Abstract: Systems and methods for clock synchronization are disclosed in which a primary node generates special physical laver clock sync symbols from the output of a reference clock and inserts the clock sync symbols within a symbol stream to one or more secondary nodes. Upon receiving a symbol stream, a secondary node can extract the clock sync symbols from the stream to synchronize its local clock with the reference clock of the primary node. In particular, the clock sync symbols can be inserted into the symbol stream at any arbitrary symbol location, e.g., even between consecutive symbols of a symbol encoded data frame. The clock sync symbols can also replace some control symbols in the symbol stream, such as idle or comma symbols. Accordingly, the clock sync symbols can be inserted into a symbol stream at fixed intervals, irregular intervals, or at any arbitrary time for high resolution clock synchronization.

Abstract: A method and apparatus for fraud detection during transactions using identity graphs are described. A method includes receiving, at a commerce platform system, a transaction from a user having initial transaction attributes and transaction data. The method also includes determining, by the commerce platform system, an identity associated with the user associated with additional transaction attributes not received with the transaction. Furthermore, the method includes accessing a feature set associated with the initial transaction attributes and the additional transaction attributes that includes machine learning (ML) model features for detecting transaction fraud.

Abstract: A distributed computing system, such as may be used to implement an electronic trading system, supports a notion of fairness in latency. The system does not favor any particular client. Thus, being connected to a particular access point into the system (such as via a gateway) does not give any particular device an unfair advantage or disadvantage over another. That end is accomplished by precisely controlling latency, that is, the time between when request messages arrive at the system and a time at which corresponding response messages are permitted to leave. The precisely controlled, deterministic latency can be fixed over time, or it can vary according to some predetermined pattern, or vary randomly within a pre-determined range of values.

Abstract: A method and apparatus for fraud detection during transactions using identity graphs are described. The method may include receiving, at a commerce platform system, a transaction from a user having initial transaction attributes and transaction data. The method may also include determining, by the commerce platform system, an identity associated with the user, wherein the identity is associated with additional transaction attributes not received with the transaction. Furthermore, the method may include accessing, by the commerce platform system, a feature set associated with the initial transaction attributes and the additional transaction attributes, wherein the feature set comprises machine learning (ML) model features for detecting transaction fraud.

Abstract: Systems and methods for clock synchronization are disclosed in which a primary node generates special physical layer clock sync symbols from the output of a reference clock and inserts the clock sync symbols within a symbol stream to one or more secondary nodes. Upon receiving a symbol stream, a secondary node can extract the clock sync symbols from the stream to synchronize its local clock with the reference clock of the primary node. In particular, the clock sync symbols can be inserted into the symbol stream at any arbitrary symbol location, e.g., even between consecutive symbols of a symbol encoded data frame. The clock sync symbols can also replace some control symbols in the symbol stream, such as idle or comma symbols. Accordingly, the clock sync symbols can be inserted into a symbol stream at fixed intervals, irregular intervals, or at any arbitrary time for high resolution clock synchronization.

Abstract: An electronic trading system and corresponding method are based on a point-to-point mesh architecture. The electronic trading system comprises a gateway, core compute node, and sequencer. The core compute node performs an electronic trading matching function. The gateway transmits a message to the core compute node via a first direct connection. The gateway transmits the message via a second direct connection to the sequencer which, in turn, transmits a sequence-marked message to the core compute node via a third direct connection. The core compute node determines relative ordering of the message among other messages in the electronic trading system based on the sequence-marked message to complete the electronic trading matching function, deterministically.

Abstract: Systems and methods are disclosed herein that provide low latency data communication with improved physical link layer reliability through repeated physical layer retransmission of data over a data connection whenever the connection is idle (i.e., no new data to send). In some embodiments, the transmission of administrative control symbols (e.g., “idles,” “commas”, etc.) can be suppressed or otherwise reduced to allow some of the available connection bandwidth to be used for data redundancy and fault tolerance through the repeated retransmission of data as described in more detail below. Accordingly, instead of executing time-consuming, error correcting routines, a receive node can discard the erroneous data frame and process at least one repeat frame that carries the same data payload. Sequence numbers and/or other repeat indicators can be used to distinguish original frames from repeat frames and/or for identifying which frames carry the same data payload.

Abstract: An electronic trading system and corresponding method are based on a point-to-point mesh architecture. The electronic trading system comprises a gateway, core compute node, and sequencer. The core compute node performs an electronic trading matching function. The gateway transmits a message to the core compute node via a first direct connection. The gateway transmits the message via a second direct connection to the sequencer which, in turn, transmits a sequence-marked message to the core compute node via a third direct connection. The core compute node determines relative ordering of the message among other messages in the electronic trading system based on the sequence-marked message to complete the electronic trading matching function, deterministically.

Abstract: A distributed computing system, such as may be used to implement an electronic trading system, supports a notion of fairness in latency. The system does not favor any particular client. Thus, being connected to a particular access point into the system (such as via a gateway) does not give any particular device an unfair advantage or disadvantage over another. That end is accomplished by precisely controlling latency, that is, the time between when request messages arrive at the system and a time at which corresponding response messages are permitted to leave. The precisely controlled, deterministic latency can be fixed over time, or it can vary according to some predetermined pattern, or vary randomly within a pre-determined range of values.

