Sap Se

Sap Ag

Peter Eberlein

Volker Driesen

Bare Said

Jens Odenheimer

Arne Harren

Markus Schmidt-Karaca

Martijn De Boer

Timm Falter

Martin Hartig

Baré Said

Florian Mueller

Juergen Sattler

Cheng Wang

Oliver Klemenz

Rainer Brendle

Joachim Gaffga

Guenter Pecht-Seibert

Horst Schaude

Hartwig Seifert

Helmut Prestel

Hilmar Demant

Andrey Engelko

Volker Wiechers

Thomas Vogt

Olaf Meincke

Stefan Rossmanith

Bjoern Mielenhausen

Frank Schertel

Klaus Kopecz

Markus A Peter

Guenter Zachmann

Andreas Poth

Eckhard Farrenkopf

Robert Getzner

Joerg Schmidt

Wolfgang Janzen

Johannes Bechtold

Michael Engler

Tobias Brandl

Lars Spielberg

Silke Storch

Alexander Rieder

André Wagner

Thierry Tung

Lukas Gross

Momchil Atanasov

Martina Galabova

Shabana Ansari

Ramesh B G

Holger Neu

Hai Anh Pham

Florian Tack

Heiko Ettelbrueck

Jurgen Sattler

Christoph Lübbe

Voker Wiechers

Various examples are directed to systems and methods for executing a computer-automated process using trained machine learning (ML) models. A computing system may access first event data describing a first event. The computing system may execute a first ML model to determine an ML characterization of the first event using the first event data. The computing system may also apply a first rule set to the first event data to generate a rule characterization of the first event. The computing system may determine an output characterization of the first event based at least in part on the rule characterization of the first event and determine to deactivate the first rule set based at least in part on the ML characterization of the first event.

Computer-automated processing with rule-supplemented machine learning

A computer-implemented method includes translating into a routing configuration, tenant-specific preferences for primary and secondary datacenter locations. A service mesh is set up for communication between services within and across the primary and secondary datacenter locations. Service persistencies with endpoints in datacenter locations are used to configure replication agents between the service persistencies. Using service endpoints, configuring Virtual Services that implement the service mesh. An Ingress Gateway is configured to route end user requests into the service mesh to a first service instance in the tenant-selected primary datacenter. According to the tenant-specific preferences, data replication is configured to copy data to redundant storage. Using endpoints of persistent storage replication agents for each service persistence in the tenant-selected primary datacenter, configuring persistent storage replication agents for each service persistence in the tenant-selected primary datacenter.

Tenant-specific disaster recovery

In an example embodiment, access to a data set in a data lake can be specified using several approaches, based on the metadata and information attached. The metadata may be replicated from the original data source of the underlying data, and additional metadata may be modeled and stored to construct linkage information between data types. This linkage information may be used to automatically grant access to users to additional objects that are linked to objects that the user has explicit access to.

Deriving and augmenting access control for data lakes

Implementations include a schema stack management system that enables zero-downtime during execution of maintenance procedures on application systems having schema stacks including one or more customer-provided schema extensions.

Schema stack management system for managing extensions through upgrades

Systems and methods include reception of a first service call from a first service at a second service, generation, at the second service, of a response to the first service call, determination, at the second service, of an amount of hardware resources consumed by the second service during generation of the response to the first service call, generation, at the second service, of a first resource consumption token identifying the amount of hardware resources consumed by the second service during generation of the response to the first service call, determination of a first identifier of a type of hardware on which the second service is executing, and return of the response, the first resource consumption token and the first identifier of the type of hardware to the first service.

Holistic measurement of microservice-based application resource consumption

In an example embodiment, a solution is provided that allows a user to submit a document. Information can be obtained from the document using optical character recognition (OCR) or other techniques. This information can then be used to identify one or more workflows that pertain to the document. The one or more workflows may be ranked using machine learning techniques and presented to the user. Once the user selects a desired workflow, the information obtained from the document can then be used to automatically complete at least a portion of the workflow, for example by prefilling one or more fields in a form.

Document initiated intelligent workflow

A request is received to create a graph database from one or more relational databases. For each relational database, data objects in the relational database are identified. For each data object, a graph data object corresponding to the data object is created. The graph data object is linked to the data object. A set of associated data objects in the relational database are determined, and for each associated data object, an associated graph data object is created if a graph data object corresponding to the data object does not exist. For each created graph data object, a graph data relation object is created that represents a relationship between the graph data object and the associated graph data object. Created graph data objects, associated graph data objects, and graph data relation objects are stored in the graph database. The graph database is provided to one or more applications.

