Practical Course Cloud Databases

A distributed key-value database system for the master seminar "Cloud Databases".

Consists of three main parts:

• distributed, replicated graph DB Servers
• configuration service that monitors health state of DB's to take action on failure
• a client to query and mutate data (query style inspired by GraphQL)

NO PRODUCTION USAGE Implemented for a seminar in a hackthon like way. Should only be used as a reference for a refactoring session.

SETUP

Run tests: ant test

Build jars for server, client and ecs: ant build-jar

Run the server java -jar server.jar --port 50000 --cache-type LFU --cache-size 10 --log-level WARN

Run the client java -jar client.jar

Run the ECS (make sure ecs.config file is available in the root folder) java -jar ecs.jar

Folder structure

• src contains all source code
• test contains all tests
• logs contains logs
• db contains dbs
• docs contains the performance evaluation
• tools contains performance tools

Milestone 5 - Graph Database Extension

The datastore is extended as a graph database which could store structured data and maintain associations between them. This gives the ability to perform powerful queries over data which was not possible with the previous simple keyValue store. The existing KVClient is extended with two command MUTATE to save/modify data and QUERY to query data

• run the server
  java -jar server.jar --port 50000 --cache-type LFU --cache-size 10 --log-level WARN

• run the client java -jar client.jar

• We can create new data using MUTATE command

  The syntax

      MUTATE <documentId> <key1>:<property1>,<key2>,<property2>,.....
  

  Example : Consider the following data that, we need to store

      {id:[email protected],firstName:John,lastName:Doe,Occupation:Engineer}
  

  The query for storing data

      MUTATE [email protected] firstName:John,lastName:Doe,Occupation:Engineer
  

  You can also specify JSON objects as properties

      MUTATE [email protected] age:30,name:{firstName:Philip,lastName:Crow}
  

• MUTATE command also can be used to add new properties or modify existing properties

  The Syntax

      MUTATE <documentIdToModify> <existingkey1>:<modifiedProperty1>,<newKey02>:<newProperty2>,....
  

  Example :

      MUTATE [email protected] lastName:Davidson,age:45
  

  The above query would modify the existing property lastName and also add a new property age

• We can query data using QUERY command

  The Syntax

      QUERY <documentId> <key1>,<key2>,...
  

  Example : Consider following data already in the graph database with a documentId [email protected]

      [email protected]:{firstName:John,lastName:Doe,Occupation:Engineer}
  

  The query to retrieve firstName and lastName

      QUERY [email protected] firstName,lastName
  

  The result

      {
        firstName: John,
        lastName: Doe
      }
  

• Creating associations between data

  The Syntax

      MUTATE <documentId> <key>:<associatedDocumentId>
  

  Example : Consider we have below data already in the database

      [email protected]:{firstName:John,lastName:Doe,Occupation:Engineer}
      [email protected]:{firstName:Anna,age:34}
  

  Associate [email protected] as John's best friend

      MUTATE [email protected] bestFriend:[email protected]
  

  Now we can query this data

      QUERY [email protected] bestFriend
  

  The result would be

      {
        bestFriend: [email protected]
      }
  

  We can query the properties of associated objects using FOLLOW

  Example :

      QUERY [email protected] bestFriend|FOLLOW{firstName,age}
  

  Which would give

  {
    bestFriend: {
      firstName: Anna,
      age: 32
    }
  }