Handling Tenant-Specific Non-Functional Requirements through a Generic SLA

Aouzal, Khadija; Hafiddi, Hatim; Dahchour, Mohamed

doi:10.5220/0006791403830391

Cited by 1 publication

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Domain Engineering process aims at defining the commonalities and variabilities regarding an applicationoraservice.Inthecontextofnon-functionalvariabilityinSaaSservices,thisprocess beginswithadomainanalysistoretrievenon-functionalrequirementsofeachtenant.Thesenon-functionalrequirementsarethendealtwithasfeaturesandmodeledusingtheEFM.Thislatterisa tree-structureddiagramthatmodelsqualityattributesasmandatoryoroptionalfeaturesalongwith possibleattributesspecifyingextra-propertiesofthesefeatures,andthatspecifiestheconstraints between the features. In the proposed approach, the EFM is in conformance with NFVariability metamodel defined in a previous work (Aouzal et al, 2018). This metamodel allows modeling QoScharacteristicsasvariablefeaturesthatcanbeeithermandatoryoroptional,andtheinter-and intra-dependenciesbetweenthem.Figure2showstheinstanceofthismetamodel,theEFM,forthe documentmanagementapplication.InthisEFM,bothcoarse-grainedandfine-grainedvariabilityare identified.Obviously,theEFMisconstructedforeachservicetheapplicationiscomposedof.Due tosimplicityreasons,thedocumentmanagementapplicationisconsideredasoneservice.…”

Section: Domain Engineeringmentioning

confidence: 99%

Policy-Driven Middleware for Multi-Tenant SaaS Services Configuration

Aouzal

Hafiddi

Dahchour

2019

International Journal of Cloud Applications and Computing

Self Cite

View full text Add to dashboard Cite

The multi-tenancy architecture allows software-as-a-service applications to serve multiple tenants with a single instance. This is beneficial as it leverages economies of scale. However, it does not cope with the specificities of each tenant and their variability; notably, the variability induced in the required quality levels that differ from a tenant to another. Hence, sharing one single instance hampers the fulfillment of these quality levels for all the tenants and leads to service level agreement violations. In this context, this article proposes a policy-driven middleware that configures the service according to the non-functional requirements of the tenants. The adopted approach combines software product lines engineering and model driven engineering principles. It spans the quality attributes lifecycle, from documenting them to annotating the service components with them as policies, and it enables dynamic configuration according to service level agreements terms of the tenants.

show abstract

Section: Domain Engineeringmentioning

confidence: 99%