摘要:
Computers are provided with a totally ordered, durable shared log. Shared storage is used and can be directly accessed by the computers over a network. Append-log operations are made atomic in the face of failures by committing provisional append ordering information onto a log. The log may comprise multiple flash packages or non-volatile memory devices, referred to as segments, although any shared storage device(s) may be used. Each log record is a multi-page stripe, where each page of a stripe is written to a different segment. Fault-tolerant protocol variants append stripes to the log, such that stripes are totally ordered in the log and each stripe is written atomically.
摘要:
Computers are provided with a totally ordered, durable shared log. Shared storage is used and can be directly accessed by the computers over a network. Append-log operations are made atomic in the face of failures by committing provisional append ordering information onto a log. The log may comprise multiple flash packages or non-volatile memory devices, referred to as segments, although any shared storage device(s) may be used. Each log record is a multi-page stripe, where each page of a stripe is written to a different segment. Fault-tolerant protocol variants append stripes to the log, such that stripes are totally ordered in the log and each stripe is written atomically.
摘要:
A prediction tree for estimating values of a network performance measure. Leaf nodes of the prediction tree are associated with networked computing devices and interior nodes are not necessarily representative of physical network connections. Values are assigned to edges in the prediction tree and the network performance measure relative to two computing devices represented by two nodes of the tree is estimated by aggregating the values assigned to the edges in the path in the prediction tree joining the two edges. Mechanisms for adding nodes representing computing devices to the prediction tree, for identifying a closest node representing a computing device in the prediction tree, for identifying a cluster of devices represented by nodes of the tree, and for rebalancing the prediction tree are provided.
摘要:
A remote procedure call chain is provided that replaces multiple consecutive remote procedure calls to multiple servers from a client by allowing a client to specify multiple functions to be performed consecutively at multiple servers in a single remote procedure call chain. The remote procedure call chain is executed by a sequence of multiple servers. Each server executes a service function and a chaining function of the remote procedure call chain. The chaining function uses the state of the remote procedure call chain in the sequence of servers to determine the next server to receive the remote procedure call chain, and the service function to be executed by that server. After the last service function is performed, the last server in the sequence of servers sends the results of the executed service functions to the client that originated the remote procedure call chain.
摘要:
The present invention extends to methods, systems, and computer program products for allocating and reclaiming resources within a rendezvous federation. Embodiments of the invention include formulating and routing messages in accordance with cached two-way agreements between adjacent nodes on a ring of nodes. Cached two-way agreements divide responsibility for at least unoccupied identifiers between the adjacent nodes on a ring. When a message is received at a node, the node can refer to a cached two-way agreement to determine if it or an adjacent node is responsible for the message. The node then sends the message to the next appropriate component in accordance with the cached two-way agreement.
摘要:
An enhancement to a web browser offers an extension mechanism for web applications to utilize client-side resources, thereby moving more extensibility and flexibility to the client-side. The web browser may expose some control interfaces to meet various requirements from web applications. Using the extension mechanism, the web applications are able to offload the cloud servers, function when offline, and improve performance. The extension mechanism also provides users with full control to the data passing through their web browsers.
摘要:
A primary node is elected from among a plurality of nodes in a replica set of nodes. The primary node accepts and processes client data access requests. The replica set includes the primary node and other secondary nodes. The primary node receives client data access requests and assigns a data sequence number to each client data access request that mutates state in the order the client data access requests are received. The data sequence numbers include a linearized processing order that is followed by each of the nodes in the replica set. The primary node sends the mutating client data access requests including any corresponding data sequence numbers to the secondary nodes. The primary node receives, from a threshold number of secondary nodes, an acknowledgement indicating reception of the client data access request. The primary node commits the data mutating access request.
摘要:
The present invention extends to methods, systems, and computer program products for a joining node to join a ring of nodes within a rendezvous federation. Embodiments include detecting a neighborhood of nodes on the ring of nodes. The joining node indicates its intent to take id-space ownership for a portion of the id-space between the joining node and a selected immediately adjacent node. The joining node indicates an intent to monitor the selected node. The joining node receives three indications, the first indicating acceptance of the joining node's intent to take id-space ownership for a portion of the id-space between the joining node and the selected node, the second indicating acceptance of the joining node's intent to monitor the selected node, the third indicating the first selected node's intent to monitor the joining node. The joining node indicates acceptance of the selected node's intent to monitor the joining node.
摘要:
In some embodiments, it is determined that a primary node has been elected from among a plurality of nodes in a replica set of nodes. The primary node accepts and processes client data access requests. The replica set includes the primary node and other secondary nodes. The primary node receives client data access requests. The primary node assigns a data sequence number to each client data access request that mutates state in the order the client data access requests are received. The data sequence numbers include a linearized processing order that is to be followed by each of the nodes in the replica set. The primary node sends the mutating client data access requests including any corresponding data sequence numbers to the secondary nodes. The primary node receives, from a threshold number of secondary nodes, an acknowledgement indicating reception of the client data access request. The primary node commits the data mutating access request.
摘要:
The present invention extends to methods, systems, and computer program products for allocating and reclaiming resources within a rendezvous federation. Embodiments of the invention include formulating and routing messages in accordance with cached two-way agreements between adjacent nodes on a ring of nodes. Cached two-way agreements divide responsibility for at least unoccupied identifiers between the adjacent nodes on a ring. When a message is received at a node, the node can refer to a cached two-way agreement to determine if it or an adjacent node is responsible for the message. The node then sends the message to the next appropriate component in accordance with the cached two-way agreement.