US20110138050A1 | 2011-06-09 |
Claims Following is the claim for this invention: - 1. In this invention we provide a service which is invoked before making any kind of database transactions in the cloud. Each transaction in the cloud might be updating the row object in database or an object in object store or a value object in key- value store, and in addition to this there will be active object like a user account object triggering the transaction along with auxiliary objects like credit card object which might result in a purchase of a new product object and hence updating the database for the same. Now all these objects from user account to credit card to product have a relationship with each other. Hence we will use a fast, reliable, high performance graph database optimized for storing such n number of relationships and querying the graph with minimum possible latency. Also we support all different kinds of queries for graph database. Here as part of transactional service we provide recommendation engine to provide personalized and relevant recommendations based on past transactions. Like in the example of user account object triggering the purchase of a product object with the help of credit card object, we recommend different product objects to the user from his purchase history. All recommendations are forwarded to a log service which appends all recommendations with timestamp. Later user can revisit the logs and analyse if any recommendation must be considered for a transaction. Also as part of transactional service we validate all transactions and raise timestamped alarms or events to the log service with transaction details if any invalid transaction was tried by the client. Let us suppose we have a user object with Social Security Number as the unique attribute which carried out a purchase for a product object with the help of a credit card object and we have stored the relationships among these objects in our graph database .One example of validation use case would be if another user object with a different unique Social Security Number attribute tries to carry out a purchase with the same credit card object then that transaction will be blocked by our service. After validating the transaction and providing any recommendations if possible the transactional service does the database transaction. Also users based on the application can decide if they want recommendations or validation or both for all their application transactions. User can register the application with the transactional service in cloud with a unique key or application identifier. Application will forward the transaction request to the transactional service in cloud by appending the unique key or application identifier. The above novel technique of providing recommendations and validations for each transaction in the cloud is the claim for this invention. |
In this invention we provide a service which is invoked before making any kind of database transactions in the cloud. Each transaction in the cloud might be updating the row object in database or an object in object store or a value object in key- value store, and in addition to this there will be active object like a user account object triggering the transaction along with auxiliary objects like credit card object which might result in a purchase of a new product object and hence updating the database for the same. Now all these objects from user account to credit card to product have a relationship with each other. Hence we will use a fast, reliable, high performance graph database optimized for storing such n number of relationships and querying the graph with minimum possible latency. Also we support all different kinds of queries for graph database. Here as part of transactional service we provide recommendation engine to provide personalized and relevant recommendations based on past transactions. Like in the example of user account object triggering the purchase of a product object with the help of credit card object, we recommend different product objects to the user from his purchase history. All recommendations are forwarded to a log service which appends all recommendations with timestamp. Later user can revisit the logs and analyse if any recommendation must be considered for a transaction. Also as part of transactional service we validate all transactions and raise timestamped alarms or events to the log service with transaction details if any invalid transaction was tried by the client. Let us suppose we have a user object with Social Security Number as the unique attribute which carried out a purchase for a product object with the help of a credit card object and we have stored the relationships among these objects in our graph database .One example of validation use case would be if another user object with a different unique Social Security Number attribute tries to carry out a purchase with the same credit card object then that transaction will be blocked by our service. After validating the transaction and providing any recommendations if possible the transactional service does the database transaction. Also users based on the application can decide if they want recommendations or validation or both for all their application transactions. User can register the application with the transactional service in cloud with a unique key or application identifier.
Application will forward the transaction request to the transactional service in cloud by appending the unique key or application identifier.
Next Patent: KIT AND METHOD FOR CALIBRATING LARGE VOLUME 3D IMAGING SYSTEMS