Abstract: Techniques are provided for customizing authentication and for handling pre-authentication and post-authentication plug-ins in an access management system. Users may want to access a protected resource, such as an application, and apply customizations to the protected resource. The customizations can be applied through the use of plug-ins, such as pre-authentication and post-authentication plug-ins. After it is determined that the user has permissions to apply a specified plug-in, analysis is performed to ensure that the plug-in complies with system requirements and that the criteria for implementing the plug-in has been satisfied. A browser session and control of the application can then be forwarded to the user.

Abstract: A software developer proxy tool accesses microservice applications for a software development project by connecting the developer proxy tool to a common port on a computer network. The tool implements software and hardware to register a plurality of the microservice applications on connection ports that connect to the developer proxy tool at an address for the common port. Data requests among the microservices are handled by the developer proxy tool via the common port. The tool sequentially queries selected microservice applications on the respective connection ports to determine availability for completing a request. The tool receives responses back from microservices and directs the responses back to the requesting program. Failed requests trigger use of remote or third party microservice applications that may be available over an internet connection.

Abstract: The present disclosure is generally directed to an apparatus and method for a personalized user interface. The method may include a method for generating a personalized user interface is described. The method may include receiving user data and classifying the user data to a membership tier. The method may include generating a unique token as a function of the membership tier, associating the unique token with a plurality of service identifiers; and generating a personalized user interface as a function of a modified token. Generating the personalized user interface may include generating a plurality of event handlers and associated display elements, wherein each event handler is associated with a service identifier of the plurality of service identifiers, and the personalized user interface configures a user device to display each display element. Moreover, the method may include receiving an activation of an event and initiating an automated process.

Abstract: Disclosed is a system for converting a high-level runtime model to a low-level runtime model where the high-level runtime model runs on a client computer system, and the low-level runtime model runs on a server computer system. The server system has installed thereon a pool of hardware accelerators, and the low-level runtime model is targeted to the pool of accelerators. Outputs of the low-level runtime model are returned to the high-level runtime model as if the high-level runtime model computed the outputs.

Abstract: Delivering file system namespace features without on-disk layout changes is described. An iterator in a file system identifies an anomaly in a hierarchical data structure, which includes a root node at a highest level in a file set, includes directories of connected nodes corresponding to other levels in the file set, and includes leaf nodes at a lowest level in the file set. The iterator stores an annotation, about the anomaly, into an in-memory storage that is external to the file system. An asynchronous agent in the file system identifies the annotation, about the anomaly, which is stored in the in-memory storage, thereby enabling the asynchronous agent to correct the anomaly.

Abstract: The invention provides a method of initiating code including (i) storing an application having first, second and third functions, the first function being a main function that calls the second and third functions to run the application, (ii) compiling the application to first and second heterogeneous processors to create first and second central processing unit (CPU) instruction set architecture (ISA) objects respectively, (iii) pruning the first and second CPU ISA objects by removing the third function from the first CPU ISA objects and removing first and second functions from the second CPU ISA objects, (iv) proxy inserting first and second remote procedure calls (RPC's) in the first and second CPU ISA objects respectively, and pointing respectively to the third function in the second CPU ISA objects and the second function in the first CPU ISA objects, and (v) section renaming the second CPU ISA objects to common application library.

Abstract: A technology agnostic distributed publication subscription as a service (DPAAS) server-based environment is used as a wrapper for abstracting technology, protocol and/or vendor specific implementations of publication/subscriber functionality. Users may interface with the DPAAS system an application programming interface to connect and/or disconnect, subscribe and unsubscribe from different publication/subscriber services. Further, selection of underlined technology, protocol and/or other vendor specific solutions can be facilitated by leveraging a self-service onboarding portal. Applications register to the DPAAS, along with any underlying technologies used by the source service, so that communications may be seamless to the end user of the subscribing application. In doing so, the DPAAS solution is vendor, protocol, and/or technology independent.

Abstract: The present application relates to an in-memory computing module and method, and an in-memory computing network and a construction method therefor. The in-memory computing module comprises at least two computing submodules, and low latency can be achieved when computing units in the computing submodules access memory units. Multiple computing submodules present a symmetric layer design, and such a symmetric layer structure facilitates the construction of a topology network so as to achieve large-scale or ultra-large-scale computation. The memory capacity of the memory units in each computing submodule can be customized, and designed flexibly. These computing submodules are in a bonding connection, and the data bit width after the bonding connection may be positive integer multiple of the data bit width of the computing units, so that high data bandwidth is achieved.

