apex-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Siyuan Hua (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (APEXMALHAR-2244) Optimize WindowedStorage and Spillable data structures for time series
Date Thu, 22 Sep 2016 16:40:21 GMT

    [ https://issues.apache.org/jira/browse/APEXMALHAR-2244?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15513770#comment-15513770

Siyuan Hua commented on APEXMALHAR-2244:

Each spillable DS implementation use a SpillableStateStore to store things and we can make
ManagedTimeUnifiedStateImpl implement the store as well and it can take some time extract
function to get/calculate time and time buckets from each V/KV data.  And the Store can be
setup by the WindowedOperator, correct? 

> Optimize WindowedStorage and Spillable data structures for time series
> ----------------------------------------------------------------------
>                 Key: APEXMALHAR-2244
>                 URL: https://issues.apache.org/jira/browse/APEXMALHAR-2244
>             Project: Apache Apex Malhar
>          Issue Type: Sub-task
>            Reporter: David Yan
>            Assignee: Siyuan Hua
> The spillable data structures currently does not make any assumption about the key that
is used in Managed State, and as a result, it uses ManagedStateImpl to interface with Managed
State and uses time buckets that are based on the apex window id. But for WindowedStorage
used by WindowedOperator, the key to the storage is a window, which is event time based. Using
the default ManagedStateImpl would be very inefficient for event time based keys, since it
would write data that would belong to the same window to different time buckets.
> On a high level, the below summarizes roughly what needs to be done:
> 1. a way to tell the spillable data structures to use the ManagedTimeUnifiedStateImpl
> 2. a way to tell the spillable data structures how to extract the timestamp from the
key. Note that in the case of WindowedOperator, the timestamp should be the end timestamp
of the window (beginTimeMillis + durationMillis), not the begin timestamp.
> 3. a way to tell the spillable data structures how to assign the time bucket given that
> 4. with point 3, the spillable implementations of WindowedStorage will need to take a
config parameter that says how much time (in millis) is each time bucket
> 5. only purge a time bucket when all keys that belong to that time bucket are removed
and the apex window id of the first window in which the keys are all removed has been committed

This message was sent by Atlassian JIRA

View raw message