Automatic creation and synchronization of graph database objects

In an implementation, a mesh assembler receives a request for a new cloud-computing application tenant, where the request includes an assembly profile. The mesh assembler calls a deployment manager for applications and services that need to be custom deployed according to the assembly profile. The mesh assembler calls a routing configurator, which looks up services with certain attributes and configures services with certain attributes into a service mesh.

User demand-specific service mesh assembly of multi-tenant, single-tenant, and extension services

In an implementation, a shutdown coordinator sends an instruction to a router to no longer assign new sessions to an application instance, where the application instance is to be shut down. The shutdown coordinator sends a request to the application instance to export associated sessions for handover to a new application instance. The shutdown coordinator receives an indication from the application instance that no further requests can be processed by the application instance since all associated sessions have been exported to an external session storage. The shutdown coordinator instructs the router to redirect requests to the application instance to the new application instance. The shutdown coordinator shuts down the application instance. The shutdown coordinator deletes remaining sessions of all associated sessions from the external session storage after a defined timeout period.

Coordinated shutdown of cloud computing applications with session handover

In an example embodiment, a process is introduced into a machine learned model where additional results are output by the machine learned model in addition to those results that would be obtained through use of the trained model itself. In some example embodiments, these additional results may be random or semi-random to introduce results that might otherwise not have been recommended by the machine learned model. By introducing such additional results in a controlled way, it becomes possible to reduce biases caused by a self-reinforcing feedback loop while still presenting users with accurate machine learned model results.

Sensitivity in supervised machine learning with experience data

Systems and methods include determination of first constraints on columns of database tables of a test tenant system based on metadata associated with the columns of database tables, determination of second constraints on the columns of database tables based on data stored in the columns of database tables of the test tenant system, identification of inconsistent data stored in columns of database tables of a plurality of tenant systems based on the first constraints and the second constraints, adding of records to the database tables of the test tenant system based on the identified inconsistent data, and execution of a test based on the database tables of the test tenant system including the added records.

Generation of inconsistent testing data

Implementations include querying metadata of data objects to define a sub-set of data objects, each data object in the sub-set of data objects including a generic text field and/or an attachment field, and, for each data object in the sub-set of data objects, processing historical data of a data object to identify a set of data types, the historical data stored within a field of a table of a database system, providing a recommendation for a first extension corresponding to a first data type, and receiving user input indicating acceptance of the recommendation for the first extension and, in response, automatically providing extension code that is executable to add a field extension to the table and to modify a UI of an application for input of values corresponding to the first data type, and executing the extension code to deploy the extension and to modify the UI.

Selective recommendation and deployment of extensions in low-code approach

Systems and methods include determination of a plurality of call map clusters, each call map cluster associated with a respective type of external call, two or more of a plurality of services, and a service call model modelling, for each of the two or more of the plurality of services, the issuance of outgoing service calls in response to received service calls. A bot is deployed proximate to each of the plurality of services, and each deployed bot is associated with a respective configuration to issue outgoing service calls in response to received service calls as modelled by the service call models for each of one or more services proximate to the deployed bot.

Landscape model verification system

Various examples are directed to systems and methods for configuring an application that uses a plurality of microservices. A distributed configuration deploy service (DCDS) may generate a first set of microservice configuration parameter values for a first microservice of the plurality of microservices based at least in part on a first microservice configuration model for the first microservice and at least in part on a second configuration model for a second microservice of the plurality of microservices. A configuration application programming interface (API) may determine that the first set of microservice configuration parameter values do not conflict with first live data at the first microservice. The configuration API may also lock the first microservice; apply the first set of microservice configuration parameter values to the first microservice; and release the locking of the first microservice.

Configuration for application using microservices

Computer-implemented systems and computer-implemented methods include the following. A request to train a machine-learning (ML) model is received at a training broker. Anonymized data for training the model is obtained by the training broker from each individual data source of a plurality of data sources. The anonymized data is accessed through a data science schema being provided by anonymization of sensitive information of production data from each individual data source. Access to the anonymized data is provided to a data vendor for training the ML model using the anonymized data.

Machine learning on distributed customer data while protecting privacy

Growing community of inventors

Malsch, Germany

Peter Eberlein