Broker@Cloud: Enabling Continuous Quality Assurance and Optimization in Future Enterprise Cloud Service Brokers

Kahina Hamadache, Stelios Pantelopoulos
Singular Logic
Topics recommended for the 2016-2017 Work Programme: 

Highly Dynamic Cloud Infrastructure: as the Cloud becomes more and more omnipresent and intent to serve an even larger audience, targeting large companies as well as isolated user, the necessity for Cloud Services to frequently adapt and provide a large spectrum of performances is arising. In this perspective, it will be necessary for cloud service providers to have fine-grained and highly dynamic, fast down- and up-scalable capacities. Cloud Services providers already have the ability to scale up or down their services, but in order to remain competitive in a more fiercer and fiercer market, they will need such highly dynamic mechanisms. Cloud and Internet-of-Things: the forthcoming and foretold explosion of the Internet-of-Things will bring new challenges for the cloud. Indeed, if current cloud services consumers are already numerous, the Internet of Things aims at connecting everyday life’s objects to the web and making them smarter through data sharing and adaptation. This will inevitably imply an important increase on the consumption and needs for computing power, storage, networking capabilities, etc. In this perspective it would be interesting to consider how this overhead will impact the cloud service providers, and if a differentiation needs to be made between the human and “things” use of cloud services. Certification mechanisms for Cloud Services: as the Cloud Market is growing bigger, the number of actors is also growing, making it even more difficult for services consumers to choose a service or another. To tackle this issue, research on cloud computing has considered the existing reputation techniques and is exploring specific trails. A potential perspective for the evolution of the cloud would be to deliver certificates to trusted cloud vendors and brokers, based on their sustained performance, respect of SLAs and overall efficiency. Such approach is perfectly in line with the outcomes of the Broker@Cloud project, as project’s results should be used as a basis to allow such sustainability and meet certifications requirements.

Projects major results: 

Multiple dissemination activities are scheduled: participation to ESOCC 2014 workshop [3], CrossCloud Brokers [4], CloudCom 2014 [5] The project has reached an important implementation milestone as the different cloud service description and interfaces components of the framework are now available (internally). The continuous quality assurance and optimisation mechanism are currently being finalised and will be soon available. On this aspect, the main contribution will be the release of the PuLSar (Preference cLoud Service Recommender) component built on the top of SMICloud [6] and using Analytic Hierarchy Process to reach decisions. Failure prevention and recovery asset is also under implementation. The DiVA reasoner [7], at the centre of the module has been re-implemented in Java for portability, simplicity and performance concerns.

Potential exploitation strategy: 

The consortium has worked for a long time now on the elaboration of a consistent and efficient exploitation strategy, allowing a wide diffusion and an optimisation of profits, spin-offs and else. In this perspective, it was decided some months ago to propose the implemented results as open sources modules under Apache [1] Dual License. Hence, organisations wishing to reuse Broker@Cloud’s code and publish their developments as Open Source will be granted a free license, while organisations using Broker@Cloud code within a proprietary non-open source software will have to pay for a non-free license. The consortium is currently working on two distinct topics: harmonising the individual efforts and exploitation strategies of the different partners; and in a second time it is working on the release schedule of the different modules implemented through the project. This last point is especially important as they proposed modules may face competition from similar projects. In this perspective the consortium is also actively working on the actions to undertake in case a similar product is released before the readiness of Broker@Cloud assets.

An update since the last Concertation meeting (March 2014): 

A list of projects closely related to Broker@Cloud interests has been created to help considering the potential bridges and collaborations with B@C. Among the final list of six projects (CAST, DIVA, FIWARE, MODAClouds, MOTIVE and PaaSage), 2 of them have been identified as being sufficiently close to foresee a collaboration as they share some of B@C concerns (cloud service description, cloud service failure prevention and recovery, optimisation of cloud services) : MODAClouds and PaaSage. It is expected that the bridging between the different projects will be eased by the participation of one of the consortium member (SINTEF) in all of them, smoothly ensuring the liaison and coordinated efforts. In this perspective a shared workshop between PaaSage and Broker@Cloud will be held at Cloud Services Brokerage 2014 [2] in order to refine the cooperation points between the projects and plan the future actions. CSB 2014 will also be a valuable opportunity for the consortium to make contact with additional projects as well as consider the latest evolution on Cloud Brokerage research and estimate the impact of the project.