Ebay, Incorporated

Paypal, Inc.

Other

Rami El-Charif

Ashok Srinivasa Murthy

Rema Hariharan

Sathyamangalam Ramaswamy Venkatramanan

Sanjay Pundlkrao Ghatare

Debashis Saha

Muhammad Faisal Rehman

Steven Aihua Chen

Olivier G Dumon

Armond Bigian

Tim Kimmet

Sangjin Lee

Bhaven Avalani

Jian Xu

Connie Wong Yang

Guanglei Song

Jean-Christophe Martin

Jiang Xu

Timothy Gordon Kimmet

Alan Honkwan Ngai

Ankit Khera

Mohit Soni

Meghdoot Bhattacharya

Taiping Yu

Fei Li

Xin Li

Suresh Mathew

Vaibhav Desai

Sami Ben-Romdhane

Shankar Jothi

Connie Y Yang

Krishnakanth Batta

Sami Ben Romdhane

Jigar Chandrakant Desai

Muthukumar Rethinakaleeswaran

Varna Puvvada

Guanglie Song

A load balancer receives a sequence of requests for computing service and distributes the requests for computing service to a computing node in an ordered list of computing nodes until the computing node reaches its maximum allowable compute capability. Responsive to an indication that the computing node has reached its maximum allowable compute capability, the load balancer distributes subsequent requests for computing service to another computing node in the ordered list. If the computing node is the last computing node in the ordered list, the load balancer distributes a subsequent request for computing service to a computing node other than one of the computing nodes in the ordered list of computing nodes. If the computing node is not the last computing node in the ordered list, the load balancer distributes a subsequent request for computing service to another computing node in the ordered list of computing nodes.

Software load balancer to maximize utilization

Systems and methods of cloud deployment optimization are disclosed. In some example embodiments, a method comprises running original instances of an application concurrently on original servers to implement an online service, receiving, by the original instances of the application original requests for one or more functions of the online service, receiving a command to deploy a number of additional instances of the application, transmitting synthetic requests for the function(s) of the online service to one of the original servers according to a predetermined optimization criteria, deploying the number of additional instances of the application on additional servers using a copy of the original instance of the application, and running the deployed additional instances of the application on their corresponding additional servers concurrently with the original instances of the application being run on their corresponding original servers.

System and method for cloud deployment optimization

A provisioning machine may receive a request that an application be executed while distributed according to a distribution constraint among various devices. The provisioning machine may access a topological model that represents multiple devices configured as a single cloud-based application server and defines a first group of devices that have the same redundancy status (e.g., active or backup). In addition, the topological model may define a second group of devices that have the same functional role (e.g., executing a particular component of the application). A device may be a member of both the first group and the second group. The provisioning machine may determine a size of the first group according to the distribution constraint. Based on the determined size of the first group, the provisioning machine may configure (e.g., provision) the first group of devices as a subset of the multiple devices of the server.

Provisioning of devices

A system and method for adaptively identifying and correcting issues in a computing system, such as a distributed node computing system, are described. The method includes receiving node data from a group of nodes, the node data describing one or more operational characteristics of a node. The operational characteristics may include CPU load, memory load, latency, or other operational data that describes node performance. Reachability data for the group of nodes is generated by trying to contact each node. Code version data is generated for each node that identifies which version of code applications in the node are running. The nodes are grouped into clusters using density-based clustering to identify outliers. A correlation is determined between the reachability, code version, and outlier data to identify problems and issue corrective actions.

Computational node adaptive correction system

Methods and systems for simulating a search, for the purpose of evaluating one or more scoring functions used in ordering item listings for presentation in a search results page are described. Consistent with some embodiments, a simulation platform includes a real-time simulation module that receives search result sets for search queries that result in the conclusion of a transaction. The result set is then processed by the simulation platform with one or more test scoring functions, such that the resulting position of the item listing that has resulted in the transaction can be compared with the actual position at which the item listing was displayed in the actual search results. For each test scoring function, an average rank shift metric is determined, and displayed, thereby providing a metric with which to base decisions about which scoring functions to use in the production system.

Methods and systems for selecting an optimized scoring function for use in ranking item listings presented in search results

Methods and systems for deriving a scoring function, for use in ordering a set of search results, are described. Consistent with some embodiments, a simulation platform includes an offline simulator module that receives search result sets for search queries that result in the conclusion of a transaction. The result set is then processed by the offline simulation platform to derive a set of weighting factors for use with one or more parameters in a parametric scoring function. The weighting factors are determined, for example, by specifying an equation setting one or more constraints and then solving the equation for the weighting factors.

Methods and systems for simulating a search to generate an optimized scoring function

In various example embodiments, a system and method for managing a server cluster are provided. An example method may include scheduling a first job on a first node, using a first resource manager, establishing a service for a second resource manager on a second node, wherein the service is allocated node resources of the second node, and attempting to schedule a second job on the first node, using the first resource manager. The method may include preempting the service on the second node, using the second resource manager, in response to the attempt to schedule the second job on the first node, and deallocating the node resources of the second node from the service. The method may include advertising, using a node manager of the first resource manager, available node resources of the second node, and scheduling the second job on the second node, using the first resource manager.

Systems and methods for resource sharing between two resource allocation systems

A data storage system may be configured to allocate replica-sets in a balanced manner and mark some of these balanced replica-sets as being spares. As one or more drives or machines fail, the data storage system may move all copies of an affected replica-set to a marked spare replica-set and mark the affected replica-set as being inactive or invalid. As the failed drives are replaced, the data storage system may reconfigure those inactive replica-sets and use them as new spares. The data storage system may implement a coordinator module that handles the balancing and allocation of spares within a sub-cluster. The coordinator may also reallocate entire replica-sets across sub-clusters to maintain balance at the cluster level.

Allocation of replica-sets in a storage cluster

Growing community of inventors

San Jose, CA, United States of America

Rami El-Charif