wiki:GEMINIv1.1Tutorial

Version 4 (modified by Ahmed El-Hassany, 11 years ago) (diff)

Add VM instructions

Basic Experiment with GEMINI v1.1 I&M Tools: Instructions

Preparation

This tutorial assumes that attendees have a basic knowledge of ProtoGENI/InstaGENI, and we recommend that interested participants also consider attending the InstaGENI tutorial offered at GEC 15.

Active participants will need a laptop equipped with a recent version of Virtual Box. If you are unable to bring one, you may partner with someone else.

Steps

0) Overview of experiment

The instructions on this page are designed for the GEMINI Tutorial at GEC 15. This tutorial covers how to instrumentize a GENI slice using the GEMINI I&M system.

The GEMINI I&M system provides an easy way to instrumentize a GENI slice with passive and active measurements. The current version of GEMINI provides host monitoring metrics (cpu, memory, network) and tools to perform on-demand and regular active measurements for throughput, one-way latency and round-trip latency (both with jitter and loss).

Users can specify through their request RSpec which nodes should have passive instrumentation (e.g. host monitoring) and which nodes should active measurement services installed. While most passive instrumentation is always on, active measurements must be configured by the user. This is done through a web interface available on a instrumentized slice. A key service for this to work in GEMINI is UNIS: Unified Network Information Service.

UNIS and Topology-based Configuration: The Unified Network Information Service (UNIS) is the most important
component of the GEMINI architecture. For those familiar with perfSONAR, this service is the combination of
the Lookup Service and the Topology Service of the perfSONAR framework. UNIS stores the topology data of GENI
slices, and services register themselves so that they can be found dynamically. The configuration of the
active measurement services is done through annotations on the topology data. A web interface is provided to
configure the active measurements and push the configuration to UNIS.

0.1) Goals

In this tutorial you will learn all the steps necessary to fully utilize the GEMINI I&M system, from instrumentizing and bootstrapping your slice to configuring measurements to visualizing and archiving the data.

0.2) Configuration

For the tutorial given at GEC15, the slices will be pre-made and instrumentized. The information for the slice will be given to you on an info sheet and you can also find your information here. Make sure you are accessing your slice!

0.3 Process (flow chart)

TODO

1) Establish experiment environment

1.1) Preparing the User workspace Virtual Machine

  • Install VirtualBox

Download the VirtualBox software from http://www.virtualbox.org.

If you already have VirtualBox installed on your machine, make sure it is version 4.1 or above.

  • Download VirtualBox VM image for tutorials

Download the VirtualBox VM image (GEC15_Tutorials_Final.ova) from the website https://www.planet-lab.org/GEC15/

  • Install the GEC15_Tutorials_Final.ova virtual machine image

Start up VirtualBox, select File->Import Appliance..., and follow the instructions. Accept the default VM settings during the import.

To run the virtual machine, go to the Oracle VM VirtualBox Manager window, select the VM and click the green arrow labeled Start at the top of this window.

If the install was successful, you should see the logon screen for the Ubuntu OS.

1.2) Gather necessary keys, certificates and credentials

1.3) Verify availability of desired aggregates

1.4) Verify availability of desired software images/packages

1.5) Verify availability of necessary operational services, and access to those services

2) Obtain slice of GENI resources, install I&M tools and experiment services

2.1) Select target aggregates 2.2) Select desired software images/packages for I&M tools and experiment services

2.3) Obtain certs, credentials, passwords and keys

2.4) Formulate slice topology for experiment, and build request rspec

2.5) Acquire resources and load images/packages for I&M tools and experiment services

2.6) Start selected software processes

2.7) Confirm slice with installed I&M tools and experiment services

3) Configure I&M tools and experiment services

3.1) Configure selected I&M services

  • objectives:
    • binds services together, e.g., between each MP and the MC
    • configures baseline services
    • verifies that they are ready to run

  • at completion: configured I&M services, ready to run and collect data

3.2) Configure selected experiment services

  • objectives:
    • binds services together to realize overall reference/actual experiment
    • configures baseline services
    • verifies that they are ready to run

  • at completion: configured experiment services, ready to run and process data, which can be observed by I&M services

4) Run and orchestrate I&M services and experiment services to complete run of experiment

4.1) Initial setup: start basic host measurements and basic ping active network measurements

  • objectives:
    • verify functionality of hosts
    • verify topology of slice
  • observe measurements with a real-time presentation service

  • continue measurements throughout duration of the test/tutorial/experiment

  • at completion:
    • functionality of hosts and topology of slice has been verified throughout duration of the experiment

4.2) Continuity test: for a limited time, run iperf active network measurements

  • objectives:
    • verify ability of slice to carry traffic expected from experiment
  • observe measurements with a real-time presentation service

  • once satisfactory measurements have been observed, stop continuity test

  • at completion:
    • capability of slice to carry traffic expected from experiment has been verified

4.3) Instrument and run experiment:

  • objectives:
    • gather measurements during experiment that allow experiment goals to be met

  • include desired measurement points within hosts and/or experiment services to instrument test/tutorial/experiment

  • begin to run and orchestrate measurement services

  • begin to run and orchestrate experiment services

  • observe measurements with a real-time presentation service, to verify expected operation of experiment

  • collect all measurements for duration of experiment

  • stop experiment services, when this run of the experiment has been completed

  • stop measurement services

  • at completion:
    • one run of experiment has been completed
    • real-time look at measurements has verfied expected operation of experiment
    • a full set of measurements has been collected, for later analysis and presentation
    • collected measurments have been transfered to storage service, so that slice resources can be released (if desired)

4.4) Store collected measurements and other artifacts from test/tutorial/experiment in storage service

  • at completion:
    • collected measurements and other artifacts have been transfered to storage service
    • collected measurements and other artifacts are available for later analysis
    • slice resources can then be released at any time, without loss of any measurements or artifacts

5) Analyze and visualize measurement results after completing run of experiment

  • if necessary, retrieve measurement results from archive service

  • analyze and format results as desired, for visualization with presentation service

  • as appropriate, store analyzed results and/or visualization in storage service

6) Move selected collected measurements and other artifacts from storage service to long-term archive service

  • identify archived objects with peristent identifier

  • include policy for sharing with others

  • allow retrieval for further analysis and visualization

7) Release experiment resources

Attachments (41)