|
|
| [Participants](participants) | [Scope](themes) | [Publications](publications) | [Contact](contact) |
|
|
|
|----|----|----|----|
|
|
|
|
|
|
# MAPE (monitor, analysis, planning, execution) in cloudified networks (Mgmt)
|
|
|
|
|
|
In Cloud Computing, one of the goals of component based software development is to facilitate re-utilization of software elements through the composition of loosely couple entities. Component based applications are naturally designed to make them reconfigurable and adaptable through simple modification of bindings. Several component models have been proposed and used in this regard. In many cases, however, the reconfiguration process must be done offline and requires stopping and redeploying the entire application.
|
|
|
|
|
|
The separation of concerns between functional (F) and non-functional (NF) parts allows us to design the two behaviors in separate ways and delay the integration of the two aspects until deployment time, this is the same separation proposed, in this case, by SDN networks: This decoupling permits the management concerns to be designed by an expert that does not need to know about the functional behavior, and, conversely, the programmer of the functional part needs not to focus on the non-functional concerns.
|
|
|
|
|
|
As in Software Defined Networks, in cloud components concerns like composition, bindings, monitoring, reconfiguration, security, load balancing, and in general self-management behaviors can be designed without having to know the details of the functional part, focusing only on the component management. Most of the NF reconfiguration frameworks are based in the Monitor/ Analyze/ Planner/ Execute (MAPE) model for autonomic computing, firstly presented by IBM in their work “An architectural blueprint for autonomic computing”.
|
|
|
|
|
|
Given the experience brought by the partner from Chili in both areas (see section B7), that is, measuring of QoS from mobile devices, and the research and development of MAPE in Cloud Components, the proposal is to study the applicability of the MAPE framework in cloudified networks, obtaining a recommendation of pros and cons of applying this framework for cloudified virtual autonomic behaviors.
|
|
|
|
|
|
|