality of input/output rates, a plurality of scores, and a plurality of selection rates;

determining based on the scoring information, by the computing device, that a first input/output rate of the plurality of input/output rates for a first rebuilding rate of the plurality of rebuilding rates exceeds a difference threshold compared to an initial first input/output rate for the first rebuilding rate, wherein the initial first input/output rate has a corresponding first score and a first selection rate;

adjusting, by the computing device, a plurality of initial input/output rates based on the first input/output rate to produce a plurality of updated input/output rates;

generating, by the computing device, an updated plurality of scores for the plurality of rebuilding rates based on the plurality of updated input/output rates; and

implementing, by the computing device, the rebuilding in accordance with the updated plurality of scores.

2. The method of claim 1 further comprises:

the plurality of input/output rates correspond to the set of storage units.

3. The method of claim 1 further comprises:

a first plurality of input/output rates correspond to a first storage unit of the set of storage units; and

a second plurality of input/output rates correspond to a second storage unit of the set of storage units.

4. The method of claim 1 , wherein the adjusting comprises:

applying a learning rate function to the plurality of initial input/output rates and the plurality of input/output rates to produce the plurality of updated input/output rates.

5. The method of claim 1 , wherein the adjusting comprises:

applying a first learning rate function for the first rebuilding rate to a first input/output rate of the plurality of initial input/output rates to produce a first updated input/output rate of the plurality of updated input/output rates; and

applying a second learning rate function for a second rebuilding rate to a second input/output rate of the plurality of initial input/output rates to produce a second updated input/output rate of the plurality of updated input/output rates.

6. The method of claim 1 further comprises:

calculating, by the computing device, an initial input/output rate for each rebuilding rate of the plurality of rebuilding rates to produce the plurality of initial input/output rates.

7. A computing device of a dispersed storage network (DSN) comprises:

memory;

an interface; and

a processing module, wherein the processing module is operably coupled to the memory and the interface, and wherein the processing module is operable to:

obtain scoring information for a rebuilding for one or more storage units of a set of storage units of the DSN, wherein the scoring information includes a plurality of rebuilding rates, a plurality of input/output rates, a plurality of scores, and a plurality of selection rates;

determine based on the scoring information that a first input/output rate of the plurality of input/output rates for a first rebuilding rate of the plurality of rebuilding rates exceeds a difference threshold compared to an initial first input/output rate for the first rebuilding rate, wherein the initial first input/output rate has a corresponding first score and a first selection rate;

adjust a plurality of initial input/output rates based on the first input/output rate to produce a plurality of updated input/output rates;

generate an updated plurality of scores for the plurality of rebuilding rates based on the plurality of updated input/output rates; and

implement the rebuilding in accordance with the updated plurality of scores.

8. The computing device of claim 7 , wherein the processing module is further operable to:

utilize the plurality of input/output rates for the set of storage units.

9. The computing device of claim 7 , wherein the processing module is further operable to:

utilize a first plurality of input/output rates for a first storage unit of the set of storage units; and

utilize a second plurality of input/output rates for a second storage unit of the set of storage units.

10. The computing device of claim 7 , wherein the processing module is operable to adjust the plurality of initial input/output rates by:

applying a learning rate function to the plurality of initial input/output rates and the plurality of input/output rates to produce the plurality of updated input/output rates.

11. The computing device of claim 7 , wherein the processing module is operable to adjust the plurality of initial input/output rates by:

applying a first learning rate function for the first rebuilding rate to a first input/output rate of the plurality of initial input/output rates to produce a first updated input/output rate of the plurality of updated input/output rates; and

applying a second learning rate function for a second rebuilding rate to a second input/output rate of the plurality of initial input/output rates to produce a second updated input/output rate of the plurality of updated input/output rates.

12. The computing device of claim 7 wherein the processing module is further operable to:

calculate an initial input/output rate for each rebuilding rate of the plurality of rebuilding rates to produce the plurality of initial input/output rates.