Repository for cloud computing projects and study files. Contains mainly cloud solutions on Java. Some projects use C++.

An implementation of a movie recommendation system that runs on Spark engine. The recomendations are made by a custom collaborative filtering algorithm based on rates given by users. The algorithm calculates movie cosine similarity scores in a distributed way. The implementation is done on Python and uses PySpark framework.

This is a capstone project for making different computations on Big Dataset (ca 200 000 000 records) using Hadoop and Kafka + Spark Streaming frameworks, as well as Cassandra NoSQL Database. Solutions are mainly made on Java. The data processing is done with R.

The solutions for programming assignments for Coursera Cloud Applications course. It contains the MapReduce, Apache Storm, HBase, Giraph and Spark MlLib solutions. The solutions are mainly made on Java.

Creating a concept of system for dealing with multiple sources of sensors data. The project is made on Java. It contains a UDP Values Generator which emulates sensors on client side and sends the data to cloud. The data flow is being processed by Apachi NiFi cluster and then send to Hadoop Ecosystem cluster for further reworking. The following systems are used on Hadoop side: Spark Streaming, Spark, HDFS, HBase, Kafka.

The implementation of Gossip Style membership protocol for P2P connections. It implements the gossip style heart beating that is used in some real cloud systems. Project is implemented on C++. Protocol satisfies following: Completeness all the time (every non-faulty process, must detect every node join, failure, and leave), accuracy of failure detection.

The implementation of a Fault-Tolerant Key-Value store that has ring-based structure (similar to Cassandra). The store provides: CRUD Operations, Load-Balancing, Fault-Tolerance (via replication), Quorum consistency level and Stabilization after failure functionality. The implementation is made on C++.