Efficient application-aware disaster recovery

摘要

Embodiments are directed to efficiently backing up portions of data and to performing a scoped data recovery. In an embodiment, a computer system preloads data images with data corresponding to various different software applications or operating systems. The computer system chunks the data images, so that each data image is divided into multiple data chunks, and where each data chunk is represented by a hash value. The computer system then receives, from a user, portions of delta data representing data differences between the received user data and the data in the preloaded data images. The computer system also generates a logical backup representation that includes the data chunk hash values for the preloaded data images as well as the delta data received from the user. This logical representation allows restoration of the user's data using only the hash values and the delta data.

主权项

1. At a computer system including at least one processor and a memory, in a computer networking environment including a plurality of computing systems, a computer-implemented method for efficiently backing up one or more portions of data, the method comprising:
an act of preloading one or more data images configured for implementation in data backups for a plurality of users, the data images comprising data corresponding to one or more software applications or operating systems; an act of chunking the preloaded data images, such that each data image is divided into a plurality of data chunks, each data chunk being represented by a hash value; an act of receiving, from a first user, one or more portions of delta data representing data differences between the first user's data and the data in the preloaded data images; and an act of generating a logical backup representation that represents a backup of the first user's data, the logical backup representation including the data chunk hash values for the preloaded data images and the delta data received from the user, wherein the logical representation allows restoration of the first user's data using the hash values to access the chunked preloaded data images and the delta data.