公开(公告)号：GB2514555A

公开(公告)日：2014-12-03

申请号：GB201309484

申请日：2013-05-28

Applicant: IBM

Inventor： CIDECIYAN ROY DARON ,  JELITTO JENS ,  SARAFIJANOVIC SLAVISA ,  STANEK JAN

IPC: G06F3/06 ,  G06F11/14 ,  G06F17/30

Abstract: Data objects, to be stored on a storage medium, are divided into data segments or chunks. A hash of each data segment is generated. The hash tag and the physical location at which the segment is to be stored on the medium are stored in an index. When a new data segment is processed, its hash is used to determine whether an identical segment is already stored on the medium. If so, the new data segment is not stored. Otherwise, the new data segment is stored in a location, which is physically close to other data segments in the same data object. Consecutive segments of a data object may be stored in a single extent on the medium. The data object may be a file. The storage medium may be a sequential access medium, such as a magnetic tape.

公开(公告)号：GB2517913A

公开(公告)日：2015-03-11

申请号：GB201315434

申请日：2013-08-30

Applicant: IBM

Inventor： JELITTO JENS ,  MITTELHOLZER THOMAS ,  SARAFIJANOVIC SLAVISA ,  SORNIOTTI ALESSANDRO ,  STANEK JAN

IPC: G06F17/30 ,  G06F3/06

Abstract: A method includes providing (S10), by at least one client, files to be stored on a remote server. The method then includes evaluating (S20) popularity of the files (e.g. the number of copies provided by different users). And storing (S30) the files on the server in a form that depends on the popularity of the files, files with a first level of popularity being stored in a deduplicated form and files with a second level of popularity being stored in an encrypted form, the first level of popularity being higher than the second level of popularity. More popular files may be seen as not intended to be confidential and therefore storable in a duplicated form. The method may further include repeatedly providing copies of the same file by different users, thereby increasing popularity of the file, and comprises storing said copies in a semantically secure encrypted form before the file reaches the first level of popularity, and in the deduplicated form after said file has reached the first level of popularity. The method may further include, upon the file reaching the first level of popularity, decrypting previously stored copies and performing deduplication on said copies.