Microsoft Technology Licensing, LLC

Other

Microsoft Corporation

Sunita Shrivastava

Joseph M Joy

Rod N Gamache

Nathan P Myhrvold

Gideon Andreas Yuval

Michael T Massa

Gor Nishanov

Steven J Friedman

Karen A Hargrove

David Bruce Lomet

Chittur Padmanabhan Subbaraman

Rohit Jain

Philip Alan Bernstein

David Allen Dion

Rajsekhar Das

John D Vert

Rushabh A Doshi

Gillian Furse

Philip A Bernstein

Glenn R Peterson

Christopher L Darling

Chittur P Subbaraman

A method and system for increasing server cluster availability by requiring at a minimum only one node and a quorum replica set of replica members to form and operate a cluster. Replica members maintain cluster operational data. A cluster operates when one node possesses a majority of replica members, which ensures that any new or surviving cluster includes consistent cluster operational data via at least one replica member from the immediately prior cluster. Arbitration provides exclusive ownership by one node of the replica members, including at cluster formation, and when the owning node fails. Arbitration uses a fast mutual exclusion algorithm and a reservation mechanism to challenge for and defend the exclusive reservation of each member. A quorum replica set algorithm brings members online and offline with data consistency, including updating unreconciled replica members, and ensures consistent read and update operations.

Consistent cluster operational data in a server cluster using a quorum of replicas

A dependency filter object for use in a server cluster is described. In an implementation, a method includes bringing online one or more resource objects included in a dependency tree. Each of the resource objects corresponds to a resource available in a server cluster. The dependency tree includes a dependency filter object that specifies a dependency of a first one of the resource objects on a second one of the resource objects. The bringing online of the first resource object is managed according to the specified dependency on the second resource object.

Dependency filter object

In a first exemplary media implementation, one or more processor-accessible media include processor-executable instructions that, when executed, direct a system to perform actions that include: accumulating host status information at multiple hosts; and sending the accumulated host status information from the multiple hosts. In a second exemplary media implementation, one or more processor-accessible media include processor-executable instructions that, when executed, direct a system to perform actions that include: receiving host status information from multiple hosts; and making load balancing decisions responsive to the received host status information. In a third exemplary media implementation, one or more processor-accessible media include processor-executable instructions that, when executed, direct a system to perform actions that include: determining health and load information on a per application basis; and selecting an application from among multiple applications responsive to the health and load information.

Network load balancing with host status information

A failover scope comprises a node collection in a computer cluster. A resource group (e.g., application program) is associated with one or more failover scopes. If a node fails, its hosted resource groups only failover to nodes identified in each resource group's associated failover scope(s), beginning with a first associated failover scope, in order, thereby defining an island of nodes within which a resource group can failover. If unable to failover to a node of a resource group's first failover scope, failover is attempted to a node represented in any next associated failover scope, which may require manual intervention. Failover scopes may represent geographic sites, whereby each resource group attempts to failover to nodes within its site before failing over to another site. Failover scopes may be managed by the cluster runtime automatically, e.g., an added node is detectable as belonging to a site represented by a failover scope.

Failover scopes for nodes of a computer cluster

In a server cluster, a system and method is provided for mitigating redundant resource failure notifications and other problems resulting from late handling of messages. Traditional resource management can result in the generation of redundant resource failure notifications that trigger unnecessary recovery actions, or cause other cluster problems such as performing an action that has previously been handled as part of failure recovery. The present invention tracks resource failures and eliminates recovery actions for redundant resource failure notifications. An incarnation number is passed to a resource each time it is called, and is incremented whenever a resource failure notification is delivered. Failure notifications having an incarnation number lower than the current incarnation number are discarded. Message processing similarly uses an incarnation number to distinguish between queued messages that correspond to those from a currently healthy node and those from a previous incarnation of the node, which no longer have meaning.

Use of incarnation number for resource state cycling

A method and system for increasing server cluster availability by requiring at a minimum only one node and a quorum replica set of replica members to form and operate a cluster. Replica members, independent from the nodes, maintain cluster operational data. A cluster operates when one node possesses a majority of replica members, which ensures that any new or surviving cluster includes consistent cluster operational data via at least one replica member from the immediately prior cluster. Arbitration provides exclusive ownership by one node of the replica members, including at cluster formation, and when the owning node fails. Arbitration uses a fast mutual exclusion algorithm and a reservation mechanism to challenge for and defend the exclusive reservation of each member. A quorum replica set algorithm brings members online and offline with data consistency, including updating unreconciled replica members, and ensures consistent read and update operations.

