CloudPowers is a wrapper around certain AWS services. It was developed with the need instead of the resource in mind so even though AWS is the only service provider included now, it shouldn't be the only one forever. There are several pieces of the Cloud Power module to talk about:
-
SelfAwareness: Gathers data about "self"
-
Synapse: Handles communication needs
-
Delegator: dynamically include Ruby executables from S3 and more...
-
Each module in Cloud Powers is in charge of a specific type of task that helps bring projects together and communicate with the outside world.
Below is a breakdown of the installation, common services provided and an example or 2 of popular methods. Better docs are on the way
Add this line to your application's Gemfile:
gem 'cloud_powers'then execute:
$ bundle
Or install it yourself as:
$ gem install cloud_powers
then either: * set environment variables that matches the below group * fill out a .env file and load it from the class that is using CloudPowers, like this Notes: * The code does its best in many cases to make a good guess at some of the configuration for you but if you set them in your system environment variables or in the .env file, it'll have a much better
require 'dotenv'
Dotenv.load('path to your .env file')things you need for pre-v1:
# Aws access:
AWS_ACCESS_KEY_ID=""
AWS_SECRET_ACCESS_KEY=""
# Aws areas and auth-related locations:
AWS_REGION=""
# Aws Build info:
AMI_NAME="for Cerebrums to create a node"
# Aws kinesis stream names
STATUS_STREAM="e.g. kinesis stream name"
# Aws s3 buckets etc
TASK_STORAGE="e.g. s3 object name"
# Aws sqs queue addresses
BACKLOG_QUEUE_ADDRESS=""
WIP_QUEUE_ADDRESS=""
FINISHED_QUEUE_ADDRESS=""* AWS resources that should be used by many services, e.g. Delegator _on EC2_ uses S3, EC2 and SQS to gain knowledge about the
context it will be working in.
- Helps a node figure out what task it should be running, where its executables are, gathers it/them and loads them.
- Lives in the Job and Task instantiation chain
- useful shared methods, like one that turns a string into snake_case
- gets and sets info about the instance -> Neuron/Cerebrum/etc)
get_awareness!- retrieves and sets all metadata from the EC2 instance and a few other things like the instance hostname (can find the instance IP from here).
- See EC2 metadata for details on the instance/node metadata that is set.
- Additionally, the instance public hostname, id and a few others are set using other-than-EC2-metadata methods.
- S3
- A Synapse is used for communicating, usually between nodes and an external source for status updates but the Synapse can handle any kind of information that needs to be passed.
- Architecture
- The Synapse is a communications module that is broken up into separate types of communication, for separate needs.
- There are 2 modules, soon to be 3, inside the Synapse module:
* like a task list or a message board for asynchronous communication
* Board <Struct>:
* interface with Queue config, data, name, etc.
```Ruby
default_workflow = Workflow.new
board = Board.new(default_workflow)
board.name
=> 'backlog'
board.address
=> 'http://aws-url/backlog-board-url'
board.next_board # useful because this is a simple state-machine implementation
=> 'wip'
```
Example usage:
1. Give the entire stream name or a symbol or string that is found in the .env for the name of the Queue
2. provide a block that will create the record that gets sent to the board
```Ruby
poll(board_name <string|symbol>, opts <optional config Hash>) { block }
```
or
```Ruby
poll(:backlog) { |msg, stats| Task.new(instance_id, msg) if stats.success? }
```
or
```Ruby
poll(:backlog, wait_time: 30, delete: false) do
edited_message = sitrep.merge(foo: 'bar')
update = some_method(edited_message)
end
```
* Good for real time information passing, like status updates, results reporting, operations, etc.
* The Pipe module is for communicating via streams. Piping is meant to be a way to communicate status, problems and other general info or huge result sets and things of that nature. There can be very high traffic through the pipe or none at all. Very soon (probably V-0.2.6 or 7), Cerebrums will be data consumers, to the Java KCL and MultiLangDaemon level, so keeping messages straight is done via partition ID. The partition ID of any message is to identify which node the message is about or the "batch" or "group" or any other concept like that. So the instance ID is used in nodes like the Neuron and Cerebrum and other identifiers that are deemed best are used in other projects as a batch ID.
Example usage:
```Ruby
pipe_to(stream_name <string/symbol>) { &block }
```
```Ruby
pipe_to(:status_queue) { sitrep(content: 'workflowComplete') }
```
and for multiple records (KCL)
```Ruby
flow_to(stream_name <string/symbol>) { &block }
```
```Ruby
flow_to(:status_queue) do
interesting_instances = neurons.map do |neuron|
return neuron if neuron.workflow.done?
end
find_efficient_neurons(interesting_instances) # this gets sent through the Pipe
end
```
* Allows the nodes to have a collective awareness of each other, the environment they work in and other Jobs (Coming soon...)
After checking out the repo, run bin/setup to install dependencies. Then, run rake spec to run the tests. You can also run bin/console for an interactive prompt that will allow you to experiment.
To install this gem onto your local machine, run bundle exec rake install. To release a new version, update the version number in version.rb, and then run bundle exec rake release, which will create a git tag for the version, push git commits and tags, and push the .gem file to rubygems.org.
Bug reports and pull requests are welcome on GitHub at https://github.com/adam-phillipps/cloud_powers.
The gem is available as open source under the terms of the MIT License.