Abstract: An information processing system includes a plurality of terminal devices, a client device, and a cloud including one or a plurality of server devices. The client device or the cloud includes: a processing instruction unit for designating a specific terminal device from among the plurality of terminal devices and instructing the specific terminal device to perform processing in accordance with a prescribed program; an authentication information storage unit for retaining authentication information necessary for carrying out the prescribed program; and an authentication information providing unit for providing, to the specific terminal device designated by the processing instruction unit, authentication information or, as a substitute therefor, temporary information, necessary for executing the processing in accordance with the program instructed by the processing instruction unit.

Abstract: Plugins that are independently written are executed in a collaborative manner to analysis a log. A plugin executing with respect to a particular node of a hierarchical data structure determines values for a set of keys based on information of the particular node and/or any ancestor nodes, and information stored in a shared repository. The plugin stores the values for the keys as additional information of the particular hierarchical node and/or into the shared repository. The plugin does not access information of non-ancestor nodes when executing with respect to the particular hierarchical node. Each plugin writes into and retrieves from the shared repository using the shared naming convention, thereby sharing information. The sequence of execution of the plugins is not dependent on dependencies amongst the plugins. If a dependent plugin requiring an output from a requisite plugin is first executed, the dependent plugin is flagged as pending and subsequently re-executed.

Abstract: Executing geospatial data workflows includes receiving a representation of a first workflow. It further includes receiving a request to execute a second workflow, the second workflow including a reference to the first workflow. At least one of the first workflow and the second workflow comprises at least one geospatial transformation operation. It further includes determining that the first workflow referenced by the second workflow is permitted to be accessed. It further includes, based at least in part on determining that the first workflow is permitted to be accessed, executing the second workflow, including executing the representation of the first workflow comprising the at least one geospatial transformation operation.

Abstract: Utilizing role types that extend underlying types. Within a program definition, a computer system identifies a statement assigning an object of an underlying type to a role type that extends the underlying type. The underlying type implements a first interface, and the role type implements a second interface. Based on the role type extending the underlying type, the computer system assigns the object of the underlying type to the role type. Based on assigning the object of the underlying type to the role type, the role type is usable to interact with the object of the underlying type using both the first interface and the second interface.

Abstract: A distributed sender driven Admission Control System (ACS) is described herein, leveraging Weighted-Fair Quality of Service (QoS) queues, found in standard NICs and switches, to guarantee RPC level latency service level objectives (SLOs) by a judicious selection of QoS weights and traffic-mix across QoS queues. ACS installs cluster-wide RPC latency SLOs by mapping LS RPCs to higher weight QoS queues, and coping with overloads by adaptively apportioning LS RPCs amongst QoS queues based on measured completion times for each queue. When the network demand spikes unexpectedly to predetermined threshold percentage of provisioned capacity, ACS achieves a latency SLO that is significantly lower than the state-of-art congestion control at the 99.9th-p and admits significantly more RPCs meeting SLO target when RPC sizes are not aligned with priorities.

Abstract: A content processing method is provided, including: obtaining first content blocks, the first content block including structured data; transmitting, through a single bidirectional communication link established based on an application layer protocol, the first content blocks in a streaming manner according to a first sequence for obtaining the first content blocks; receiving, through the single bidirectional communication link, second content blocks returned in a streaming manner, the second content blocks being obtained by performing content type conversion on the first content blocks, transmission of the first content blocks and reception of the second content blocks being asynchronously performed via the single bidirectional communication link; and outputting the second content blocks sequentially according to a second sequence for receiving the second content blocks.

Abstract: The present disclosure involves systems, software, and computer implemented methods for integrated data privacy services. An example method includes determining to initiate an integrated end of purpose protocol for an object. An end-of-purpose query is provided to multiple applications that requests each application to determine whether the application is able to block the object. End-of-purpose statuses are received, in response to the end-of-purpose query, that each indicate whether a respective application is able to block the object. The end-of-purpose statuses are evaluated to determine whether an aligned end of purpose has been reached for the object. In response to determining that the aligned end of purpose has been reached for the object, a block command is provided to each application that instructs the application to locally block the object in the application.

