Abstract: The disclosed computer-implemented method for marking application-consistent points-in-time may include intercepting, by an I/O filter, a write request from a guest virtual machine to a virtual machine disk and queueing the write request in an I/O filter queue. The method may include sending the write request to the virtual machine disk and receiving a write completion message from the virtual machine disk. The method may also include sending, in response to the write completion message, the write request to an I/O daemon, and queueing the write request in an I/O daemon queue. The method may further include sending the write completion message to the guest virtual machine, and sending the write request to a backup gateway such that the backup gateway mimics writes to the virtual machine disk. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: The disclosed computer-implemented method for marking application-consistent points-in-time may include intercepting, by an I/O filter, a write request from a guest virtual machine to a virtual machine disk and queueing the write request in an I/O filter queue. The method may include sending the write request to the virtual machine disk and receiving a write completion message from the virtual machine disk. The method may also include sending, in response to the write completion message, the write request to an I/O daemon, and queueing the write request in an I/O daemon queue. The method may further include sending the write completion message to the guest virtual machine, and sending the write request to a backup gateway such that the backup gateway mimics writes to the virtual machine disk. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: An online system generates dynamically optimized content items composed of creatives selected from a set of creatives provided by a content provider according to a set of rules associated with the creatives. Creatives include the title, image, video, descriptive text and other different types of components. The online system also receives rules describing one or more actions that can be performed on each of the creatives and under what condition for the assembly of the content item. For a target user of the content item, the online system applies the rules to remove creatives that violate the rules. Each creative that satisfies the rules is analyzed and ranked based on the likelihood that the target user will interact with a content item that includes that particular creative. For a different user, a different sponsored content item having different creatives chosen from the same set of creatives is generated.

Abstract: An online system generates dynamically optimized sponsored content for a target user of the online system. Each sponsored content item comprises optimal creatives selected for a target user from a set of creatives provided by a content provider. Each type of creative (e.g. title, image, video, descriptive text), has a trained creative model to generate a prediction score for a creative of the same type based on the features of the creative and the characteristics of the target user. The prediction score of a creative indicates the likelihood that the target user will interact with a sponsored content item that includes that particular creative. The online system selects a creative for each type and assembles the selected creatives into a sponsored content item for the target user. For a different user, a different sponsored content item having different creatives chosen from the same set of creatives is generated.

Abstract: An online system receives a set of creatives provided by a content provider, and presents one or more pseudo-assembled content items composed of the different combinations of the received creatives on a user interface to the content provider. A pseudo-assembled content item includes one or more creatives to be included in a final content item that are placed in their positions in the display interface, but the content item has not yet undergone assembly or creation. The positions of the creatives are defined by one or more placement rules provided by the content provider. The content provider can interact with the user interface to swap different creatives into the content item. The content provider can visually preview different content item candidates assembled from different permutations of creatives from the set of creatives of the content item before creating the final content item.

Abstract: The disclosed computer-implemented method for facilitating analytics on remotely stored data sets may include (1) identifying, within a secondary storage system, a secondary copy of a data set duplicated from a primary copy of the data set stored in a primary storage system, (2) generating a set of virtual objects that represent at least a portion of the secondary copy of the data set, (3) exposing the set of virtual objects to a remote analytics engine via a network such that the portion of the secondary copy of the data set appears to be stored locally on the remote analytics engine, and then (4) enabling the remote analytics engine to perform at least one analytics job on the portion of the secondary copy of the data set by way of the set of virtual objects via the network. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: The disclosed computer-implemented method for efficient placement of virtual machines may include (1) allocating space in a cache shared by a group of virtual machines to add a new virtual machine, (2) receiving data requests from the new virtual machine for the cache, (3) recording each of the data requests as a cache hit or a cache miss in a list based on availability of the data in the cache, (4) determining a ratio of cache hits to cache misses for the new virtual machine based on the recorded data requests, and (5) placing the new virtual machine in the group of virtual machines when the ratio of cache hits to cache misses exceeds a threshold, such that the data backup device efficiently utilizes the cache for servicing the data requests from the new virtual machine. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: The disclosed computer-implemented method for load balancing backup data may include (1) receiving a request to backup files in a multi-node computing cluster, (2) identifying a backup distribution of the files among multiple backup clients, (3) reading an initial data block of a current file from a data node in the cluster, (4) reading a copy of the initial data block of an additional file from another data node in the cluster, (5) reading a subsequent data block of the current file from the data node in the cluster, and (6) balancing backup of the current and additional files among the data node and the another data node by reading a copy of a subsequent backup data block of the additional file from the another data node in the multi-node computing cluster. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: The disclosed computer-implemented method for load balancing backup data may include (1) receiving a request to backup files in a multi-node computing cluster, (2) identifying a backup distribution of the files among multiple backup clients, (3) reading an initial data block of a current file from a data node in the cluster, (4) reading a copy of the initial data block of an additional file from another data node in the cluster, (5) reading a subsequent data block of the current file from the data node in the cluster, and (6) balancing backup of the current and additional files among the data node and the another data node by reading a copy of a subsequent backup data block of the additional file from the another data node in the multi-node computing cluster. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: The disclosed computer-implemented method for accelerating access to data may include (1) monitoring, at a data-caching system, read requests of a first data-accessing system for a dataset managed by a data-management system, (2) identifying a pattern of the read requests of the first data-accessing system, (3) monitoring, at the data-caching system, read requests of a second data-accessing system for the dataset managed by the data-management system, (4) determining that a pattern of the read requests of the second data-accessing system resembles the pattern of the read requests of the first data-accessing system, and (5) using a portion of the dataset accessed by the read requests of the first data-accessing system to pre-warm a cache of the second data-accessing system. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: An online system displays a content item generated from the ranked candidate creatives after an ad campaign or during the running of an ad campaign of the content item. The online system may present the content item through a display interface and allows the content provider to view the content item that is composed of different combinations of candidate creatives in view of performance statistics of the creatives, e.g. click through rate, number of “Likes”, number of audiences reached by the ad campaign of the content item, and viewers' engagement with the individual creatives or the content item as a whole. The online system can select a number of top ranked creatives for a content item for a target audience and adjust the selected creatives' ranking scores based on the performance statistics associated with the selected creatives.

