Intelligent memory block replacement

Abstract

A framework for intelligent memory replacement of loaded data blocks by requested data blocks is provided. For example, various factors are taken into account to optimize the selection of loaded data blocks to be discarded from the memory, in favor of the requested data blocks to be loaded into the memory. In some implementations, correlations between the requested data blocks and the loaded data blocks are used to determine which of the loaded data blocks may become candidates to be discarded from memory.

Description

technical field

[0001] The present disclosure relates generally to databases, and more specifically, to replacing or exchanging data in memory. Background technique

[0002] In-memory databases were developed to take full advantage of modern hardware to improve performance. By maintaining all relevant data in main memory (eg, random access memory (RAM)), data processing operations can be significantly accelerated.

[0003] More and more enterprises and institutions have applied in-memory databases to process and analyze big data (bigdata). Since all relevant data is kept in memory, in-memory databases undoubtedly surpass traditional databases in terms of computing speed. However, keeping all the data in memory can also be problematic when the available memory size (RAM) is less than the total data size of the database. For example, in this case less than the entire database is loaded into memory at a time. Memory replacement techniques are generally used in such situatio...

Examples

Embodiment Construction

[0014] According to example implementations, various techniques and systems for replacing or exchanging data in memory are disclosed. For example, refer to figure 1 , the techniques and systems describe exchanging, replacing, exchanging (etc.) a data column 102 currently in memory 104 with a requested data column 106 that is retrieved (e.g., from storage device 108) Load into data column 102 position. For ease of discussion, this disclosure describes various techniques with reference to the illustrated computing environment (system) 100 . However, the description is also applicable to other computing environments, networks, other forms and types of computing models, and the like.

[0015] Traditional memory replacement algorithms can be too complex to be useful, or too simple with too few considerations. For example, factors such as the amount of access to a data block 102 , the memory size of a data block 102 , and dependencies between different data blocks 102 may have a ...

technical field

[0001] The present disclosure relates generally to databases, and more specifically, to replacing or exchanging data in memory. Background technique

[0002] In-memory databases were developed to take full advantage of modern hardware to improve performance. By maintaining all relevant data in main memory (eg, random access memory (RAM)), data processing operations can be significantly accelerated.

[0003] More and more enterprises and institutions have applied in-memory databases to process and analyze big data (bigdata). Since all relevant data is kept in memory, in-memory databases undoubtedly surpass traditional databases in terms of computing speed. However, keeping all the data in memory can also be problematic when the available memory size (RAM) is less than the total data size of the database. For example, in this case less than the entire database is loaded into memory at a time. Memory replacement techniques are generally used in such situatio...

Examples

Embodiment Construction

[0014] According to example implementations, various techniques and systems for replacing or exchanging data in memory are disclosed. For example, refer to figure 1 , the techniques and systems describe exchanging, replacing, exchanging (etc.) a data column 102 currently in memory 104 with a requested data column 106 that is retrieved (e.g., from storage device 108) Load into data column 102 position. For ease of discussion, this disclosure describes various techniques with reference to the illustrated computing environment (system) 100 . However, the description is also applicable to other computing environments, networks, other forms and types of computing models, and the like.

[0015] Traditional memory replacement algorithms can be too complex to be useful, or too simple with too few considerations. For example, factors such as the amount of access to a data block 102 , the memory size of a data block 102 , and dependencies between different data blocks 102 may have a ...

Images