| Interface | Description |
|---|---|
| DataLoadingService |
All data loading services should implement this interface.
|
| DataSet<K,V> |
A view of the data set to be loaded
at an evaluator as an iterable of
key value pairs.
|
| EvaluatorToPartitionStrategy<V extends org.apache.hadoop.mapred.InputSplit> |
Interface that tracks the mapping between evaluators and the data partitions
assigned to those evaluators.
|
| Class | Description |
|---|---|
| DataLoader |
The driver component for the DataLoadingService
Also acts as the central point for resource requests
All the allocated evaluators pass through this and
the ones that need data loading have a context stacked
that enables a task to get access to Data via the
DataSet. |
| DataLoadingDriverConfiguration | |
| DataLoadingRequestBuilder |
Builder to create a request to the DataLoadingService.
|
| DataLoadingRequestBuilder.LoadDataIntoMemory | |
| DataLoadingRequestBuilder.NumberOfDesiredSplits | |
| DistributedDataSet |
Represents a distributed data set that is split across data centers.
|
| ResourceRequestHandler |
Copyright © 2017 The Apache Software Foundation. All rights reserved.