[
https://issues.apache.org/jira/browse/IGNITE-16594?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Andrey N. Gura updated IGNITE-16594:
------------------------------------
Description:
Need to create infrastructure to calculate hash code for colocation key.
Apache Ignite uses partitioning by a hash function. The hash function should
have relatively even distribution. Because of it we can’t use primitive and
very fast commutative hash functions (e.g. based on XOR). So the hash function
should respect the order of fields that are included into the affinity key.
It should be taken into account that the code for hash code calculation should
be accessible from thin clients.
was:
Need to create infrastructure to calculate hash code for colocation key.
Apache Ignite uses partitioning by a hash function. The hash function should
have relatively even distribution. Because of it we can’t use primitive and
very fast commutative hash functions (e.g. based on XOR). So the hash function
should respect the order of fields that are included into the affinity key.
> Colocation key hash code calculation
> ------------------------------------
>
> Key: IGNITE-16594
> URL: https://issues.apache.org/jira/browse/IGNITE-16594
> Project: Ignite
> Issue Type: Improvement
> Reporter: Yury Gerzhedovich
> Priority: Major
> Labels: ignite-3
>
> Need to create infrastructure to calculate hash code for colocation key.
> Apache Ignite uses partitioning by a hash function. The hash function should
> have relatively even distribution. Because of it we can’t use primitive and
> very fast commutative hash functions (e.g. based on XOR). So the hash
> function should respect the order of fields that are included into the
> affinity key.
> It should be taken into account that the code for hash code calculation
> should be accessible from thin clients.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
