Sun Microsystems, Inc.

Oracle America, Inc.

Raphael J Rom

Glenn Carter Scott

Germano Caronni

Amit Narayan Gupta

Israel Cidon

Asad Khamisy

Moshe Sidi

Phanindra Jujjavarapu

Man-Tung Tony Hsiao

Christoph L Schuba

Clifford James Buckley

Asad Khamisky

Tony Hsiao

A method for re-routing a request in a distributed system, that includes sending the request for an originating object, receiving the request at a root node of the originating object, determining whether the root node comprises a first forwarding pointer associated with the originating object, determining a first target object using the first forwarding pointer, if the first forwarding pointer is associated with the originating object, and re-routing the request to the first target object, wherein the distributed system implements an overlay network for message delivery.

Symbolic links in a distributed system

A method for managing object evolution in a distributed object store (DOS) involving requesting an update of an object, wherein the object includes an active globally unique identifier (AGUID) object and at least one version globally unique identifier (VGUID) object, wherein the at least one VGUID object includes a first generation number and a first serializer name, locating a first serializer using the first serializer name, wherein the first serializer is associated with the first generation number, obtaining an order of the update using the first serializer, and creating a new VGUID object, wherein the new VGUID object includes a new version number, the first generation number, and the first serializer name.

Method and system for managing independent object evolution

A method for locating a resource involves maintaining a capacity object configured to store server information for a plurality of servers, transmitting a resource request from a client to an indirection server comprising the capacity object, selecting one of the plurality of servers to service the resource request using server associated with the capacity object to obtain a selected server, and redirecting the resource request from the indirection server to the selected server.

Efficient resource balancing through indirection

A method for locating a free resource involves maintaining an address space containing a plurality of regions, wherein each of the plurality of regions is mapped with a server to obtain a mapping, transmitting a request for the free resource from a client to a request address that belongs to one of the plurality of regions, determining a selected server using the mapping, and directing the request for the free resource to the selected server.

Resource location by address space allocation

A method for storing a first copy of an object, including obtaining the object including a globally unique identifier (GUID), generating a first derived GUID using the GUID of the object, storing the first copy of the object identified by the first derived GUID in a root node of the object, publishing possession of the first copy of the object identified by the first derived GUID by the root node of the object, associating a first shadow root with the first copy of the object, and storing the first copy of the object identified by the GUID in the first shadow root.

Method and system for creating and using shadow roots

A method for monitoring a target node in a distributed system, that includes determining a plurality of neighbor nodes of the target node, determining a plurality of neighbor watch nodes, wherein the plurality of neighbor watch nodes are selected from the plurality of neighbor nodes, monitoring at least one selected from the group consisting of data sent by the target node and data received by the target node, using at least one of the plurality of neighbor watch nodes to obtain tracking information, and determining, using at least one the plurality of neighbor watch nodes, an action to perform using the tracking information and a response policy, wherein the action is specified in the response policy, wherein the distributed system implements an overlay network for message delivery.

Method and apparatus for monitoring a node in a distributed system

A method for approving a response or a decision of an observed node in a distributed system that includes generating at least one selected from the group consisting of the response and the decision by the observed node, forwarding the at least one selected from the group consisting of the response and the decision to at least one of the plurality of chaperones associated with the observed node based on a chaperone scheme, and approving the least one selected from the group consisting of the response and the decision by the plurality of chaperones using a chaperone voting policy and a chaperone approval policy to obtain at least one selected from the group consisting of a approved response and a approved decision, wherein the distributed system implements an overlay network for message delivery, and wherein the observed node and the plurality of chaperones communicate using the overlay network.

Chaperones in a distributed system

A method for serializer maintenance and coalescing in a distributed object store (DOS) including a first partition and a second partition, involving requesting an update of an object, wherein the object includes an active globally unique identifier (AGUID) object and at least one version globally unique identifier (VGUID) object, wherein the least one VGUID object includes a first generation number and a first serializer name, determining whether a first serializer is located in the first partition using the first serializer name, wherein the first serializer is associated with the first generation number, if the first serializer is not located in the first partition, constructing a second serializer using the first serializer name, assigning a second generation number to the second serializer, obtaining an order of the update to the object using the second serializer, and creating a new VGUID object.

Method for serializer maintenance and coalescing

The invention involves a method for aggregating data in a distributed system. The method includes specifying an aggregation scope including a plurality of nodes, where the plurality of nodes includes a destination node and a plurality of non-destination nodes. The method also includes defining an aggregation tree for the plurality of nodes using at least one Distributed Hash Table, where the destination node is the root of the aggregation tree and where each of the plurality of non-destination nodes is one selected from a group of an interior node and a leaf node in the aggregation tree. The method also includes determining a final aggregation result for the aggregation tree using an aggregation function and aggregation data stored on the destination node and the plurality of non-destination nodes.

