Method for reading data by utilizing multilevel cache and multilevel cache apparatus for caching data

Abstract

The invention provides a method for reading data by utilizing a multilevel cache and a multilevel cache apparatus for caching the data. The data loss risk can be lowered; the time for obtaining cached data can be shortened; the data processing time in a calculation process can be shortened; and the program throughput can be increased. According to the method for reading the data by utilizing the multilevel cache, the multilevel cache comprises N cache regions; the N cache regions have different priorities; and N is an integer greater than 1. The method is characterized by comprising the steps of trying to read required data from the N cache regions according to a sequence of priorities from high to low; and when the data is successfully read in one cache region, returning the data and synchronously caching the data to the cache regions with the priorities higher than the priority of the cache region in which the data is read.

Description

technical field [0001] The invention relates to the technical field of computers and computer software, in particular to a method for reading data by using a multi-level cache and a multi-level cache device for caching data. Background technique [0002] In the process of Storm streaming calculation, subsequent calculations need to rely on historical calculation results. For example, when we calculate the real-time PV (abbreviation of page view, that is, page views) value of the website, each PV value calculation needs to rely on The historical PV value. At this time, the historical calculation results (ie: intermediate results) need to be cached. The existing method is to directly cache the calculation results in a single database, such as Redis, HBase, etc. When you need to use the cached intermediate results for calculations, read them directly from the database. [0003] Take intermediate results cached in Redis as an example (refer to figure 1 ), read cached data dire...

AI Technical Summary

Problems solved by technology

[0012] 1. Both Redis and HBase are single-database storage, and single-database storage may have the risk of losing cached data, and there may be situations where the cached data cannot be obtained if the database is not connected;

[0013] 2. Both Redis and HBase need to obtain the stored cache value through the network. Obtaining the stored cache value through the network depends on the network condition, which is more limited, and the online cache takes a long time

Images