Updating cached database query results

• US 9,235,620 B2
• Filed: 08/14/2012
• Issued: 01/12/2016
• Est. Priority Date: 08/14/2012
• Status: Active Grant

• Current Assignee: Amadeus (Amadeus Global Travel Distribution S.A.)
• Original Assignee: Amadeus (Amadeus Global Travel Distribution S.A.)
• Inventors: Ciabrini, Damien, Legrand, Guillaume, Janin, Benoit, Isnardy, Luc, Maillot, Nicolas, Robelin, Charles-Antoine, Daniello, Rudy
• Primary Examiner(s): Richardson, James E
• Application Number:
  US13/585,286
• Publication Number:
  US 20140052750A1
• Time in Patent Office: 1,246 Days
• Field of Search: 707/948
• US Class Current: 1/1
• CPC Class Codes: G06F 16/24552 Database cache management
• CPC Class Codes: G06Q 10/02 Reservations, e.g. for tickets, services or events

Abstract

A data cache platform maintains pre-computed database query results computed by a computation platform based on data maintained in the computation platform and is configured to determine probabilities of the cached database query results being outdated, to automatically issue re-computation orders to the computation platform for updating cached database query results on the basis of the determined probabilities of the pre-computed database query results being outdated and to receive the updated pre-computed database query results as results of the re-computation orders. The probability determination depends on a probabilistic model and on the occurrence of asynchronous real-time events. The real-time events are indeterministic with regard to the expiration of the cached database query results and only have a probabilistic influence on the discrepancies between the cached database query results maintained in the data cache platform and presumed actual database query results.

Claims

1. A method of updating query results in each of a plurality of distributed application platforms using a data cache platform for maintaining a first cache of query results and a computation platform for computing the query results, each query result including a pre-computed price for a travel recommendation, the method comprising:

determining, by the data cache platform, probabilities of the pre-computed prices of the query results in the first cache being outdated based on a probabilistic model that estimates likelihoods that the pre-computed prices of the query results in the first cache are inaccurate based at least in part on meta-data attributes of the query results in the first cache;

automatically issuing, by the data cache platform, re-computation orders to the computation platform for updating the query results in the first cache on a basis of the determined probabilities of the pre-computed prices of the query results in the first cache being outdated, wherein the query results in the first cache having a higher probability of being outdated than others are ordered to be re-computed;

in response to receiving the re-computation orders, computing, by the computation platform, updated query results and transmitting the updated query results to the data cache platform and each of the application platforms, each of the application platforms maintaining a respective second cache of query results;

receiving, at the data cache platform, the updated query results as results of the re-computation orders;

enriching, by the data cache platform, the updated query results with the meta-data attributes; and

updating, by the data cache platform, the query results in the first cache with the enriched query results,

wherein the query results in each second cache are updated with the updated query results independently of the content of the second cache as a result of the re-computation orders issued by the data cache platform.

2. The method of claim 1, wherein the probabilistic model models a volatility of data maintained in the computation platform based on statistical historic data.

3. The method of claim 1, wherein the data cache platform, when determining the probabilities of the pre-computed prices of the query results in the first cache being outdated and issuing the re-computation orders, considers grids of the query results in the first cache corresponding to groups of adjacent sets of data maintained in the computation platform.

4. The method of claim 1, wherein the data cache platform issues the re-computation orders based on an amount of available computation resources at the computation platform.

5. The method of claim 1, wherein the computation platform maintains information on travel availability and fares, and the data cache platform maintains priced travel recommendations calculated from the information on the travel availability and the fares.

6. The method of claim 1 further comprising:

receiving the updated query results at each of the application platforms; and

updating the query results in the respective second cache of each of the application platforms with the received query results.

7. The method of claim 1 wherein:

determining the probabilities of the pre-computed prices of the query results in the first cache being outdated is further based on an occurrence of one or more asynchronous real-time events,

the one or more asynchronous real-time events are indeterministic with regard to accuracy of the query results in the first cache and only have a probabilistic influence on the likelihoods the query results in the first cache maintained by the data cache platform are inaccurate, and

the probabilities determined based on the probabilistic model are amended on the occurrence of the one or more asynchronous real-time events.

