Abstract: System and methods are described for specifying, deploying, and using plugins during replication of a database. The method includes receiving a replication plugin service configuration specifying a plurality of plugins, getting a first database plugin specified by the replication plugin service configuration, getting one or more reader plugins specified by the replication plugin service configuration, deploying the first database plugin and the one or more reader plugins to a relay communicating with a source database, and getting data from the source database using the first database plugin and the one or more reader plugins.

Abstract: A first copy of a database object may be constructed prior to applying a designated database update request to a first database. The database object may be represented by a plurality of database table rows in the first database, and the request may update one or more of the plurality of rows. The request may be applied to the first database to update the one or more of the plurality of rows in the first database. A second copy of the database object may be constructed after applying the database update request to the first database. A message identifying a difference between the first copy of the database object and a second copy of the database object may be transmitted to a remote computing device.

Abstract: Techniques are disclosed relating to the distribution of database key permissions. A database system may distribute first permission information to a plurality of database nodes that identifies a distribution of key range permissions to ones of the plurality of database nodes. A given key range permission being distributed to a given database node may permit that database node to write records whose keys fall within a key range associated with the given key range permission. The database system may receive, from a first database node, a request for a first key range permission provisioned to a second database node. The database system may modify the first permission information to derive second permission information that provisions the first key range permission to the first database node instead of the second database node. The database system may distribute the second permission information to ones of the plurality of database nodes.

Abstract: Examples include a system and computer-implemented method to receive a notification from an application programming interface (API) of creation of a just in time (JIT) grant, the JIT grant defining a request for a user to be authorized to access a cluster according to a JIT policy; determine if access to the cluster by the user is authorized according to the JIT policy; grant access to the user to the cluster when access is authorized according to the JIT policy; and send a notification to the API that access by the user to the cluster is granted.

Abstract: In accordance with embodiments disclosed herein, there are provided mechanisms and methods for automating deployment of applications in a multi-tenant database environment. For example, in one embodiment, mechanisms include managing a plurality of machines operating as a machine farm within a datacenter by executing an agent provisioning script at a control hub, instructing the plurality of machines to download and instantiate a lightweight agent; pushing a plurality of URL (Uniform Resource Locator) references from the control hub to the instantiated lightweight agent on each of the plurality of machines specifying one or more applications to be provisioned and one or more dependencies for each of the applications; and loading, via the lightweight agent at each of the plurality of machines, the one or more applications and the one or more dependencies for each of the one or more applications into memory of each respective machine.

Abstract: In accordance with embodiments, there are provided mechanisms and methods for facilitating dynamic adjustment of web product-based performance factors in database systems according to one embodiment. In one embodiment and by way of example, a method includes identifying a web product and metadata associated with the web product, where the metadata reveals one or more performance factors capable of impacting performance of the web product. The method may further include generating one or more virtual toggles representing the one or more factors, and facilitating adjustment to the one or more virtual toggles to facilitate one or more changes to one or more metrics corresponding to the one or more performance factors, where the one or more changes to the one or more metrics reflect variations in the performance of the web product.

Abstract: Methods and systems for making effective use of system resources. A plurality of requests for access to a resource are received. Each request has an associated group of features. The group of features for each request is analyzed to collect observations about the plurality of requests. A function to predict an outcome of a subsequent request is generated based on the observations. Resources are allocated to service the subsequent request based on the function.

Abstract: In various embodiments, a data storage system maintains a data store for a plurality of groups of users, where the data store includes a first section that is accessible by a first group of users, but not a second group of users, via the data storage system. In response to receiving, from a first user of the first group of users, a request to share data with the second group of users, the data storage system sends a request to a second user of the second group of users. Subsequent to receiving an acceptance from the second user, the data is stored in a second object in a second section of the data store that is accessible by the second group of users, but not the first group of users, via the data storage system.

Abstract: Implementations are described for automatically storing, in a blockchain, metrics that relate to payments. In one implementation, permissions that identify a set of metrics are retrieved from a database. The permissions are stored by a tenant of a multi-tenant system to grant permission to the multi-tenant system to store the metrics for a second tenant. A metric is selected from the set of metrics, based on a determination that a value of the metric is to be updated responsive to a payment in a transaction between the first and second tenants where the payment has been successfully processed. The value of the metric is determined for the second tenant based on data relating to the payment. The value of the metric is stored, in the blockchain, with an identifier that uniquely identifies the second tenant in the blockchain.

Abstract: In a computing environment, a set of executing processes each having associated resources are provided. Aggregate resources for the computing environment include multiple different types of resources. A utilization level for each of the resources within the computing environment is evaluated to determine an unconsumed capacity for each of the resources below a utilization threshold. The utilization threshold is resource-dependent. An indication of at least a portion of unconsumed capacity for each of the resources below the utilization threshold is gathered. The unconsumed portion for each of the resources below the utilization threshold is exposed for consumption by other executing processes.