Abstract: An improved unit test framework that validates large datasets generated by a data management system is described herein. Typical unit test frameworks validate functions. However, the improved unit test framework validates the underlying data. For example, after each step of a data transformation process implemented by the data management system, the data management system can execute a data unit test that loads data sets into memory, checks a set of preconditions, and applies unit test logic to the loaded data sets. In some embodiments, the data management system executes the data unit tests asynchronously with the data transformation processes and therefore do not interfere with the data transformation processes. Rather, the data management system generates and transmits a notification when any step of the data transformation process fails a particular data unit test.

Abstract: An apparatus for decoding and rendering or otherwise processing, manipulating, or outputting a variety of digital data is provided. The apparatus may include a browser engine configured to receive first encoded digital data. The apparatus may include a browser extension configured to intercept the first encoded digital data transmitted in response to a first request from the browser engine. The browser extension may be configured to intercept a first decoder written in a platform-independent syntax for decoding at least a portion of the first encoded digital data. The browser extension may be configured to point the browser engine to one or more local storage locations of the first encoded digital data and the first decoder.

Abstract: Systems and methods for gateway communications between non-distributed ledger systems and distributed ledger systems are disclosed. In one embodiment, a method for gateway communication may include: (1) receiving, at a listener comprising a computer processor and from an application in a first domain, a message for a destination distributed ledger system that is one of a plurality of distributed ledger systems in a second domain; (2) generating, by the listener, a payload for the message, wherein the payload identifies the destination distributed ledger system for the message; and (3) encrypting, by the listener, the payload and passing a payload message comprising the encrypted payload to one of a plurality of adapters in a gateway communication service using an API. The adapter may receive the payload message, decrypt the payload, identifies the destination distributed ledger system, and communicate the decrypted payload to the destination distributed ledger system.

Abstract: In some aspects, an edge-based data collection system discovers, collects, processes, and forwards data in an observability pipeline system. In some implementations, an edge agent of the observability pipeline system runs on a computer node. The edge agent identifies processes running on the computer node; identifies files on the computer node that the processes have opened for writing; accesses log discovery parameters of the observability pipeline system; selects a plurality of files from the identified files according to the log discovery parameters; generates a list of discovered log files that includes a path and a name for each of the plurality of files; adds the list of discovered log files to a list of monitored log files to be monitored by the observability pipeline system; and then monitors the plurality of files to generate input for the observability pipeline system.

Abstract: In some implementations, a user interface for an application is displayed using a web browser instance on a client device. An input is received to present data on the user interface in a particular view. In response to the input, a first web worker thread corresponding to the web browser instance obtains data from a server, and executes first library routines to store the data in local storage at the client device. A second web worker thread, which corresponds to the web browser instance and the user interface, accesses the data from the local storage by using one or more second library routines, and processes the data to convert to a presentation format corresponding to the particular view. The second web worker thread stores the processed data in the local storage by using one or more third library routines, and provides the processed data for display on the user interface.

Abstract: A device having a Graphics Processing Unit (GPU) may be configured to selectively run in a normal mode or a timing testing mode. In the timing testing mode the device is configured to disrupt timing of processing that takes place on the GPU while running an application with the GPU and test the application for errors in device hardware component and/or software component synchronization while the device is running in the timing testing mode.

Abstract: Provided is a memory management system based on a non-uniform memory access, which includes: a first persistent memory disposed in a first node; a second persistent memory disposed in a second node physically distinguished from the first node; a first journaling process disposed in the first persistent memory; and a second journaling process disposed in the second persistent memory.

Abstract: Systems and methods for intercepting an operation requested by a user (e.g., print a document, cut text, copy an image, paste a hyperlink, embed an audio clip, save as a document in a new location, etc.) and performing one or more actions based on the level of security associated with the originating program, originating file, intended program, and/or intended file for that operation. As such, the disclosed systems and methods may enable consistent data security to be applied to a particular data item regardless of the location of that data item or the operations performed on the same.

Abstract: Methods and apparatus for efficient data transfer within a user space network stack. Unlike prior art monolithic networking stacks, the exemplary networking stack architecture described hereinafter includes various components that span multiple domains (both in-kernel, and non-kernel). For example, unlike traditional “socket” based communication, disclosed embodiments can transfer data directly between the kernel and user space domains. Direct transfer reduces the per-byte and per-packet costs relative to socket based communication. A user space networking stack is disclosed that enables extensible, cross-platform-capable, user space control of the networking protocol stack functionality. The user space networking stack facilitates tighter integration between the protocol layers (including TLS) and the application or daemon. Exemplary systems can support multiple networking protocol stack instances (including an in-kernel traditional network stack).

