METHOD AND SYSTEM FOR PROCESSING DATA QUERIES

• US 20130262443A1
• Filed: 03/30/2012
• Published: 10/03/2013
• Est. Priority Date: 03/30/2012
• Status: Active Grant

• Current Assignee: British Telecommunications PLC (BT Group PLC), Emirates Telecommunications Corporation, Khalifa University Of Science, Technology & Research
• Original Assignee: British Telecommunications PLC (BT Group PLC), Emirates Telecommunications Corporation, Khalifa University of Science and Technology
• Inventors: Leida, Marcello, Afzal, Ali, Majeed, Basim, Taylor, Paul
• Granted Patent: US 9,639,575 B2
• Time in Patent Office: Days
• US Class Current: 707/722
• CPC Class Codes: G06F 16/2452 Query translation
• CPC Class Codes: G06F 16/24534 Query rewriting; Transformation

Abstract

The invention relates to a method and system that provide a high performance and extremely scalable triple store within the Resource Description Framework (or alternative data models), with optimized query execution. An embodiment of the invention provides a data storage and analysis system to support scalable monitoring and analysis of business processes along multiple configurable perspectives and levels of granularity. This embodiment analyses data from processes that have been already executed and from ongoing processes, as a continuous flow of information. This embodiment provides defining and monitoring processes based on no initial domain knowledge about the process and such that the process will be built only from the incoming flow of information. Another embodiment of the invention provides a grid infrastructure that allows storage of data across many grid nodes and distribution of the workload, avoiding the bottleneck represented by constantly querying a database.

Claims

1. A method of processing a query on a data set stored on a plurality of nodes, the method including the steps of:

dividing the query into one or more atoms which define individual queries on the data set;

calculating the execution cost of each atom in the query;

determining one or more query paths which set out an order in which one or more atoms are to be executed using said calculated execution costs and interdependence between said atoms;

determining a query execution plan which is a set of said query paths which can be executed in parallel;

executing said atoms on each of said nodes in accordance with said query execution plan; and

combining the results of each query path to produce a result set that is the answer to said query.

2. A method according to claim 1 wherein the data set stores the data as triples comprising a subject, a predicate and an object, wherein the subject and object are variables and the predicate describes the relationship between the subject and the object.

3. A method according to claim 2, further including the step of storing data into the data set.

4. A method according to claim 3, wherein the step of storing data includes the step of converting the data in each triple from character strings to numeric representations.

5. A method according to claim 4 wherein the numeric representations are sequence numbers.

6. A method according to claim 5 wherein the step of storing data includes the sub-steps of:

parsing data to be stored to determine a triple containing character strings;

comparing each of said character strings against a plurality of stored character strings and their associated sequence numbers;

if a character string matches a stored character string, using the sequence number associated with said stored character string to replace the character string in the triple;

if a character string does not match any stored character string, obtaining a new sequence number, replacing the character string in the triple with said new sequence number and storing said character string associated with said new sequence number; and

storing said triple.

7. A method according to claim 1 wherein, where a query path contains a plurality of atoms, when said step of executing is performed for an atom which is not a first atom in a query path, said atom is only executed on the results of previously executed atoms in the query path.

8. A method according to claim 7 wherein, where the size of results from an atom exceeds a predetermined threshold, a subset of the results from said atom are passed to the subsequent atom or atoms in the query path before the execution of said atom is completed.

9. A method according to claim 1 wherein the calculation of the execution cost for each atom is carried out by assessing the number of instances in said data set of one or more constants contained within said atom.

10. A method according to claim 2 wherein the calculation of the execution cost is carried out by determining, for each atom, a weight, which for an atom i(j, k, l) is

W(j_j,k_i,l_i)=min{wgt(j_s),wgt(k_p),wgt(lo)}

where wgt(x_pos) is the number of matching triples with value x in the subject, predicate or object position pos if x is a constant, or the total size of the data set if x is a variable, and s, p and o are the subject, predicate and object positions respectively.