Method and system for consistent cluster operational data in a server cluster using a quorum of replicas

A method and system wherein following a partitioning of a server cluster, operational subgroups arbitrate for possession of a quorum resource that determines cluster representation, wherein the arbitration is biased by a relative weight of the subgroup. The weight may be relative to the original cluster weight, or submitted as a bid that is relative to other possible subgroup weights. The biasing gives subgroups that are better capable of representing the cluster an arbitration advantage over lesser subgroups. The biasing weight of each subgroup may be determined by node count and/or by a calculation of the subgroup's resources. The arbitration may be delayed based on the relative weight, or alternatively, the arbitration may comprise a bidding process in which a subgroup's bid is based on the subgroup's relative weight.

System for determining at subgroup of nodes relative weight to represent cluster by obtaining exclusive possession of quorum resource

A custom palette comprises colors on a plurality of luminance levels consisting of colors of uniform luminance. The spacing in luminance between levels varies in relation to the luminance, the spacing between levels being closer at low luminance and increasing at greater luminance. The palette preferably includes a gray color from each luminance level, highly saturated colors at the vertices of every alternate level, and a selection of lower saturation colors encircling the gray color on each level. The selection of lower saturation colors preferably includes a higher concentration of flesh and sky colors.

Luminance sensitive palette

A method and system for forming a new cluster of servers having state data that is consistent with the state data of a cluster that previously existed. The previous cluster preserves its state data by recording transactions to a log file in conjunction with a sequence number maintained by each node. The node that recorded the transaction then increments its copy of the sequence number and attempts to replicate the transaction to other nodes of the cluster. If the transaction is replicated, each other node commits the transaction and increments its sequence number. If the node fails before it can replicate the transaction, then no other nodes will increment their sequence numbers, whereby the next logged transaction has the same sequence number as the previously logged, but orphaned replicated transaction. When a new node forms a cluster, it unrolls the log file so as to become consistent with the state of the previous cluster. Logged transactions having duplicate sequence numbers indicate that the first such transaction was not fully replicated. To remain consistent, the first transaction is discarded and not used to update the state of the new cluster.

Method and system for discarding locally committed transactions to ensure consistency in a server cluster

A method and system for transparently failing over a legacy application from a first system to a second system of a server cluster by tracking and checkpointing changes to application configuration information stored in a system's local registry. When an application running on the first system makes a change to the application configuration information in a subtree of the registry, the change is detected and a snapshot of the subtree's data is taken. The snapshot is written to a storage device shared by systems of the cluster, such as a quorum disk. When the application is failed over to a second system, the snapshot for that application is retrieved from the quorum disk and written to the registry of the second system in a corresponding subtree. The application is then run on the second system using the most-recent application configuration information as modified by the other system in the cluster.

Method and system for transparently failing over application configuration information in a server cluster

A method and system for communicating modification information to servers in a server cluster. Local changes, such as modifications to a resource requested at one node, are associated into a single transaction. A master node, such as the node that owns the set of resources corresponding to the modifications in the transaction requests permission from a locker node to replicate the transaction. When permission to replicate the transaction is received from the locker node, the master node replicates the transaction by requesting each node in the cluster, one node at a time, to commit the transaction. Any node that does not commit the transaction is removed from the cluster, ensuring consistency of the cluster. Failure conditions of any node or nodes are also handled in a manner that ensures consistency.

Method and system for replicated and consistent modifications in a

A high color resolution image is accurately approximated with a low resolution image through a unique combination of ordered dithering and error diffusion. The true color of each pixel is modified with error from previously rendered pixels and then dithered to an intermediate color of 15 bits. The intermediate color is then mapped to the nearest displayable color in a displayable color palette using a precomputed look-up table. Any error between a displayed color of a pixel and its true color is calculated and spread among neighboring pixels not yet rendered.

Method and apparatus for mapping colors in an image through dithering

A high color resolution image is accurately approximated with a low resolution image through ordered dithering to colors of a displayable color palette. The colors of the displayable color palette are located at points of a grid aligned with three orthogonal axes in a three dimensional color model, including a luminance axis between black and white. The grid points are defined by the intersections of three sets of parallel planes, each set being perpendicular to one of the three axes. The set of planes perpendicular to the luminance axis are spaced nonuniformly, preferably according to a cubic model. The ordered dithering maps the true color of a pixel in the image to the color in the displayable palette at one of eight adjacent grid points. When ordered dithering of a pixel's true color would map the true color to a grid point outside the color gamut, the pixel is instead mapped to the nearest color on the color gamut's border to the grid point.

Growing community of inventors

Redmond, WA, United States of America

Sunita Shrivastava