Table Of Contents

Previous topic

Use TinySVM

Next topic

Overview

This Page

Document

Overview

Tutorial

Speed up user programs

Tutorial 1-1: Run a user program as a RaSC service

  • Explains how to reduce overhead of startup of user programs and call the programs on remote computers.
  • Overviews the basic settings of a RaSC service.

Tutorial 1-2: Execute a user program in parallel

  • Explains how to execute user programs in parallel on one computer, using a multi-core CPU.
  • Explains parallel execution inside RaSC and usecases.

Tutorial 1-3: Call a user program from various programming languages

  • Provides some examples of calling a RaSC service from Perl, Python and Ruby.

Tutorial 1-4: Run your analysis tool as an RaSC service (in preparation)

  • Explains how to define interface tailored to the function of a user program.

Improve usability and compatibility

Tutorial 2-1: Work with various network protocols

  • Explains how to use various network protocols, JSON RPC, ProtocolBuffers, and SOAP in additon to MessagePack RPC.
  • Shows usage of JSON RPC through command line or the Web browser interface.

Tutorial 2-2: Call through ProtocolBuffers

  • Explains how to use ProtocolBuffers, a high-speed binary protocol developed by Google.
  • Shows a Java client program to call a RaSC service through ProtocolBuffers.

Tutorial 2-3: Call by SOAP

  • Explains how to a call RaSC service with SOAP, which is the major protocol for SOA (service-oriented architecture).
  • Shows the SOAP standalone client soapUI to call a RaSC service.

Distributed execution using multiple computers

Tutorial 3-1: Distributed execution with load balancing

  • Executes RaSC services on multiple nodes to process large scale data, and aggregates the results.
  • Introduces load balancing using a round robin.

Tutorial 3-2: Monitor distributed execution (in preparation)

  • Explains monitoring over RaSC services in a distributed environment, using Fluentd and growthforecast.