
-
Spark Continuous Processing
Continuous Processing is Apache Spark's new Execution engine that allows very low latency(in milliseconds) event at a time processing. In earlier versions, streaming was done via micro-batching. In continuous processing, Spark launches long-running tasks that continuously read, process and write data. In this blog, we are going to do an early peek at this still experimental feature in Apache Spark that is going to be available in version 2.3.
-
A Tour of Spark Structured Streaming
Structured Streaming is Apache Spark's streaming engine which can be used for doing near real-time analytics. In this blog we explore Structured Streaming by going through a very simple use case.
-
Deep Learning - ANN, RNN, LSTM networks
Long Short Term Memory(LSTM) model is a type supervised Deep Neural Network that is very good at doing time-series prediction. In this blog, we do a step by step exploration of it's architecture starting from the basic NN, then RNN leading to LSTM.
-
Session Windows in Apache Flink
Apache Flink's Session Windows allows messages to be windowed into sessions. In this blog, we will create a streaming application that counts number of Clicks made by each user within a session using EventTimeSession windows.
-
Kafka Streams - Part 2
This is continuation of the blog post - "Hello Kafka Streams". In this blog we build a bit more complicated application that demonstrates the use of flatMapValues, branch, predicate, selectKey, through, join and also see how to create a custom SerDe using Kryo.