System and method for object-based continuous data protection

A system for protecting data in a protected storage system, the system comprising:

• a management unit configured to:
• store, in an object-based storage, a replication of the protected storage system from a predetermined time point before a time window, in a plurality of fixed-size replicated content objects of a target disk, wherein a replicated content object stores data of an address apace of the protected storage system, wherein a replicated content object can be re-written but cannot be modified;
• obtain data stored on the protected storage system following a write operation in the protected storage system, and metadata related to the stored data;
• maintain a journal comprising a plurality of fixed-size journal content objects stored in the object-based storage and a plurality of fixed-size metadata objects stored in the object-based storage, by:
• including a copy of the obtained data in a content entry in one or more of the journal content objects;
• including at least some of the metadata in a metadata entry in one or more of the metadata;
• and
• associating the metadata entry with the content entry;
• and
• periodically apply the data received before a predetermined time window from the journal content objects to at least one replicated content objects by obtaining the at least one replicated content object, modifying the at least one replicated content object and re-writing the at least one replicated content object,
• wherein the management unit is configured to periodically apply the data by:
• examining metadata objects in the journal to identify entries related to last writes to offsets, that occurred before the time window;
• generating a region-locator, wherein the region-locator includes a list of the offsets and, for an offset in the list, a reference to a content entry in a journal content object that includes data of the last write to the offset;
• using the region-locator to copy content of all last writes to a respective address space from the content entries to the corresponding replicated content object;
• and
• removing the copied content entries from the journal.