ANAGRAMMER ⁉️

An anagram finder framework for Apache Mesos.

See the accompanying slides for more context.

ANAGRAMMER consists of three main components:

• FinderExecutor extends mesos.Executor
• DefinerExecutor extends mesos.Executor
• RenderingFinder extends mesos.Scheduler and launches tasks with the executors

Quick Start with Vagrant

Requirements

• VirtualBox 4.1.18+
• Vagrant 1.3+
• git (command line tool)

Start the mesos-demo VM

$ wget http://downloads.mesosphere.io/demo/mesoscon.box -O /tmp/mesoscon.box
$ vagrant box add --name mesos-demo /tmp/mesoscon.box
$ git clone https://github.com/mesosphere/ANAGRAMMER.git
$ cd ANAGRAMMER
$ vagrant up

Now that the VM is running, you can view the Mesos Web UI here: http://10.141.141.10:5050

You can see that 1 slave is registered and you've got some idle CPUs and Memory. So let's start the ANAGRAMMER!

Run ANAGRAMMER in the mesos-demo VM

$ vagrant ssh
vagrant@mesos:~ $ cd hostfiles

# Start the scheduler with the seed word, the mesos master ip and optionally a task limit
vagrant@mesos:hostfiles $ python anagrammer.py word 127.0.1.1:5050 42

# See results
vagrant@mesos:hostfiles $ less result.dot
# <Ctrl+C> to stop..

Shutting down the mesos-demo VM

# Exit out of the VM
vagrant@mesos:hostfiles $ exit
# Stop the VM
$ vagrant halt
# To delete all traces of the vagrant machine
$ vagrant destroy

Anagrammer Architecture (Follows RENDLER framework)

Finder Executor

• Interprets incoming tasks' task.data field as a word
• Fetches the anagrams for that word and extracts them from the document
• Sends a framework message to the scheduler containing the finder result (the anagrams).

Definer Executor

• Interprets incoming tasks' task.data field as a word
• Fetches the anagram, and find the definition of it, if it's a real word
• Sends a framework message to the scheduler containing the definer result (the definition).

Intermediate Data Structures

We define some common data types to facilitate communication between the scheduler and the executors. Their default representation is JSON.

results.FinderResult(
    "1234",                                 # taskId
    "foo",                                  # word
    ["foo", "oof"]                          # anagrams
)

results.DefinerResult(
    "1234",                                 # taskId
    "foo",                                  # word
    "definition of foo"                     # definition
)

Anagrammer Scheduler

Data Structures

• finderQueue: list of words
• definerQueue: list of words
• processedwords: set or words
• finderResults: list of word tuples
• definerResults: map of words to definitions

Scheduler Behavior

The scheduler accepts one word as a command-line parameter to seed the definer and finder queues.

1. For each word, create a task in both the definer queue and the finder queue.

2. Upon receipt of a finder result, add an element to the finder results adjacency list. Append to the definer and finder queues each word that is not present in the set of processed words. Add these enqueued words to the set of processed words.

3. Upon receipt of a definer result, add an element to the definer results map.

4. The finder and definer queues are drained in FCFS order at a rate determined by the resource offer stream. When the queues are empty, the scheduler declines resource offers to make them available to other frameworks running on the cluster.