Abstract: A method for providing a notification of applications that relate to a business process is provided. The method includes collecting identification data that relates to each of a plurality of applications connected to the at least one processor; storing the identification data; receiving a user request for information that includes a list of first applications that relate to the business process; analyzing the identification data to determine whether each respective one of the plurality of applications is to be included in the list; and generating the list based on a result of the analyzing and notifying the user of the list. The method also includes collecting network data that relates to each of the plurality of applications; storing the network data; and analyzing the network data to determine whether each respective one of the plurality of applications is to be added to the list.

Abstract: An agent selection component that selects and activates artificial agents for interacting with a user on behalf of various entity based on user interaction with a real environment. The agent selection component detects user interaction with the real environment, and evaluates the detected user interaction. Based on that evaluation, the agent selection component selects an artificial agent that acts for an entity from amongst multiple artificial agents that act for different entities (e.g., to answer questions, to place orders, to schedule, or the like). The agent selection component then causes the selected artificial agent to activate to interact with the entity. Thus, different interactions with a real environment may result in the agent selection component permitting the user to interface with artificial agents for different entities.

Abstract: Interworking between different layer two (L2) medias using network tunnels is provided by receiving, at a virtual gateway network element (GNE), a packet from a first L2 media type network for transmission to a second L2 media type network under the control of a different entities, wherein the first and second L2 media type networks are incompatible for direct packet transmission; removing, at the virtual GNE, ethernet encapsulation from the packet; re-encapsulating, at a network device located between the virtual GNE and a packet network, the packet in a multi-protocol label switching (MPLS) encapsulation; forwarding, over the packet network, the packet from the network device to a digital communication channel (DCC) associated with the second L2 media type network; replacing, at the DCC, the MPLS encapsulation with a link access protocol (LAP) encapsulation; and transmitting the packet encapsulated with the LAP encapsulation to the second L2 media type network.

Abstract: Example communication methods and example apparatuses are provided. One example method includes obtaining, by an adaptor client, a southbound interface protocol type, where the adaptor client manages two or more protocol adaptors. The adaptor client can then select a first protocol adaptor from the two or more protocol adaptors based on the southbound interface protocol type, where the first protocol adaptor implements a first protocol.

Abstract: Devices and techniques for non-blocking external device calls are described herein. Specifically, when a processor receives an instruction with a no-return indication from a thread for a device, the processor can increase a counter corresponding to the thread based on the no-return indication. The processor can then continue execution of the thread without waiting for a return value from the device. When a return value is received for the instruction, the processor can decrement the counter. While the counter is not zero, the processor prevents the thread from completing (exiting).

Abstract: A technology agnostic distributed publication subscription as a service (DPAAS) server-based environment is used as a wrapper for abstracting technology, protocol and/or vendor specific implementations of publication/subscriber functionality. Users may interface with the DPAAS system an application programming interface to connect and/or disconnect, subscribe and unsubscribe from different publication/subscriber services. Further, selection of underlined technology, protocol and/or other vendor specific solutions can be facilitated by leveraging a self-service onboarding portal. Applications register to the DPAAS, along with any underlying technologies used by the source service, so that communications may be seamless to the end user of the subscribing application. In doing so, the DPAAS solution is vendor, protocol, and/or technology independent.

Abstract: The invention relates to a method for compiling code adapted for secondary offloads in a graphics processing unit (GPU). The method, performed by a processing unit, includes: reconstructing execution codes in a first kernel into a second kernel. The second kernel includes an operation table including entries, and computation codes. The computation codes include a portion of the execution codes, and synchronization hooks, and each synchronization hook includes information indicating one entry of the operation table. An order of the portion of the execution codes and the synchronization hooks in the computation codes matches an order of the execution codes in the first kernel, thereby enabling a compute unit (CU) in the GPU to execute the computation codes, and an engine in the GPU to instruct a component inside or outside of the GPU to complete a designated operation in accordance with content of each entry in the operation table.

