Overview
Spark Streaming is an extension of the core Spark API that enables scalable, high-throughput, fault-tolerant stream processing of live data streams.
Data can be ingested from many sources like
- Kafka,
- Flume,
- Twitter Streaming and REST APIs,
- ZeroMQ,
- Amazon Kinesis, or
- TCP sockets,
- etc
and can be processed using
complex algorithms expressed with high-level functions like map,
reduce, join and window.
Finally, processed data can be pushed out to filesystems, databases, and live dashboards. In fact, you can apply Spark's
- machine learning and
- graph processing algorithms on data streams.
Internally, it works as follows:
- Spark Streaming receives live input data streams and
- divides the data into batches,
- which are then processed by the Spark engine
- to generate the final stream of results in batches.
Spark Streaming provides a high-level abstraction called discretized stream or DStream, which represents a continuous stream of data. DStreams can be created either from input data streams from sources such as Kafka, Flume, and Kinesis, or by applying high-level operations on other DStreams. Internally, a DStream is represented as a sequence of RDDs.
This guide shows you how to start writing Spark Streaming programs with DStreams. You can write Spark Streaming programs in Scala, Java or Python (introduced in Spark 1.2), all of which are presented in this guide.
Here, we will focus on Streaming in Scala.
Spark Streaming is a near-real-time micro-batch stream processing engine as opposed to other real-time stream processing frameworks like Apache Storm. Typically 'near-real-time' in Spark Streaming can be in the order of seconds as opposed to milliseconds, for example.
// === Configuration to control the flow of the application === val stopActiveContext = true // "true" = stop if any existing StreamingContext is running; // "false" = dont stop, and let it run undisturbed, but your latest code may not be used // === Configurations for Spark Streaming === val batchIntervalSeconds = 1 val eventsPerSecond = 1000 // For the dummy source // Verify that the attached Spark cluster is 1.4.0+ require(sc.version.replace(".", "").toInt >= 140, "Spark 1.4.0+ is required to run this notebook. Please attach it to a Spark 1.4.0+ cluster.")
Discretized Streams (DStreams)
Discretized Stream or DStream is the basic abstraction provided by Spark Streaming. It represents a continuous stream of data, either the input data stream received from source, or the processed data stream generated by transforming the input stream. Internally, a DStream is represented by a continuous series of RDDs, which is Spark's abstraction of an immutable, distributed dataset (see Spark Programming Guide for more details). Each RDD in a DStream contains data from a certain interval, as shown in the following figure.
// This is the dummy source implemented as a custom receiver. No need to fully understand this. import scala.util.Random import org.apache.spark.streaming.receiver._ class DummySource(ratePerSec: Int) extends Receiver[String](StorageLevel.MEMORY_AND_DISK_2) { def onStart() { // Start the thread that receives data over a connection new Thread("Dummy Source") { override def run() { receive() } }.start() } def onStop() { // There is nothing much to do as the thread calling receive() // is designed to stop by itself isStopped() returns false } /** Create a socket connection and receive data until receiver is stopped */ private def receive() { while(!isStopped()) { store("I am a dummy source " + Random.nextInt(10)) Thread.sleep((1000.toDouble / ratePerSec).toInt) } } }
Transforming and Acting on the DStream of lines
Any operation applied on a DStream translates to operations on the
underlying RDDs. For converting
a stream of lines to words, the flatMap operation is applied on each
RDD in the lines DStream to generate the RDDs of the wordStream DStream.
This is shown in the following figure.
These underlying RDD transformations are computed by the Spark engine. The DStream operations hide most of these details and provide the developer with a higher-level API for convenience.
Next reduceByKey is used to get wordCountStream that counts the words in wordStream.
Finally, this is registered as a temporary table for each RDD in the DStream.
var newContextCreated = false // Flag to detect whether new context was created or not // Function to create a new StreamingContext and set it up def creatingFunc(): StreamingContext = { // Create a StreamingContext - starting point for a Spark Streaming job val ssc = new StreamingContext(sc, Seconds(batchIntervalSeconds)) // Create a stream that generates 1000 lines per second val stream = ssc.receiverStream(new DummySource(eventsPerSecond)) // Split the lines into words, and then do word count val wordStream = stream.flatMap { _.split(" ") } val wordCountStream = wordStream.map(word => (word, 1)).reduceByKey(_ + _) // Create temp table at every batch interval wordCountStream.foreachRDD { rdd => rdd.toDF("word", "count").createOrReplaceTempView("batch_word_count") } stream.foreachRDD { rdd => System.out.println("# events = " + rdd.count()) System.out.println("\t " + rdd.take(10).mkString(", ") + ", ...") } ssc.remember(Minutes(1)) // To make sure data is not deleted by the time we query it interactively println("Creating function called to create new StreamingContext") newContextCreated = true ssc }
// Stop any existing StreamingContext // The getActive function is proviced by Databricks to access active Streaming Contexts if (stopActiveContext) { StreamingContext.getActive.foreach { _.stop(stopSparkContext = false) } } // Get or create a streaming context val ssc = StreamingContext.getActiveOrCreate(creatingFunc) if (newContextCreated) { println("New context created from currently defined creating function") } else { println("Existing context running or recovered from checkpoint, may not be running currently defined creating function") } // Start the streaming context in the background. ssc.start() // This is to ensure that we wait for some time before the background streaming job starts. This will put this cell on hold for 5 times the batchIntervalSeconds. ssc.awaitTerminationOrTimeout(batchIntervalSeconds * 5 * 1000)
SDS-2.x, Scalable Data Engineering Science
Last refresh: Never