11. A method according to claim 1 wherein a query path is determined by:

setting the atom with the lowest execution cost as the first atom in the path;

setting as a subsequent atom in the path an atom which is dependent on said first atom in that it shares one or more variables with said first atom; and

repeating until all atoms which are dependent on the first atom or any subsequent atoms have been set in the path.

12. A method according to claim 11 wherein, where there are a plurality of atoms which are dependent, the subsequent atoms are set in order of execution cost of those atoms.

13. A method according to claim 1 wherein a query path may contain atoms which can be executed in parallel to each other as well as atoms which can be executed sequentially.

14. A method according to claim 1 wherein the step of executing includes executing the first atom in a query path on each node only in respect of data stored on said node

15. A method according to claim 1 wherein the query includes a filter which is a limit on the value of one or more variables in the data set, and the step of dividing further includes assigning said filter to each of the atoms which relate to said variables.

16. A method according to claim 15 wherein if said filter is disjunctive, the method includes the further steps of:

determining whether it is possible to assign said filter as a whole to an atom which features all of the variables in the filter, and if so, assigning it to all such atoms;

if it is not possible to so assign said filter, dividing said filter into a plurality of conjunctive filters and assigning a query path to each of said conjunctive filters.

17. A method according to claim 16 further including the step of, when said filter cannot be assigned to any atom, applying said filter to the results of said step of combining.

18. A system for storing business process data, the system having a plurality of nodes which are connected to each other via a network, each node having a data storage and a data processing device, wherein:

the system stores business process data as a data set in a distributed fashion on the data storages of the plurality of nodes; and

the data processing device of each node is arranged to processing a query on the data set by;

dividing the query into one or more atoms which define individual queries on the data set;

calculating the execution cost of each atom in the query;

determining one or more query paths which set out an order in which one or more atoms are to be executed using said calculated execution costs and interdependence between said atoms;

determining a query execution plan which is a set of said query paths which can be executed in parallel;

executing said atoms on each of said nodes in accordance with said query execution plan;

combining the results of each query path to produce a result set that is the answer to said query.

19. A system according to claim 18 wherein the data set stores the data as triples comprising a subject, a predicate and an object, wherein the subject and object are variables and the predicate describes the relationship between the subject and the object and at least one of the data processing devices is arranged to converting the data in each triple from character strings to numeric representations prior to storing the data.

20. A system according to claim 19 wherein the data processing device is arranged to monitor business process messages and to

parse data to be stored to determine a triple containing character strings;

compare each of said character strings against a plurality of stored character strings and their associated sequence numbers;

if a character string matches a stored character string, use the sequence number associated with said stored character string to replace the character string in the triple;

if a character string does not match any stored character string, obtain a new sequence number, replacing the character string in the triple with said new sequence number and storing said character string associated with said new sequence number; and

store said triple.

21. A system according to claim 18 wherein, where a query path contains a plurality of atoms, the data processing device is arranged to execute atoms which are not a first atom in a query path, only on the results of previously executed atoms in the query path.

22. A system according to claim 21 wherein, where the size of results from an atom exceeds a predetermined threshold, the data processing device is arranged to execute a subsequent atom or atoms in the query path on a subset of the results from said atom before the execution of said atom is completed.

23. A system according to claim 18 wherein the data processing device calculates the execution cost for each atom by assessing the number of instances in said data set of one or more constants contained within said atom.

24. A system according to claim 18 wherein the data processing device determines a query path by:

setting the atom with the lowest execution cost as the first atom in the path;

setting as a subsequent atom in the path an atom which is dependent on said first atom in that it shares one or more variables with said first atom; and

repeating until all atoms which are dependent on the first atom or any subsequent atoms have been set in the path.

25. A system according to claim 18 wherein the data processing device on each node is arranged to execute the first atom in a query path on the respective node only in respect of data stored on said node.