Abstract: Systems and method are provided for receiving, at a virtualized management operation engine (VMOE) of a distributed system, a management task request from a client for one or more services. The management task request may be to an orchestrator of a first application server of one or more application servers of the distributed system. The first application server having the orchestrator may determine whether the management task request is valid. The first application server having the orchestrator may generate a request message based on the valid management task request. A message bus may broadcast the generated request message to one or more agents of the one or more application servers of the distributed system. The one or more agents may execute an operation based on the request message when the request message is intended for the one or more agents.

Abstract: An apparatus includes a removable media interface circuit and a processor. The removable media interface may be configured to read and write files to a non-volatile storage medium. The processor may be configured to generate encoded data and manage file operations involving storing the encoded data on the non-volatile storage medium to minimize a number of file fragments.

Abstract: A method apparatus for resource allocation and related products are provided. The method includes the following. An operating system acquires at least one system event associated with a target application in response to detecting that the target application is running in a foreground of the mobile terminal. The operating system identifies an internal running scenario of the target application according to the at least one system event. The operating system acquires a strategy for performance improvement corresponding to the internal running scenario. The operating system adjusts allocation of system resources of the target application according to the strategy for performance improvement.

Abstract: A database system supporting persistent queries, using an enhanced persistent query service and various data sources. On receiving a request to create a persistent query from a client software application, the persistent query service: creates a query virtual table; parses the persistent query; creates a plurality of intermediate virtual tables; establishes listeners for the query virtual table; creates a plurality of data source virtual tables; causes the plurality of data source virtual tables to retrieve initial data from data sources; and propagates data via intermediate virtual tables to the persistent query virtual table.

Abstract: A query referencing a function associated with a remote software component is received by a network-based data warehouse system. Temporary security credentials corresponding to a role at a cloud computing service platform are obtained. The role has permission to send calls to a web endpoint corresponding to the remote software component. A request comprising input data and electronically signed using the temporary security credentials is sent to a web Application Programming Interface (API) management system of the cloud computing service platform. The request, when received by the web API management system, causes the web API management system to invoke external functionality provided by the remote software component at the web endpoint with respect to the input data. A response comprising a result of invoking the external functionality is received from the web API management system, and the result data is processed according to the query.

Abstract: A database system supporting persistent queries, using an enhanced persistent query service and various data sources. On receiving a request to create a persistent query from a client software application, the persistent query service: creates a query virtual table; parses the persistent query; creates a plurality of intermediate virtual tables; establishes listeners for the query virtual table; creates a plurality of data source virtual tables; causes the plurality of data source virtual tables to retrieve initial data from data sources; and propagates data via intermediate virtual tables to the persistent query virtual table.

Abstract: Services can be managed across containers. A management service can obtain or compile configuration information for containerized applications and containerized services that are hosted on a computing device. The configuration information can define how a containerized application is dependent on a containerized service. Using the configuration information, the management service can establish data paths between containers to enable container services running in the containers to perform cross-container communications by which a containerized application in one container can access a containerized service in another container. The management service may also enable a container service to perform communications by which a containerized application can access services provided by the host operating system.

Abstract: A method of deploying a virtual network function of a network service in a data center having a cloud management server running a cloud computing management software to provision virtual infrastructure resources of the data center to at least one tenant, includes generating at least first and second API calls to the cloud computing management software in response to external commands received at the data center to deploy a virtual network function, and executing at least the first and second API calls by the cloud computing management software to deploy the virtual network function. The cloud computing management software creates at least one virtual machine by executing the first API call and at least one virtual disk by executing the second API call.

Abstract: In an example embodiment, a data model infrastructure is implemented as a service rather than as an always-running server. Specifically, one of the technical issues with past implementations is that the data models are deployed onto a server that is intended to be “always running”, even if there are no requests to the server. This utilizes memory and processing power. While it may be useful to have an always running server for commonly used applications, for applications that are infrequently used (e.g., 10 times a month), it can mean that memory and processing power is wasted. Thus, by implementing the data model infrastructure as a service rather than an always-running server, the service can be launched only when actually needed, saving both memory and processing power.

Abstract: The technology disclosed relates to maintaining a cache of effective properties in an identity management system employing a graph. In particular, it relates to handling vertex/edge and/or graph topology updates in accordance with update notification requirements configured from a schema and, in conjunction with detecting updating of vertex/edge attributes and/or graph topology, recalculating effective attributes in accordance with the configured notification requirements.