8. The method of claim 7 further comprising:

analyzing, at the data cache platform, whether the one or more asynchronous real-time events are represented in the probabilistic model;

for each of the one or more asynchronous real-time events which are not represented in the probabilistic model, issuing re-computation orders for corresponding query results in the first cache; and

for each of the one or more asynchronous real-time events which are represented in the probabilistic model, accumulating such real-time events over a certain period of time, comparing the actually occurred and accumulated real-time events with their representation in the probabilistic model and, if the actually occurred and accumulated real-time events deviate from their representation in the probabilistic model to a predetermined extent, issuing re-computation orders with respect to potentially affected query results in the first cache.

9. The method of claim 7 wherein the one or more asynchronous real-time events comprise flight fare changes, airplane seat availability changes, client flight ticket requests and/or flight cancellations.

10. A system for updating query results in each of a plurality of distributed application platforms, each query result including a pre-computed price, the system comprising:

at least one processor; and

at least one memory coupled to the at least one processor and including program code that, when executed by the at least one processor, causes the system to;

determine probabilities of the pre-computed prices of query results in a first cache being outdated based on a probabilistic model that estimates likelihoods that the pre-computed prices of the query results in the first cache are inaccurate based at least in part on meta-data attributes of the query results in the first cache;

automatically issue re-computation orders for updating the query results in the first cache on a basis of the determined probabilities of the pre-computed prices of the query results in the first cache being outdated, wherein the query results in the first cache having a higher probability of being outdated than others are ordered to be re-computed;

in response to receiving the re-computation orders, compute updated query results and transmit the updated query results to each of the application platforms, each of the application platforms maintaining a respective second cache of query results;

enrich the updated query results with the meta-data attributes; and

update the query results in the first cache with the enriched query results,

wherein the query results in each second cache are updated with the updated query results independently of the content of the second cache as a result of the re-computation orders issued by the system.

11. The system of claim 10, wherein the probabilistic model models a volatility of the pre-computed prices based on statistical historic data.

12. The system of claim 10 wherein the program code further causes the system to:

further determine the probabilities of the pre-computed prices of the query results in the first cache being outdated based on an occurrence of one or more asynchronous real-time events that are indeterministic with regard to accuracy of the query results in the first cache and only have a probabilistic influence on the estimated likelihoods of the pre-computed prices of the query results in the first cache being inaccurate;

amend the probabilities determined based on the probabilistic model on the occurrence of the one or more asynchronous real-time events;

analyze whether the one or more asynchronous real-time events are represented in the probabilistic model;

for each of the one or more asynchronous real-time events which are not represented in the probabilistic model, issue re-computation orders for corresponding query results in the first cache; and

for each of the one or more asynchronous real-time events which are represented in the probabilistic model, accumulate such real-time events over a certain period of time, compare the actually occurred and accumulated real-time events with their representation in the probabilistic model and, if the actually occurred and accumulated real-time events deviate from their representation in the probabilistic model to a predetermined extent, issue re-computation orders with respect to potentially affected query results in the first cache.

13. The system of claim 12, wherein the real-time events comprise flight fare changes, airplane seat availability changes, client flight ticket requests and/or flight cancellations.

14. The system of claim 10 wherein the program code further causes the system to:

consider grids of the cached query results corresponding to groups of adjacent sets of data when determining the probabilities of the pre-computed prices of the query results in the first cache being outdated and issuing the re-computation orders.

15. The system of claim 10 wherein the re-computation orders are issued to a computation platform, and the program code further causes the system to:

issue the re-computation orders based on an amount of available computation resources at the computation platform.

16. The system of claim 10 wherein the re-computation orders are issued to a computation platform, the system comprises a travel reservation system, the computation platform maintains information on travel availability and fares, and the system maintains priced travel recommendations calculated from the information on the travel availability and the fares.