Abstract: Systems, methods, and apparatuses for implementing certificates of authenticity of digital twins transacted onto a blockchain using Distributed Ledger Technology (DLT) in conjunction with a cloud based computing environment are described herein.

Abstract: Embodiments described herein provide a provide a fully unsupervised model for text compression. Specifically, the unsupervised model is configured to identify an optimal deletion path for each input sequence of texts (e.g., a sentence) and words from the input sequence are gradually deleted along the deletion path. To identify the optimal deletion path, the unsupervised model may adopt a pretrained bidirectional language model (BERT) to score each candidate deletion based on the average perplexity of the resulting sentence and performs a simple greedy look-ahead tree search to select the best deletion for each step.

Abstract: A system and method for generating a response in a video grounded dialogue are provided. A video-grounded dialogue neural network language model receives video input and text input. The text input includes a dialogue history between the model and a human user and a current utterance by the user. Encoded video input is generated using video encoding layers. Encoded text input is generated using text encoding layers. The encoded video input and the encoded text input are concatenated in to a single input sequence. A generative pre-trained transformer model generates the response to the current utterance from the singe input sequence.

Abstract: Within one or more instances of a computing environment where an instance is a self-contained architecture to provide at least one database with corresponding search and file system. User information from the one or more instances of the computing environment is organized as zones. A zone is based on one or more characteristics of corresponding user information that are different than the instance to which the user information belongs. User information is selectively obfuscated prior to transmitting blocks of data including the obfuscated user information. The selective obfuscation is based on zone information for one or more zones to which the user information belongs.

Abstract: A method in a cloud-based system for implementing behavioral changes to flow engine software that executes user-specific flows for a plurality of users with little or no disruption to the performance of the user-specific flows is provided. The method comprises: receiving a request to load and execute a user-specific flow; identifying an API version identifier for the user-specific flow; loading the user-specific flow; determining whether a software change to the flow engine software should be applied when executing the user-specific flow based on the API version identifier; and executing the user-specific flow using the flow engine software with the software change when it is determined, based on the API version identifier, that the software change to the flow engine software should be applied.

Abstract: The embodiments describe a generative-discriminative (GeDi) language modeling for determining a next token in a text sequence. A class conditional language model and a positive control code determine a first class conditional probability for each token candidate. The class conditional language model and a negative control code determine a second class conditional probability for the each token candidate. A logarithmic probability difference between the first class conditional probability and the second class conditional probability is determined for each token candidate. An unconditional language model determines an unconditional probability for each token candidate. A combined probability is determined by combining the unconditional probability and the logarithmic probability difference for each token candidate. The next token is selected from the token candidates based on the combined probabilities of the token candidates.

Abstract: Disclosed techniques relate to storing a key cache within a secure enclave. In some embodiments, a computing system receives, from an application, a request to access a database, where the request is associated with a particular account. The computing system then accesses, using an identifier associated with the particular account, a key cache stored in a secure enclave of a memory of the computing system to determine at least one private key associated with the request, where the key cache stores private keys of a key management system (KMS) for a plurality of accounts. The computing system performs a cryptographic operation for accessing the database within the secure enclave using the at least one private key. In various embodiments, disclosed techniques may improve the security of cryptographic private keys cached for a plurality of tenants.

Abstract: An embodiment provided herein preprocesses the input samples to the classification neural network, e.g., by adding Gaussian noise to word/sentence representations to make the function of the neural network satisfy Lipschitz property such that a small change in the input does not cause much change to the output if the input sample is in-distribution. Method to induce properties in the feature representation of neural network such that for out-of-distribution examples the feature representation magnitude is either close to zero or the feature representation is orthogonal to all class representations. Method to generate examples that are structurally similar to in-domain and semantically out-of domain for use in out-of-domain classification training. Method to prune feature representation dimension to mitigate long tail error of unused dimension in out-of-domain classification. Using these techniques, the accuracy of both in-domain and out-of-distribution identification can be improved.

Abstract: System and methods are described for creating an instance of a database in a cloud computing environment. One method includes receiving a request to deploy an instance of a database in a cloud computing environment, the request including a database deployment model defining deployment of the database instance, building a deployment model decision tree based on the database deployment model, and initializing a cloud deployment strategy using the database deployment model for enabling one or more cloud service providers of the cloud computing environment to deploy the database instance. The method further includes provisioning cloud computing resources for the database instance according to the cloud deployment strategy, initializing sets of services for the database instance according to the deployment model decision tree, creating the database instance, and indicating the database instance as active in the cloud computing environment.

Abstract: Techniques and structures to facilitate instantiating a data model, including receiving a parameterized encoding of a data model including one or more scenarios, receiving data from a source, executing one or more operations associated with each of the one or more scenarios to transform the received data and creating a plurality of records for the data model based on the transformed data.