Abstract: An online system generates dynamically optimized content items composed of creatives selected from a set of creatives provided by a content provider according to a set of rules associated with the creatives. Creatives include the title, image, video, descriptive text and other different types of components. The online system also receives rules describing one or more actions that can be performed on each of the creatives and under what condition for the assembly of the content item. For a target user of the content item, the online system applies the rules to remove creatives that violate the rules. Each creative that satisfies the rules is analyzed and ranked based on the likelihood that the target user will interact with a content item that includes that particular creative. For a different user, a different sponsored content item having different creatives chosen from the same set of creatives is generated.

Abstract: An online system receives a set of creatives provided by a content provider, and presents one or more pseudo-assembled content items composed of the different combinations of the received creatives on a user interface to the content provider. A pseudo-assembled content item includes one or more creatives to be included in a final content item that are placed in their positions in the display interface, but the content item has not yet undergone assembly or creation. The positions of the creatives are defined by one or more placement rules provided by the content provider. The content provider can interact with the user interface to swap different creatives into the content item. The content provider can visually preview different content item candidates assembled from different permutations of creatives from the set of creatives of the content item before creating the final content item.

Abstract: An online system provides feedback to a content provider creating a content item for a target audience. The feedback may include a score, recommendation, or error notification for a creative such as an image, video, or text to be included in the content item. The score indicates a likelihood that users of the online system will interact with the content item having the creative. Modifying the content item based on recommendations may result in a different score for the content item. The online system trains a machine learning model to generate the scores. The model learns which creatives are popular among particular audiences. The online system provides error notifications if the content item violates a rule. The online system can generate the content item even if there are rule violations. The feedback is displayed inline on a graphical user interface while the content provider is creating the content item.

Abstract: An online system provides a content data model to content providers for optimizing content creation. The content data model is a hierarchical model with multiple levels for content creation, e.g., campaign level, content item set level and content item level. At each level of the content data model, a content provider can specify certain information concerning the content creation at that level, such as optimization goal for each level. The information specified at each level of the content data model is applied to all elements under that level by the online system during the content creation process. With the content data model, a content provider can efficiently design creative campaigns by specifying objectives, optimization goals, target audiences and budgets, etc., at different design levels. The online system dynamically optimizes content item creation based on information about creatives to be included in a content item for a target user.

Abstract: The disclosed computer-implemented method for facilitating analytics on data sets stored in remote monolithic files may include (1) identifying, within a secondary storage system, a secondary copy of a data set duplicated from a primary copy of the data set stored in a primary storage system, (2) generating a set of virtual objects that represent at least a portion of the secondary copy of the data set, (3) exposing the set of virtual objects to a remote analytics engine via a network such that the set of individual data objects appears to be stored locally on the remote analytics engine, and then (4) enabling the remote analytics engine to perform at least one analytics job on the set of individual data objects by way of the set of virtual objects via the network. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: An online system generates dynamically optimized sponsored content for a target user of the online system. Each sponsored content item comprises optimal creatives selected for a target user from a set of creatives provided by a content provider. Each type of creative (e.g. title, image, video, descriptive text), has a trained creative model to generate a prediction score for a creative of the same type based on the features of the creative and the characteristics of the target user. The prediction score of a creative indicates the likelihood that the target user will interact with a sponsored content item that includes that particular creative. The online system selects a creative for each type and assembles the selected creatives into a sponsored content item for the target user. For a different user, a different sponsored content item having different creatives chosen from the same set of creatives is generated.

Abstract: A method and system for incremental backup is disclosed. The method includes writing a first update to a secondary node and writing a second update to the secondary node. The secondary node comprises a replication storage group (RSG). Further, the order in which the updates are written at the secondary node is consistent with the order in which the updates were written at a primary node.

Abstract: A computer-implemented method for performing incremental backups may include 1) identifying a volume of data previously subjected to a backup, 2) determining a time to perform an incremental backup of the volume, the incremental backup being incremental with respect to the backup, and, at the time of the incremental backup, 3) identifying a list of files changed on the volume since the backup, 4) identifying a map of blocks changed on the volume since the backup, and 5) including, in the incremental backup, each block within at least one file that is indicated as changed in the map of blocks. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: A system and method for determining the data that should be backed up by a plurality of backup policies are described. A global file change log that lists which files have been changed may be maintained. The global file change log may be processed to distribute the file change information into policy-specific file change logs. A global block change bitmap may also be maintained. The global block change bitmap may indicate which data blocks of the files have been changed. Performing a respective backup operation for a respective backup policy may include backing up only the data blocks that have changed in the files listed in the respective policy-specific file change log since a previous backup operation for the respective backup policy was performed. The data blocks that have changed may be specified in a respective policy-specific block change bitmap for the respective backup policy.