17. A travel reservation system comprising a data cache platform maintaining a first cache of travel recommendations computed by a computation platform in communication with a plurality of distributed application platforms each maintaining a respective second cache of travel recommendations, each travel recommendation including a pre-computed price, the data cache platform comprising:

a first processor; and

a first memory coupled to the first processor and including first program code that, when executed by the first processor, causes the data cache platform to;

determine probabilities of the pre-computed prices of the travel recommendations in the first cache being outdated based on a probabilistic model that estimates likelihoods that the pre-computed prices of the travel recommendations in the first cache are inaccurate based at least in part on meta-data attributes of the travel recommendations in the first cache;

automatically issue re-computation orders to the computation platform for updating the travel recommendations in the first cache on a basis of the determined probabilities of the pre-computed prices of the travel recommendations in the first cache being outdated, wherein the travel recommendations in the first cache having a higher probability of being outdated than others are ordered to be re-computed;

receive updated query results as results of the re-computation orders;

enrich the updated query results with the meta-data attributes; and

update the travel recommendations in the first cache with the enriched query results;

the computation platform comprising;

a second processor; and

a second memory coupled to the second processor and including second program code that, when executed by the second processor, causes the computation platform to;

in response to receiving the re-computation orders, compute the updated query results and transmit the updated query results to the data cache platform and each of the application platforms,

wherein the travel recommendations in each second cache are updated with the updated query results independently of the content of the second cache as a result of the re-computation orders issued by the data cache platform.

18. The travel reservation system of claim 17 wherein the first program code further causes the data cache platform to:

further determine the probabilities of the pre-computed prices of the travel recommendations in the first cache being outdated based on an occurrence of one or more asynchronous real-time events that are indeterministic with regard to accuracy of the travel recommendations in the first cache and only have a probabilistic influence on the likelihoods the travel recommendations in the first cache are inaccurate;

amend the probabilities determined based on the probabilistic model on the occurrence of the one or more asynchronous real-time events;

analyze whether the one or more asynchronous real-time events are represented in the probabilistic model;

for each of the one or more asynchronous real-time events which are not represented in the probabilistic model, issue re-computation orders for corresponding travel recommendations in the first cache;

for each of the one or more asynchronous real-time events which are represented in the probabilistic model, accumulate such real-time events over a certain period of time, compare the actually occurred and accumulated real-time events with their representation in the probabilistic model and, if the actually occurred and accumulated real-time events deviate from their representation in the probabilistic model to a predetermined extent, issue re-computation orders with respect to potentially affected query results in the first cache.

19. The travel reservation system of claim 17, further comprising:

the application platforms, wherein each of the application platforms comprises;

a respective third processor; and

a respective third memory coupled to the third processor and including respective third program code that, when executed by the third processor, causes the application platform to;

receive the updated query results; and

update the query results stored in the respective second cache with the received query results.

20. A computer readable storage medium having computer program instructions stored therein for maintaining cached query results including pre-computed prices, which when executed on a computer system cause the computer system to:

determine probabilities of the pre-computed prices of the query results in a first cache being outdated based on a probabilistic model that estimates likelihoods that the pre-computed prices of the query results in the first cache are inaccurate based at least in part on meta-data attributes of the query results in the first cache;

automatically issue re-computation orders for updating the query results in the first cache on a basis of the determined probabilities of the pre-computed prices of the query results in the first cache being outdated, wherein the query results in the first cache having a higher probability of being outdated than others are ordered to be re-computed;

in response to receiving the re-computation orders, compute updated query results and transmit the updated query results to a plurality of application platforms, each of the application platforms maintaining a respective second cache of query results;

enrich the updated query results with the meta-data attributes; and

update the query results in the first cache with the enriched query results,

wherein the query results in each second cache are updated with the updated query results independently of the content of the second cache as a result of the re-computation orders issued by the computer system.