It was officially released as part of the first version of the CELAR Elasticity Platform, which uses SlipStream as an applications deployment provisioner and orchestrator. Whilst the CELAR platform can be configured to work with any cloud supported by SlipStream, the development of the connector allows FCO to come pre-installed.

To fully integrate with all the workflows available in SlipStream, the connector covers the run image, build imagine, and start/terminate deployment capabilities. As part of the scalability workflows, the project tested the connector on real applications.

The code for the connector is open source, and available here. The pre-build binaries are available as RPMs for RedHat systems.

To install the connector on a pre-installed SlipStream server please run: ‘yum install slipstream-connector-fco slipstream-connector-fco-python’. Once the connector is installed, you should instantiate its instances and configure them via SlipStream’s standard https:///configuration resource.

The Platform allows the user to define deployments of their complex multi-tier distributed application, specify the application specific elasticity constraints and actions to fulfil them, provide the application custom monitoring probes, provision the application on Clouds with one button click and finally let the CELAR Platform to manage and control the scalability lifecycle domain of the application.

CELAR features c-Eclipse GUI for defining application’s topology, deployment and elasticity constraints.  User application deployments can be defined in any scripting language with an ability to incorporate any of the configuration management systems (Chef, Puppet etc.).  For the details on the installation and usage of c-Eclipse see this link

User can start with an initial multiplicity per nodes in a tier, and when load grows, the different tiers of the application will be scaled up or down accordingly. The decisions for the scaling actions are made based on the evaluation of the metrics coming from VMs and application level probes. CELAR uses JCatascopia for monitoring. Here one can find the required information on how to write and configure the user application specific monitoring probes.

The decision for scaling action is made by the Decision Making Module (DMM).  At the moment, on the IaaS level CELAR supports the following scaling actions:

- add and remove VMs

Any application level actions (e.g., rebalancing of the DB cluster, etc.) can be registered along with the elasticity constraints and the monitoring probes, which will be evaluated and applied accordingly.

CELAR uses SlipStream as applications deployment provisioner and orchestrator on clouds.  Out of the box CELAR comes with the following installed and pre-configured clouds.

– Flexiant Cloud Orchestrator 

– ~Okeanos

However, the list of supported clouds is much wider and comprises all the clouds supported by SlipStream.  After the deployment of the Platform’s core service — CELAR Server — one can easily install and configure extra cloud connectors.

 ::: Deployment of the CELAR Platform :::

Here is the deployment model of the functional components of the CELAR Platform.

c-Eclipse GUI – user desktop

* for installation you can look here

CELAR Server – cloud SaaS service. Pre-deployed into VM (or physical server).

– CELAR Manager

– SlipStream server

CELAR Orchestrator VM – automatically started from CELAR Server for each user application.

– CELAR Application Orchestrator

– SlipStream Orchestrator

– CELAR Decision Making Module

– CELAR JCatascopia monitoring Server and Web

CELAR Application VM RedHat-based (CentOS) or Ubuntu – user applications get deployed into those VMs.

– CELAR JCatascopia monitoring Agent

– SlipStream Node executor

The easiest way to roll out the CELAR Platform on a cloud infrastructure would be to import the Platform appliances into the respective cloud.  Here is the page with the links to the pre-built appliances with CELAR Server, Orchestrator, Application VMs (CentOS, Ubuntu)

http://snf-175960.vm.okeanos.grnet.gr/appls/

For manual installation of the CELAR Platform and/or building of the custom appliances one can use the following set of scripts for the CELAR Server, Orchestrator, Application VMs

https://github.com/CELAR/celar-deployment

The CELAR Platform is also available from this YUM repo as a set of RPM packages

http://snf-175960.vm.okeanos.grnet.gr/yum/

After the appliances with the CELAR Platform are available for instantiation, one should provision CELAR Server VM as a service and perform the following configuration on it as described in this readme

The CELAR Orchestrator cloud image ID should be noted and configured for the corresponding cloud connector(s) in the SlipStream service running on the CELAR Server.  For more details on the configuration