Github user tdas commented on a diff in the pull request:
https://github.com/apache/spark/pull/7648#discussion_r35958079
--- Diff:
streaming/src/main/scala/org/apache/spark/streaming/scheduler/rate/PIDRateEstimator.scala
---
@@ -0,0 +1,109 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.spark.streaming.scheduler.rate
+
+/**
+ * Implements a proportional-integral-derivative (PID) controller which
acts on
+ * the speed of ingestion of elements into Spark Streaming. A PID
controller works
+ * by calculating an '''error''' between a measured output and a desired
value. In the
+ * case of Spark Streaming the error is the difference between the
measured processing
+ * rate (number of elements/processing delay) and the previous rate.
+ *
+ * @see https://en.wikipedia.org/wiki/PID_controller
+ *
+ * @param batchDurationMillis the batch duration, in milliseconds
+ * @param proportional how much the correction should depend on the current
+ * error. This term usually provides the bulk of correction. A
value too large would
+ * make the controller overshoot the setpoint, while a small value
would make the
+ * controller too insensitive. The default value is -1.
--- End diff --
Assuming that we are switching the signs (i.e. default is 1 instead of -1),
what does proportional < 0 mean?
That is just plain wrong! You are increasing the rate when it should be
decreased. And anything more than 1 means that you are compensating more than
feasible. Consider this. say there is a workload with 1 second batches where
the system can process 5 records in 0.8 seconds. Say the rate limit was 10.
After first batch with 5 recods in 0.8 seconds process delay (no scheduling
delay), the calculation would be (assuming proportional = 1.5)
```
error = rateLimit - numRecords/procDelay 10 - (5/0.8) = 10 - 6.25 = 3.75
newRate = rateLimit - 1.5 * error = 10 - 1.5 * 3.75 = 4.375
```
4.375 is less than 5 records that the system can handle in 0.8 secs. This
is wrong, that is going to happen for proportional > 1. With proportional = 1,
the calculation is.
`new rate = 10 - 1.0 * (10 - (5/0.8)) = 6.25 `
This same as it had processed in the previous batch. Perfectly feasible.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
