Implementation method of double-layer lockless hash table with aging deletion based on collision

Abstract

The invention relates to the technical field of data processing, in particular to a collision-based double-layer lockless hash table implementation method with aging deletion. The collision-based double-layer lock-free hash table implementation method with aging deletion comprises the following steps: S1, setting a double-layer hash table; S2, carrying out lock-free operation; S3, carrying out an operation mode of a skip list and a double linked list; S4, on the basis of the step S3, executing FIFO control. The invention provides a collision-based double-layer lock-free hash table implementation method with aging deletion.

Description

technical field [0001] The invention relates to the technical field of data processing, in particular to a method for realizing a collision-based double-layer lock-free hash table with aging deletion. Background technique [0002] The structure of the general hash table is mostly like this: Create a hash table according to the given number of elements; select the hash algorithm, the candidate algorithm is crc md5 murmur.. etc.; install the candidate algorithm for the input key and keysize It is calculated to obtain a value, and then the value is calculated based on the given total number of elements to obtain the index value; jump to the index entry where the hash table is located, and perform addition, deletion, modification and query operations; for the same index value but different content The element situation is called a collision. For this situation, the linked list operation is performed, and the element is added to the end of the linked list. [0003] The above met...

AI Technical Summary

Problems solved by technology

However, the above two aging deletion methods have disadvantages. The disadvantage of using a timer for aging deletion: using a proprietary hardware timer or a third-party software timer for aging deletion will use additional memory resources, which is at least double the amount to be precise. memory resources

[0007] Lock-free operations in concurrent scenarios, deleting elements will face ABA problems. The root of the ABA problem is that cas cannot record the state of variables when modifying variables, such as modifying The number of times, whether this variable has been modified

In this way, when one thread modifies A to B, another thread will modify B to A again, causing casd to execute multiple times.

And ABA problems are usually more likely to emerge at the moment of deletion

[0008]The traditional approach is to implement a garbage collection mechanism and use freelist for final deletion. This approach has no logical problems in concurrent scenarios, but has efficiency problems, especially in resource In a limited hardware system, if a hash table with a garbage collection mechanism is established, the code size and complexity of garbage collection will greatly exceed the hash table itself

Images