Abstract: A computing device may be configured to provide a shared N4 termination function for a plurality of Session Management Function (SMF) instances of an SMF Set and interface with one or more User Plane Function (UPF) instances. To achieve this, the computing device may operate a server configured with a Remote Procedure Call (RPC)-based protocol (e.g. gRPC Remote Procedure Call or “gRPC”) for a Packet Forwarding Control Protocol (PFCP) message service for serving a plurality of client stubs respectively situated at the plurality of SMF instances. The server may perform processing associated with each client stub by receiving an RPC-based protocol message which indicates a request, unpacking the RPC-based protocol message for obtaining PFCP message request data, and operating or interfacing with a PFCP message handling client for sending a PFCP message having the PFCP message request data to an indicated one of the UPF instances.

Abstract: A method and an apparatus for running an applet are provided. The method includes: reading, in response to issuing an applet to a host application, a configuration file to determine a dynamic library referenced by the applet; loading the dynamic library referenced by the applet from a sandbox directory and/or an applet platform; and running, in response to a running instruction of the applet, a script code of the applet to call the dynamic library referenced by the applet. When the applet is run, the dynamic library referenced by the applet is loaded from the sandbox directory and/or the applet platform, and the dynamic library is no longer integrated into the applet package.

Abstract: An embodiment of the present disclosure includes an RPC architecture that includes a central manager gateway with a client-facing side that allows for client access via web services protocols such as SOAP and REST. The central manager gateway further includes a server-facing side that can communicate with a plurality of network elements, with each network element implementing a common IDL architecture and RPC manager instance. Each of the network elements, and in particular their RPC manager instance, may communicate with other RPC manager instances to ‘learn’ the network topology for the system and maintain a topology database for purposes of exposing a naming service, e.g., a CORBA naming service. The network elements may elect one master element while the others remain as slaves. The central manager gateway may automatically locate the master network element and forward client requests to the same for servicing.

Abstract: A master database module is on a master computer node. Slave database modules are on slave computer nodes connected to the master computer node via a network. A distributed database includes executable code executed by processors on the master computer node and the slave computer nodes to receive a distributed database query at the master computer node. A query execution plan is prepared at the master computer node. The query execution plan is deployed on the slave computer nodes. The query execution plan is executed on the slave computer nodes. The slave computer nodes each perform a single node query execution analysis to selectively produce a query cancellation command. The query cancellation command is propagated to the master computer node and the slave computer nodes. The query execution plan is cancelled on the master computer node and the slave computer nodes.

Abstract: In one implementation, a method for automatically generating a security policy for a controller includes receiving, by a security policy generation system and from a controller development environment, code for a device controller; selecting middleware that enforces a security policy; analyzing the code for the device controller; based at least in part on the analyzing, automatically generating the security policy; and providing the selected middleware along with the generated security policy.

Abstract: An apparatus for decoding and rendering or otherwise processing, manipulating, or outputting a variety of digital data is provided. The apparatus may include a browser engine configured to receive first encoded digital data. The apparatus may include a browser extension configured to intercept the first encoded digital data transmitted in response to a first request from the browser engine. The browser extension may be configured to intercept a first decoder written in a platform-independent syntax for decoding at least a portion of the first encoded digital data. The browser extension may be configured to point the browser engine to one or more local storage locations of the first encoded digital data and the first decoder.

Abstract: The present disclosure provides information interaction methods, apparatuses, systems, and devices, storage media and computer programs. The method includes: publishing a remote procedure call, RPC, message to a service-side; and in response to detecting that a response message with respect to the RPC message exists on the service-side, obtaining the response message, where the response message is published to the service-side by a called device indicated in the RPC message, and the response message is acquired by the called device processing the RPC message after detecting the RPC message at the service-side.

Abstract: A process for annotating a video in real-time on a mobile device. The process may include creating one or more markers, allowing a user of the mobile device to annotate the video while one or more users within a group of users are annotating the streaming video in real-time. The process may include receiving a selection from the user for which he or she seeks to annotate within the video. The process further includes displaying a text box for a frame or range of frames selected by the user seeks for annotation, and receiving a submitted text box from the user and propagating the annotations within the submitted text box to one or more users within the group in real-time.

Abstract: Techniques for partitioning data flow operations between execution on a compute instance and an attached accelerator instance are described. A set of operations supported by the accelerator is obtained. A set of operations associated with the data flow is obtained. An operation in the set of operations associated with the data flow is identified based on the set of operations supported by the accelerator. The accelerator executes the first operation.