No release in over 3 years
Low commit activity in last 3 years
There's a lot of open issues
AWS Ruby Library for interfacing with Amazon Web Services including EC2, S3, SQS, SimpleDB and most of their other services as well. By http://www.appoxy.com
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
 Dependencies
 Project Readme

DEPRECATED

Please use this gem now: https://rubygems.org/gems/aws-sdk

Appoxy AWS Library

A Ruby gem for all Amazon Web Services.

Brought to you by: Appoxy

Discussion Group

http://groups.google.com/group/ruby-aws

Documentation

Ruby Docs

Appoxy Amazon Web Services Ruby Gems

Published by Appoxy LLC, under the MIT License. Special thanks to RightScale from which this project is forked.

INSTALL:

gem install aws

Then require 'aws' in your application.

DESCRIPTION:

The AWS gems have been designed to provide a robust, fast, and secure interface to Amazon EC2, EBS, S3, SQS, SDB, and CloudFront. The AWS gems comprise:

  • Aws::Ec2 -- interface to Amazon EC2 (Elastic Compute Cloud) and the associated EBS (Elastic Block Store)
  • Aws::S3 and Aws::S3Interface -- interface to Amazon S3 (Simple Storage Service)
  • Aws::Sqs and Aws::SqsInterface -- interface to Amazon SQS (Simple Queue Service)
  • Aws::SdbInterface -- interface to Amazon SDB (SimpleDB). See SimpleRecord for an ActiveRecord like gem.
  • Aws::AcfInterface -- interface to Amazon CloudFront, a content distribution service
  • Aws::ElbInterface -- interface to Amazon Load Balancing service
  • Aws::MonInterface -- interface to Amazon CloudWatch monitoring service
  • Aws::Iam -- for AWS Identity and Access Management

To use a single piece intead of loading all of then, you can require it explicitly for example: require 'aws/sqs'.

FEATURES:

  • Full programmmatic access to EC2, EBS, S3, SQS, SDB, ELB, and CloudFront.
  • Complete error handling: all operations check for errors and report complete error information by raising an AwsError.
  • Persistent HTTP connections with robust network-level retry layer using RightHttpConnection). This includes socket timeouts and retries.
  • Robust HTTP-level retry layer. Certain (user-adjustable) HTTP errors returned by Amazon's services are classified as temporary errors. These errors are automaticallly retried using exponentially increasing intervals. The number of retries is user-configurable.
  • Fast REXML-based parsing of responses (as fast as a pure Ruby solution allows).
  • Uses libxml (if available) for faster response parsing.
  • Support for large S3 list operations. Buckets and key subfolders containing many (> 1000) keys are listed in entirety. Operations based on list (like bucket clear) work on arbitrary numbers of keys.
  • Support for streaming GETs from S3, and streaming PUTs to S3 if the data source is a file.
  • Support for single-threaded usage, multithreaded usage, as well as usage with multiple AWS accounts.
  • Support for both first- and second-generation SQS (API versions 2007-05-01 and 2008-01-01). These versions of SQS are not compatible.
  • Support for signature versions 0, 1 and 2 on all services.
  • Interoperability with any cloud running Eucalyptus (http://eucalyptus.cs.ucsb.edu)
  • Test suite (requires AWS account to do "live" testing).

THREADING:

All AWS interfaces offer three threading options:

  1. Use a single persistent HTTP connection per process. :single
  2. Use a persistent HTTP connection per Ruby thread. :per_thread
  3. Open a new connection for each request. :per_request

Either way, it doesn't matter how many (for example) Aws::S3 objects you create, they all use the same per-program or per-thread connection. The purpose of sharing the connection is to keep a single persistent HTTP connection open to avoid paying connection overhead on every request. However, if you have multiple concurrent threads, you may want or need an HTTP connection per thread to enable concurrent requests to AWS. The way this plays out in practice is:

  1. If you have a non-multithreaded Ruby program, use the non-multithreaded setting.
  2. If you have a multi-threaded Ruby program, use the multithreaded setting to enable concurrent requests to S3 (or SQS, or SDB, or EC2).
  3. For running under Mongrel/Rails, use the non-multithreaded setting even though mongrel is multithreaded. This is because only one Rails handler is invoked at time (i.e. it acts like a single-threaded program)

Note that due to limitations in the I/O of the Ruby interpreter you may not get the degree of parallelism you may expect with the multi-threaded setting.

By default, EC2/S3/SQS/SDB/ACF interface instances are created in per_request mode. Set params[:connection_mode] to :per_thread in the initialization arguments to use multithreaded mode.

== LICENSE:

Copyright (c) 2007-2009 RightScale, Inc.

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the 'Software'), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.