Data supplying system and method, and terminal

Abstract

The invention discloses a data supplying system and method, and a terminal. The data supplying system comprises at least one data cache, at least one data query module and at least one data routing module, wherein the data cache is used for storing various pieces of data; the data query module is suitable for receiving the query request of a demand side and carrying out matching query on various pieces of data in the at least one data cache according to the query request to obtain a first query result and feed the first query result back to the demand side; and the data routing module is suitable for calling a corresponding supplier server when a first query result indicates that data matched with the query request is not inquired, and loads corresponding data information to the at least one data cache according to a second query result fed back from the supplier server. Through the technical scheme of the invention, the performance of the data supplying system can be improved.

Description

technical field [0001] The present invention relates to the technical field of data processing, in particular to a data supply system and method, and a terminal. Background technique [0002] In the process of data circulation, the online real-time data transaction system can provide continuous and on-demand supply capabilities. The improvement of system performance can increase the supply of the system. Multiple data sources can not only improve the stability of data supply, but also can Data fusion capability is required. For the online real-time data transaction system, not only must have single-point high performance, but also must have scalability to improve the overall supply capacity of the system. For example, in real-time advertising bidding (Real Time Bidding, RTB), the entire process from the start of bidding to the completion of delivery needs to be completed within 100 milliseconds (ms). Since multiple steps and related parties are involved, the query and acqui...

AI Technical Summary

Problems solved by technology

[0004] However, the data supply system in the prior art has the following disadvantages: 1. The throughput is low and the QPS is low; since the same application service needs to perform multi-supplier routing services, it takes up a lot of system resources, resulting in low system processing capacity, and when the system encounters In the case of high front-end concurrency and high throughput, the system cannot support it, and the central processing unit (Central Processing Unit, CPU) usage rate remains high, which can easily lead to system downtime

2. Long response time; the original system structure requires front-end requests and back-end multi-supplier routing services at the same time, and the overall processing process is long, resulting in a long application response time and a relatively long response time in the entire system, which cannot meet the demand-side system response time Require

3. The ability to expand is weak; under the original system structure, when the system needs to improve the service capability when the system is expanded, it is difficult to expand in certain aspects according to the request of the demand side or multi-supplier routing, but only by increasing the overall server After the expansion, the expansion gain effect is not obvious, the system throughput and QPS increase non-linearly, and the system scalability is weak

4. The system reliability is poor; when the original system has problems with the external supplier system or the network fluctuates, the multi-supplier routing will slow down, resulting in the accumulation of system requests and weakening of the system processing capacity

When there is a problem with the external system, the system is prone to downtime and the system reliability is poor

5. Poor disaster tolerance; the system is deployed with a single-machine redis cache. When there is a problem with redis, the entire system will be unavailable, and it must wait for the redis cache to recover before it can be used.

When the stand-alone redis crashes, data is easily lost and difficult to recover, and the disaster tolerance of the entire system is poor

Images