Method and apparatus for data aggregation in a distributed system

The invention is directed to a flexible method of allocating groups of virtual circuits in a network as a virtual circuit bunch. A virtual circuit bunch of arbitrary number can be set up with the ease of setting up a single virtual circuit. A single virtual circuit bunch can have plural destinations. Further, a virtual circuit bunch can be routed to the same destination over different routes. Connections across the network can be initiated as separate virtual circuits and then be grouped together as a virtual circuit bunch. The implementation of virtual circuit bunches permits rapid setup of virtual circuits and reduced table sizes at individual nodes of the network. A fast connect service can be implemented using virtual circuit bunches without establishing a connection by assigning a fast connect packet to one virtual circuit of a virtual circuit bunch. Virtual circuit bunches permit cell interleaving problems found in ATM switches to be circumvented in a particularly convenient way.

Techniques for improving virtual channel management and maintenance in a network environment

The invention is a method and apparatus for determining an approximate network distance using one or more reference points. In accordance with an embodiment of the invention, the method comprises the steps of selecting at least one reference point positioned along a path between first and second points of a network, generating first distance metric information associated with at least one path associating a first point and the at least one reference point, generating second distance metric information associated with at least one path associating a second point and the at least one reference point, and determining a total approximate distance between the first point and the second point along one or more paths based on the first and second distance metric information. In one or more embodiments, the second distance metric information is published to one or more servers providing domain name server mapping information and the second distance metric information is provided to the first point along with requested mapping information. In one or more embodiments, total approximate distance information between a client and multiple servers providing a desired service is determined and a server associated with a shortest total approximate distance is selected from which to obtain the service.

Method and apparatus for determining approximate network distances using reference locations

A method for load balancing including creating a network, having a plurality of servers, to service a single multicast address using a source specific join, where the source specific join allows each of the plurality of servers to specify a source internet protocol address range that each of the plurality of servers services. Further, method includes reallocating the source internet protocol address range specified for at least one of the plurality of servers using a load balancing policy and a control multicast channel while at least one of the plurality of servers is handling communications.

Reallocating address spaces of a plurality of servers using a load balancing policy and a multicast channel

Network and host efficiencies are improved by reducing the overhead associated with establishing virtual circuits. In one approach, a request for information from a client is sent to a server using a connectionless protocol such as UDP. If the requested information satisfies a policy for return by the connectionless protocol, the response is sent that way. If the policy is not satisfied, the server may reply with a message to try a connection oriented protocol such as TCP. If no response is received at all after a certain number of tries, the client will try a connection using a connection oriented protocol. In a second approach, when a request from a client is sent using a connectionless protocol, the state information for a transaction TCP (T/TCP) connection is set up in the client, giving the server the option of responding either using the connectionless protocol or using T/TCP. Transparent proxies can be conveniently implemented which will handle asymmetric connections and efficiently return cached information.

Network and end-host efficiency for web communication

A routing engine for a node controller of a switch or for a router provides enhanced routing for unicast and multicast connections by representing preferred routing trees, associated with different bandwidths and quality of service indications, as tables having, for each entry, a node identification, a parent node identification, a link identifier and ah effective distance by which a node identified by the node identification is separated from the controller or router. The routing engine adapts quickly to changes in network configurations, such as link failure, by quickly patching a table affected by the failure in order to maintain service until a new preferred routing tree can be calculated. Recalculations of preferred routing trees are avoided by using said table to determine whether a new link or a restored link could possibly provide a more cost effective route than that specified in the existing table.

Efficient caching of routing information for unicast and multicast connections

Methods and apparatus for maintaining connectivity of nodes in a wireless LAN. Accordingly, the present invention provides a method and apparatus for maintaining connectivity in a wireless LAN. The LAN is divided into a plurality of cells. Each cell is occupied by at least one access point for communicating information between the cells, and at least one node for communicating via the LAN through the access points. To initiate a handoff from a first access point with which the node is currently communicating, the node first selects a second access point as a candidate for handoff. The node then communicates an instruction to the first access point instructing the first access point to relay a request to the second access point that the second access point accept a handoff of the node from the first access point. The first access point then relays the request to the second access point. In response, the second access point transmits its operating parameters to the node through the first access point, and the node adjusts its own parameters in response so as to establish communication between the node and the second access point, thereby accomplishing a handoff.

Growing community of inventors

Palo Alto, CA, United States of America

Raphael J Rom