System and method for operating a big-data platform
First Claim
Patent Images
1. A method for operating a big-data platform comprising:
- at a data analysis platform, receiving discrete client data;
storing the client data in a network accessible distributed storage system that includes;
storing the client data in a real-time storage system in a row format;
merging the client data into a columnar-based distributed archive storage system;
identifying a merge status for the client data merged into the archive storage system, wherein the merge status indicates a redundancy of client data between the real-time storage system and the archive storage system;
receiving a data query through a query interface; and
processing the data query by selectively interfacing with the client data from the real-time storage system and archive storage system, according to a data mapping and reduction process, wherein the real-time storage system and the archive storage system are different, wherein processing the data query comprises;
(i) converting the single data query from a relational database-type query format to a first converted query format compatible with the real-time storage system,(ii) converting the single data query from the relational database-type query format to a second converted query format compatible with the archive storage system,(iii) cooperatively querying the real-time storage system and the archive storage system by distributing, in parallel, the first converted query over the real-time storage system and the second converted query over the archive storage system,(iv) using the merge status and timestamps of the client data in the real-time storage system and the archive storage system to skip client data from either the real-time storage system or the archive storage system if the skipped data is accounted for in the other of the real-time storage system or the archive storage system, and(v) retrieving a single cohesive query result that incorporates real-time data and archive data returned from the first converted query and the second converted query, respectively,wherein merging the client data into a columnar-based distributed archive storage system comprises storing the client data in the archive storage system in a columnar format, andwherein interfacing with the client data from the archive storage system comprises;
converting, by using a query processing cluster, at least a portion of the data query to the mapping process and the reduction process; and
executing the mapping process and the reduction process by using the query processing cluster.
1 Assignment
0 Petitions
Accused Products
Abstract
A system and method for operating a big-data platform that includes at a data analysis platform, receiving discrete client data; storing the client data in a network accessible distributed storage system that includes: storing the client data in a real-time storage system; and merging the client data into a columnar-based distributed archive storage system; receiving a data query request through a query interface; and selectively interfacing with the client data from the real-time storage system and archive storage system according to the query.
21 Citations
24 Claims
-
1. A method for operating a big-data platform comprising:
-
at a data analysis platform, receiving discrete client data; storing the client data in a network accessible distributed storage system that includes; storing the client data in a real-time storage system in a row format; merging the client data into a columnar-based distributed archive storage system; identifying a merge status for the client data merged into the archive storage system, wherein the merge status indicates a redundancy of client data between the real-time storage system and the archive storage system; receiving a data query through a query interface; and processing the data query by selectively interfacing with the client data from the real-time storage system and archive storage system, according to a data mapping and reduction process, wherein the real-time storage system and the archive storage system are different, wherein processing the data query comprises; (i) converting the single data query from a relational database-type query format to a first converted query format compatible with the real-time storage system, (ii) converting the single data query from the relational database-type query format to a second converted query format compatible with the archive storage system, (iii) cooperatively querying the real-time storage system and the archive storage system by distributing, in parallel, the first converted query over the real-time storage system and the second converted query over the archive storage system, (iv) using the merge status and timestamps of the client data in the real-time storage system and the archive storage system to skip client data from either the real-time storage system or the archive storage system if the skipped data is accounted for in the other of the real-time storage system or the archive storage system, and (v) retrieving a single cohesive query result that incorporates real-time data and archive data returned from the first converted query and the second converted query, respectively, wherein merging the client data into a columnar-based distributed archive storage system comprises storing the client data in the archive storage system in a columnar format, and wherein interfacing with the client data from the archive storage system comprises; converting, by using a query processing cluster, at least a portion of the data query to the mapping process and the reduction process; and executing the mapping process and the reduction process by using the query processing cluster. - View Dependent Claims (2, 3, 4, 5, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23)
-
-
6. A method for operating a big-data platform comprising:
-
at a client data agent, collecting discrete client data and transmitting the discrete client data to the data analysis platform, wherein the client data agent is integrated into an event channel from which client data is collected, wherein the event channel is selected from a list comprising syslog, a relational database, cloud data, and sensor data; and at a data analysis platform, receiving the discrete client data; storing the client data in a network accessible distributed storage system that includes; storing the client data in a real-time storage system in a row format; merging the client data into a columnar-based distributed archive storage system, wherein merging the client data into a columnar-based distributed archive storage system comprises storing the client data in the archive storage system in a columnar format; identifying a merge status for the client data merged into the archive storage system, wherein the merge status indicates a redundancy of client data between the real-time storage system and the archive storage system; receiving a data query through a query interface; and processing the data query by selectively interfacing with the client data from the real-time storage system and archive storage system, according to a data mapping and reduction process, wherein processing the data query comprises cooperatively querying the real-time storage system and the archive storage system and distributing the data query over the real-time storage system and the archive storage system to retrieve a single cohesive query result, using the merge status and timestamps of the client data in the real-time storage system and the archive storage system to skip client data from either the real-time storage system or the archive storage system if the skipped data is accounted for in the other of the real-time storage system or the archive storage system, and wherein interfacing with the client data from the archive storage system comprises; converting, by using a query processing cluster, at least a portion of the data query to the mapping process and the reduction process; and executing the mapping process and the reduction process by using the query processing cluster.
-
-
24. A method comprising:
- at a multi-tenant data analysis platform;
receiving discrete client data, the client data being associated with a user account of the multi-tenant data analysis platform through a unique identifier; storing the client data in a network accessible distributed storage system that includes a real-time storage system and a columnar-based distributed archive storage system, the storing of the client data comprising; storing the client data in the real-time storage system in a row format; merging the client data into the archive storage system in a columnar format, the client data merged into the archive data storage system being isolated according to the user account associated with the client data; identifying a merge status for the client data merged into the archive storage system, wherein the merge status indicates a redundancy of client data between the real-time storage system and the archive storage system; receiving a data query through a query interface; and processing the data query by selectively interfacing with the client data from the real-time storage system and archive storage system, wherein processing the data query comprises; converting the data query from a relational database-type query format to a first converted query format compatible with the real-time storage system and a second converted query format compatible with the archive system, wherein the converting the data query comprises converting, by using a query processing cluster, the data query to a MapReduce mapping process and a MapReduce reduction process, cooperatively querying the real-time storage system and the archive storage system by distributing, in parallel, the first converted data query over the real-time storage system and the second converted query over archive storage system, using the merge status and timestamps of the client data in the real-time storage system and the archive storage system to skip client data from either the real-time storage system or the archive storage system if the skipped data is accounted for in the other of the real-time storage system or the archive storage system, and retrieving a single cohesive query result based on results from both the real-time storage system and the archive storage system, wherein interfacing with the client data from the archive storage system comprises; executing the MapReduce mapping process and the MapReduce reduction process by using the query processing cluster, and wherein the query processing cluster includes a cluster that is constructed to execute MapReduce processes.
- at a multi-tenant data analysis platform;
Specification