Microsoft Technology Licensing, LLC

Microsoft Corporation

Guhan Suriyanarayanan

Sarosh Cyrus Havewala

Christophe Franck Robert

Huisheng Liu

Nikolaj Skallerud Bjorner

Shobana Balakrishnan

Dan Teodosiu

James Ernest Johnson

James M Lyon

Luis Irun-Briz

Daniel J Thompson

Nikolaj S Bjørner

Chien-Lung Yu

Robert Hanz

Michael K Liptack

Todd R Manion

Anatoliy Panasyuk

Ahmed Hassan Mohamed

Sandeep Singhal

Catharine Van Ingen

Yousef A Khalidi

Ryan Yonghee Kim

David Golds

Changhoon Kim

Steve E Olsson

Gaurav Omprakash Deshpande

Anuj Bindal

Bodhi Deb

Ahmed H Microsoft Corporation Mohamed

Nikolaj S Bjerner

Embodiments are directed to a distributed computing system comprising a plurality of compute nodes for providing resources to users and a hierarchy of two or more layers of controllers coupling the compute nodes to a user interface via a control plane, wherein at least one compute node receives a local application program interface (API) call from an application running on the at least one compute node, the local API call causing the at least one compute node to configure a local resource without requiring commands from the control plane.

Data plane API in a distributed computing network

Embodiments are directed to a distributed computing system comprising a plurality of compute nodes for providing resources to users and a hierarchy of two or more layers of controllers coupling the compute nodes to a user interface via a control plane, wherein at least one compute node receives a local application program interface (API) call from an application running on the at least compute node, the local API call causing the at least one compute node to configure a local resource without requiring commands from the control plane.

A message transport system may use a publication subscription mechanism to connect nodes and transport messages through the nodes. Each node may establish connections to other nodes, and subscription requests and publication notifications may be passed across the nodes to establish paths for messages. When a message is published, the message may be passed over those connections for which a subscription is active. A path identifier may be added to the message as it is passed between nodes, and the path identifier may be used by a subscribing node for identification of the information being received. When a subscriber notification is removed, the path may be deconstructed across multiple nodes. The nodes may be arranged such that each node is agnostic to any connections past the nodes to which it is connected, and may allow any node to subscribe to any information published within the network.

Message transport system using publication and subscription mechanisms

A regional data center system corresponding to a particular region of the world. The data center system includes multiple data centers, each configured such that region-specific requests for services may be satisfied by any of the data centers. One or more of the data centers may also perform load balancing on received region-specific request. In order to perform load balancing, the receiving data center may itself process the requests, but in some cases may determine that the request is to be processed by another data center within the regional data center system.

Distributed data center technology

Automatic file replication and scanning for malware in a sharing environment is based on detection of file system changes. Only one client of the multi-client environment needs to include malware protection software for all clients to benefit. A file replication service for each client monitors a shared directory, and synchronizes files across clients at the file system driver level. When a new file appears in the shared directory of a client system, through this driver, the new file automatically gets replicated. A malware protection component operates by monitoring the directory for new or modified files. File replication causes automatic malware scanning on the client that has protection. When the file has been cleaned and re-written to the directory, the replication service senses that the file has changed and replicates the cleaned file to the other clients. Thus, the replication behavior produces a multiparty file scanning and cleansing protection scheme.

Distributed malicious software protection in file sharing environments

A block based backup system may store backed up data from multiple clients by maintaining backup databases for each client, where each backup database may include blocks of data. The backup system may perform a de-duplication process by examining the databases to identify common blocks and moving the common blocks to a common database. The common blocks may be removed from the client backup databases and a pointer substituted for the data, thus compacting the client backup databases. The de-duplication operation may be performed at an opportunistic time, such as after the clients have completed any data transfer to the backup system.

Opportunistic asynchronous de-duplication in block level backups

Files on a computing device are defined to include data and metadata relating to the data. Data from each file has been removed therefrom and stored at an alternate location and the file is in a reduced, ghosted form. Each of at least some ghosted files has been fully or partially reconstituted for use by retrieving the data thereof from the alternate location and associating such retrieved data with such ghosted file to form the reconstituted file. Upon receiving a trigger based on a triggering event, files are selected from among the fully and partially reconstituted files and never-ghosted files based on first predetermined selection criteria, and each of at least some of the selected files are re-ghosted.

