OCCIware Use Case : Datacenter as a Service

Use Case Project Leader: Scalair. 


Deliver an elastic Cloud/MultiCloud infrastructure with fast deployment and reconfiguration resources for hosted applications.


Development of the elasticity manager, that takes inputs from the Monitoring System and Roboconf.

Use Case Description

Scalair provides a complete platform (e.g., load balancer, fronts, BDD, etc.) that can adapt according to the incoming flow of visitors (load work). 

The platform developers can benefit from a self-scalable system according to two predefined strategies: 

- Strategy one:  Management of horizontal and multi-platform scalability, i.e., the multiplication of virtual machines/containers on n-tier architectures with a transparent and autonomous load balancing for developers to absorb traffic spikes contingencies. Virtual machines/containers can be spread over several cloud platforms to increase availability level. 

- Strategy two: Managing vertical scalability (increase in existing resources) to address peak loads on virtual machines/containers.

These strategies are required in several application domains, including carpooling: it can be highly stressed during weekends, making it necessary to have an appropriate and scalable infrastructure. In a typical cloud infrastructure, resources are provisioned in advance and often manually adjusted. During a load increase, resources are adjusted to meet the needs but without any anticipation and often at the explicit request of the client. The carpool use case is developed with data from Ecolutis/SNCF. 

OCCIware Datacenter Use Case

What are the specific components involved in the carpool application to ensure seamless application scalability?


Elasticity Manager is able to monitor the inflows and the behavior of the different servers. The supervision of these elements enables the platform to anticipate request's overload and make decisions to increase or to add additional resources. The Elasticity Manager makes decisions based on Monitoring Manager outputs.

Roboconf is able to provide hot (re)configuration of application nodes. For example, when a new node appears or disappears due to elasticity decisions (e.g. load-balancing, etc.), the configuration of dependent application nodes will be updated. The process is asynchronous, and depends only on the application topology for example, when autonomic decisions are made by ProActive cloud manager (it decides to add or remove a node), Roboconf performs the adequate application reconfiguration, so the global application still works.

ProActive Cloud Automation is able to make decisions concerning elasticity (e.g., add or remove nodes), based on criteria like application load, latency, SLA, etc., using a decision-making algorithm. Then, it will be able to make these decisions operational, and change the application topology to adapt to the new situation.