Abstract: A distributed computing system, such as may be used to implement an electronic trading system, controls inbound message flow rates. Limiting a per-client or per-connection inbound message rate also helps ensure fair provisioning of computing resources, so that a single client's excessive use of resources cannot overwhelm the system to such an extent that it prevents other clients from interacting with the distributed system. It is also desirable to have system-wide control of the overall inbound message rate across all client connections. Such system-wide control ensures that the distributed system as a whole can maintain the required levels of service, including offering a predictable level of access for all clients.

Abstract: Systems and methods are disclosed herein that provide low latency data communication with improved physical link layer reliability through repeated physical layer retransmission of data over a data connection whenever the connection is idle (i.e., no new data to send). In some embodiments, the transmission of administrative control symbols (e.g., “idles,” “commas”, etc.) can be suppressed or otherwise reduced to allow some of the available connection bandwidth to be used for data redundancy and fault tolerance through the repeated retransmission of data as described in more detail below. Accordingly, instead of executing time-consuming, error correcting routines, a receive node can discard the erroneous data frame and process at least one repeat frame that carries the same data payload. Sequence numbers and/or other repeat indicators can be used to distinguish original frames from repeat frames and/or for identifying which frames carry the same data payload.

Abstract: Systems and methods for clock synchronization are disclosed in which a primary node generates special physical layer clock sync symbols from the output of a reference clock and inserts the clock sync symbols within a symbol stream to one or more secondary nodes. Upon receiving a symbol stream, a secondary node can extract the clock sync symbols from the stream to synchronize its local clock with the reference clock of the primary node. In particular, the clock sync symbols can be inserted into the symbol stream at any arbitrary symbol location, e.g., even between consecutive symbols of a symbol encoded data frame. The clock sync symbols can also replace some control symbols in the symbol stream, such as idle or comma symbols. Accordingly, the clock sync symbols can be inserted into a symbol stream at fixed intervals, irregular intervals, or at any arbitrary time for high resolution clock synchronization.

Abstract: A distributed computing system, such as may be used to implement an electronic trading system, controls inbound message flow rates. Limiting a per-client or per-connection inbound message rate also helps ensure fair provisioning of computing resources, so that a single client's excessive use of resources cannot overwhelm the system to such an extent that it prevents other clients from interacting with the distributed system. It is also desirable to have system-wide control of the overall inbound message rate across all client connections. Such system-wide control ensures that the distributed system as a whole can maintain the required levels of service, including offering a predictable level of access for all clients.

Abstract: A distributed computing system, such as may be used to implement an electronic trading system, supports a notion of fairness in latency. The system does not favor any particular client. Thus, being connected to a particular access point into the system (such as via a gateway) does not give any particular device an unfair advantage or disadvantage over another. That end is accomplished by precisely controlling latency, that is, the time between when request messages arrive at the system and a time at which corresponding response messages are permitted to leave. The precisely controlled, deterministic latency can be fixed over time, or it can vary according to some predetermined pattern, or vary randomly within a pre-determined range of values.

Abstract: A document collaboration process is centered around a digital media object that is the subject of the collaboration, and makes the role of the server computer that is hosting the collaboration session more transparent to the user. As a result, a user can initiate a collaboration session without logging into the server computer and uploading the digital media object into the server computer. A user can also join a collaboration session without logging into the server computer. The process for initiating a collaboration session may be performed directly from the browser while viewing a web page, directly from an e-mail interface while viewing an e-mail, directly from an Instant Messaging Interface while participating in an IM session or directly from a file browser while viewing a list of files stored in the user computer.

Abstract: A search engine graphical user interface (GUI) displays a search result, which may contain a single page or multiple pages, in an independently updatable section, such as an inline frame. With this implementation, a user can update the display contents of a search result, e.g., by navigating to a different page of a multi-page search result, without disrupting the rest of the web page, and a user can create or add to a search listing collection by grabbing a search result using an input device and dropping it into a virtual basket for collecting and organizing search results.

Abstract: A search engine generates search results and at least one advertisement to be displayed adjacent to the search results, wherein the search results and the at least one advertisement are to be displayed in independently updateable sections. The search results include one or more search listing collections, and the advertisement that is to be displayed adjacent to a search listing collection is selected based on its relevance to the particular search listing of the search listing collection that is to be displayed therewith. The advertisements may be a collection of advertisements that are displayed one at a time in an independently updateable section.

Abstract: A search engine presents search results with user-editable abstracts and ranks search results based on the user-edited abstracts. A graphical user interface is provided to permit users to edit abstracts of search results. This enables owners of web sites and local businesses with local search listings to deliver abstracts of their products and services in their own words and improve the search ranking of their web sites or local search listings by observing abstract composing rules that are optimized to the relevance algorithm that is applied by the search engine.

Abstract: Search results generated in response to a search query includes one or more search listing collections and are ranked based on usage of search listing collections. In the case where search results including individual search listings are ranked, the usage metrics include: the number of times a search result has been added to a search listing collection, the number of times a search result has been viewed within a search listing collection, and the number of click-throughs or the click-through rate of a search result within a search listing collection.

Abstract: An image digitizer including a pixel engine connected to both a video memory and a processor. The pixel engine is configured to enable pixels to be transferred out of memory or read into memory while bypassing the processor thereby allowing more efficient transfer of data into and out of the image digitizer.