Moving data from file on storage volume to alternate location to free space

Aspects of the subject matter described herein relate to resource replication systems. In aspects, a mechanism is described that detects when local resources are stale. Local resources may be stale when the time between a last successful synchronization activity and a current time exceeds a staleness value. If the local resources are determined to be stale, the local member may refrain from certain synchronization activities until the local member is allowed to resume these certain synchronization activities. If the local resources are not stale, additional synchronization activities including synchronizing resources between the member and an upstream or downstream partner may be performed.

Resource freshness and replication

A backup file is generated to capture the current state of a computer. The computer state may be restored to the computer in a hybrid manner at a later time by accessing some files and/or data included in the backup file and downloading other files contained in a list in the backup file. By listing some files rather than including the actual files in the backup file, the backup file may be generated faster and require less processing and memory resources. When the computer state is restored, secondary files and computer state data contained in the backup file are installed to the computer. A list of primary files in the backup file is accessed, each primary file is retrieved from one or more remote publishers over a network, and the retrieved primary files are installed.

Hybrid computer restore using network service

Aspects of the subject matter described herein relate to replicating initial content in resource replication systems. In aspects, each member participating in a replica group obtains a content set state from a directory service. The content set state indicates whether the member is allowed to synchronize resources with another member participating in the replica group. If the member is designated as a primary member, the member may replicate out resources to its downstream partners. If a member is not designated as a primary member, the member may not replicate out resources to its downstream partners until it has successfully synchronized resources with an upstream partner. Fence values may be set which control which resources win (and are replicated out) during synchronization.

Pre-existing content replication

Aspects of the subject matter described herein relate to shutdown recovery for resource replication systems. In aspects, a mechanism is described in which a machine having replicated data thereon can recover from a dirty shutdown. First, the machine determines whether a dirty shutdown has occurred. If so, the machine automatically performs shutdown recovery by causing resource metadata stored by the machine to be consistent with resource data stored by the machine. This may involve fixing the resource metadata for updates to the resource data that were not flushed to disk or may involve deleting the resource metadata and restoring it from another machine replicating the data.

Shutdown recovery

Aspects of the subject matter described herein relate to resource replication systems. In aspects, a mechanism is described that maintains replicated resources on a slave member that participates in a replica group. The slave member detects that a local resource has been updated locally. If the resource was created locally in the update, the slave member deletes the local resource. Otherwise, the slave member determines an upstream partner from which to install a replacement resource and then installs the replacement resource. In aspects, this behavior undoes local updates to local resources that occur on the slave member.

Slave replica member

A computing device has a storage volume, a file system managing the storage volume, and a file stored on the storage volume and accessed by such file system. The file system receives an open command for the file from an application, notes that the file has been ghosted in that data of the file has been moved to an alternate location, and returns an error. The returned error is intercepted and the file system is commanded to retrieve ghosting information from the ghosted file. A handle corresponding to the ghosted file is delivered to the application, and a read command for the file is received from the application, including the handle, an offset, and a read length. Reconstituting of the ghosted file is initiated based on the data thereof at the alternate location, and after reconstituting the read command is responded to with the data requested.

To reconstitute a ghosted file for use, a single ghosting filter on a computing device locates ghosting information in the metadata of the ghosted file and locates within the ghosting information an identification of a particular ghosting manager of the computing device, where the identified ghosting manager is responsible for the ghosted file. Thereafter, the ghosting filter communicates to the identified ghosting manager a request to obtain data of the ghosted file from a corresponding alternate location, and the identified ghosting manager does in fact obtain such requested data from such alternate location. Upon receiving the requested data of the ghosted file from the identified ghosting manager, then, the ghosting filter reconstitutes the received data to the ghosted file. The single ghosting filter is generic to all of the ghosting managers of the computing device.

Growing community of inventors

Redmond, WA, United States of America

Guhan Suriyanarayanan