Facebook, Inc.

Mpath Interactive, Inc.

Hearme

Jeffrey Rothschild

Daniel Joseph Samuel

Michael Andrew Wolf

Robert M English

Marc Peter Kwiatkowski

Stephen M Grimm

Peter Vajgel

Jason Scott Sobel

Robert Carlton Johnson

Norman Robert Black

Steven Michael Grimm

Dhrubajyoti Borthakur

Nagavamsi Ponnekanti

Norman R H Black

Marc Peter Kwiatowski

Conrad Donald Wong

A load balancer allocates requests to a pool of web servers configured to have low queue capacities. If the queue capacity of a web server is reached, the web server responds to an additional request with a rejection notification to the load balancer, which enables the load balancer to quickly send the rejected request to another web server. Each web server self-monitors its rejection rate. If the rejection rate exceeds a threshold, the number of processes concurrently running on the web server is increased. If the rejection rate falls below a threshold, the number of processes concurrently running on the web server is decreased.

Load balancing web service by rejecting connections

Load balancing by circling back to a secondary port of a first web server after a request is rejected a threshold number of times

A variety of methods for improving efficiency in a database system are provided. In one embodiment, a method may comprise: generating multiple levels of data according to how recently the data have been updated, whereby most recently updated data are assigned to the newest level; storing each level of data in a specific storage tier; splitting data stored in a particular storage tier into two or more groups according to access statistics of each specific data; during compaction, storing data from different groups in separate data blocks of the particular storage tier; and when a particular data in a specific data block is requested, reading the specific data block into a low-latency storage tier.

LSM cache

Receiving at a secondary port of a first web-server a request that has been rejected a predetermined amount of time by multiple web-servers

A data object management scheme for storing a large plurality of small data objects (e.g., image files) in small number of large object stack file for storage in secondary storage (e.g., hard disks). By storing many individual data objects in a single object stack file, the number of files stored in the secondary storage is reduced by several orders of magnitude, from the billions or millions to the hundreds or so. Index data for each object stack file is generated and stored in primary storage to allow efficient and prompt access to the data objects. Requests to store or retrieve the data objects are made using HTTP messages including file identifiers that identify the files storing the data objects and keys identifying the data objects. A file server stores or retrieves the data object from secondary storage of a file server without converting the requests to NSF or POSIX commands.

Efficient storage and retrieval for large number of data objects

A network match making system and method is used to match users of a multi-users networked application. Each user is associated with a client computer connected to the network. Clients are selected based on attributes of their users, the clients, servers, and/or communication links. The network match maker works with three different forms of network applications: peer-to-peer, multiple clients to a single server, and multiple clients to multiple servers. In one match maker method, a client computer and one multi-user application server are matched. A list of multi-user application servers connected to the computer network that are available for a match is received. At least one communication attribute of respective data communication links between the client computer and a plurality of multi-application servers in the received list is measured. An additional attribute, such as a server attribute, is received from at least one of the plurality of multi-user application servers. A subset of one or more multi-user application servers is then selected based on the measured communication attributes and the additional attribute received from the multi-user application servers.

Network match maker

A method and apparatus to efficiently and effectively ensure that the quality of the data communications links are adequate for digital electronic game playing is provided. A number of servers are connected to the Internet. The servers periodically perform communications link quality measurements among themselves. At least one of the servers is designated as a matchmaker. When a game is requested by clients, the matchmaker acts as a game coordinator and compiles the results of the measurements. Based on the compilation and other information from its database, the matchmaker selects a few servers as candidate servers. These candidates perform data communications link quality measurements with the requesting clients. Based on these measurements, the matchmaker selects a server as the server for the requested game.

Latency server and matchmaker

A method for deploying interactive applications over a network containing host computers and group messaging servers is disclosed. The method operates in a conventional unicast network architecture comprised of conventional network links and unicast gateways and routers. The hosts send messages containing destination group addresses by unicast to the group messaging servers. The group addresses select message groups maintained by the group messaging servers. For each message group, the group messaging servers also maintain a list of all of the hosts that are members of the particular group. In its most simple implementation, the method consists of the group server receiving a message from a host containing a destination group address. Using the group address, the group messaging server then selects a message group which lists all of the host members of the group which are the targets of messages to the group. The group messaging server then forwards the message to each of the target hosts. In an interactive application, many messages will be arriving at the group server close to one another in time. Rather than simply forward each message to its targeted hosts, the group messaging server aggregates the contents of each of messages received during a specified time period and then sends an aggregated message to the targeted hosts. The time period can be defined in a number of ways. This method reduces the message traffic between hosts in a networked interactive application and contributes to reducing the latency in the communications between the hosts.

Server-group messaging system for interactive applications

A networked computer online gaming system and process arranged in a client/server online gaming architecture and utilized to run gaming programs. The client computers are configured to run a gaming client program. The server computers are coupled to the client computers via a network. The server computers run server programs including a master control program (MCP) that governs access of the server programs to the online gaming architecture, a servorum program (SV) for creating instances of a server program, a matchmaker program (MM) that supports rendezvous services, a game instances class server program (GICS) that enables games and provides user to user communication, and game upper level protocol server program (GULP) that supports the user to user communication provided by said GICS.

Online gaming architecture

A network match making system and method is used to match users of a multi-user networked application. Each user is associated with a client computer connected to the network. Clients are selected based on attributes of their users, the clients, servers, and/or communication links. The network match maker works with three different forms of network applications: peer-to-peer, multiple clients to a single server, and multiple clients to multiple servers. In one match maker method, remote server computers and client computers are matched. A match maker receives a plurality of client requests and a plurality of server requests. At least two instances of client programs and at least one instance of server programs are selected based on client and server attributes.

A network match making method and system matches an additional computer executing an instance of a client computer program with one or more other computers executing instances of client computer programs. The computers are linked by one or more communication links. A match maker computer program receives a request from the instance executing on the additional computer to match with an instance executing on another computer. A measurement request is sent to the instance executing on the additional computer. Communication attributes between the additional computer and each existing computer are measured and an existing computer is selected as a match for the additional computer if one of the communication attributes exceeds a minimum performance requirement. Existing computers can be further selected based on further attributes. If none of the communication attributes meets the minimum performance requirement, a match maker program waits for a further request. The communication attributes measured can relate to at least one parameter such as latency, packet lost rate, and bandwidth.

Growing community of inventors

Los Altos, CA, United States of America

Jeffrey Rothschild