| 摘要 |
FIELD: information technology.SUBSTANCE: method of replacing a failed node storing distributed data, comprising steps of: receiving, through a first storage node, a first set of shares generated from a data file, each share in the first set including a linear combination of pieces of a data file along with a set of coefficients used to generate the linear combination; receiving, through the first storage node, an indication of a new storage node replacing the failed node, wherein the failed node includes a second set of shares generated from a data file; through the first storage node, generating a first replacement share as a response to said indication, the first replacement share generated by: multiplying each share in the first set and set of coefficients by a random scaling value and combining the multiplied first set of shares and the multiplied set of coefficients; and through the first storage node, transferring the generated first replacement share to the new storage node. The first replacement share and at least another replacement share form a second set of shares in the new storage node, said other replacement share being generated by a second storage node.EFFECT: improved reliability of storing data.23 cl, 5 dwg |
