The Experts below are selected from a list of 645 Experts worldwide ranked by ideXlab platform
Joosen Wouter - One of the best experts on this subject based on the ideXlab platform.
-
Middleware for dynamic upgrade activation and compensations in multi-tenant Saas
2017Co-Authors: Van Landuyt Dimitri, Gey Fatih, Truyen Eddy, Joosen WouterAbstract:Multi-tenant Software as a Service (Saas) is the cloud computing delivery model that maximizes resource sharing up to the level of a single application instance, servicing many customer organizations (tenants) at once. Due to this scale of delivery, a Saas Offering, once successful, becomes difficult to upgrade and evolve without affecting service continuity, and this in turn limits its capabilities to respond to the reality of changing customer requirements. However, not all tenants are equal, and to some organizations such disruptions are more costly than to others. Supporting different quality trade-offs for different tenants is often a manual, error-prone task and far from trivial. This short paper outlines our middleware design for fine-grained, gradual and continuous evolution of multi-tenant Saas applications, providing automated and systematic support for (i) tenant-aware upgrade enactment, and (ii) compensations that allow recovering from negative side-effects of the upgrade enactment.status: accepte
-
Evolving multi-tenant Saas applications through self-adaptive upgrade enactment and tenant mediation
'Association for Computing Machinery (ACM)', 2016Co-Authors: Gey Fatih, Van Landuyt Dimitri, Joosen WouterAbstract:When successful, multi-tenant Saas applications service many customer organizations (tenants) at once, and Saas providers face the challenge of complying to the different SLAs of each of these tenants. As a consequence, evolving a Saas application is in practice one at run time to limit service disruptions, and preferably on a gradual, tenant-per-tenant basis, while taking into account the nature of the upgrade at hand but also these different tenant SLAs. The economic viability and cost-effectiveness of a Saas Offering depends strongly on two principles: (i) maximal automation of its operation, and (ii) self-service: allowing tenant organizations themselves to customize and configure different aspects of the service to their specific needs. In this position paper, we highlight the value of adopting the principles of self-adaptive systems in the design of middleware solutions that support continuous evolution of multi-tenant Saas applications as a means to implement the first principle. Furthermore, we discuss the additional challenges imposed by the second principle, more specifically for supporting tenant mediation, i.e. introducing human stakeholders such as tenant administrators into the inner control-loop of a self-adaptive system. We present the design of our middleware that addresses these challenges for the specific purpose of evolving multi-tenant Saas applications, but also discuss the relevance for self-adaptive systems that support stakeholder mediation in general.status: publishe
-
A research roadmap for service lines
2015Co-Authors: Van Landuyt Dimitri, Walraven Stefan, Joosen WouterAbstract:Software product line engineering (SPLE) and variability enforcement techniques have been applied to run-time adaptive systems for quite some years, also in the context of multi-tenant Software-as-a-Service (Saas) applications. The focus has been mainly on (1) the pre-deployment phases of the development life cycle and (2) fine-grained (tenant-level), run-time activation of specific variants. However, with upcoming trends such as DevOps and continuous delivery and deployment, operational aspects become increasingly important. In this paper, we present our integrated vision on the positive interplay between SPLE and adaptive middleware for multi-tenant Saas applications, focusing on the operational aspects of running and maintaining a successful Saas Offering. This vision, called Service Lines, is based on and motivated by our experience and frequent interactions with a number of Belgian Saas providers. We concretely highlight and motivate a number of operational use cases that require advanced variability support in middleware and have promising added value for the economic feasibility of Saas Offerings. In addition, we provide a gap analysis of what is currently lacking from the perspectives of variability modeling and management techniques and middleware support, and as such sketch a concrete roadmap for continued research in this area.status: publishe
-
Variability middleware for multi-tenant Saas applications
'Association for Computing Machinery (ACM)', 2015Co-Authors: Van Landuyt Dimitri, Walraven Stefan, Joosen WouterAbstract:Software product line engineering (SPLE) and variability enforcement techniques have been applied to run-time adaptive systems for quite some years, also in the context of multi-tenant Software-as-a-Service (Saas) applications. The focus has been mainly on (1) the pre-deployment phases of the development life cycle and (2) fine-grained (tenant-level), run-time activation of specific variants. However, with upcoming trends such as DevOps and continuous delivery and deployment, operational aspects become increasingly important. In this paper, we present our integrated vision on the positive interplay between SPLE and adaptive middleware for multi-tenant Saas applications, focusing on the operational aspects of running and maintaining a successful Saas Offering. This vision, called Service Lines, is based on and motivated by our experience and frequent interactions with a number of Belgian Saas providers. We concretely highlight and motivate a number of operational use cases that require advanced variability support in middleware and have promising added value for the economic feasibility of Saas Offerings. In addition, we provide a gap analysis of what is currently lacking from the perspectives of variability modeling and management techniques and middleware support, and as such sketch a concrete roadmap for continued research in this area.status: publishe
-
Continuous evolution of multi-tenant Saas applications: A customizable dynamic adaptation approach
'Institute of Electrical and Electronics Engineers (IEEE)', 2015Co-Authors: Gey Fatih, Van Landuyt Dimitri, Joosen Wouter, Jonckers VivianeAbstract:Applying application-level multi-tenancy in Software-as-a-Service (Saas) Offerings yields a number of compelling benefits: sharing a single instance of the application between large numbers of customer organizations increases cost efficiency and allows the Saas provider to attain true economies-of-scale benefits. There is however a main downside to this: increased sharing of resources causes the Saas application to be very difficult to modify after initial development and deployment without affecting service continuity: any change potentially affects the service levels promised to all enrolled tenant organizations and their end users. This rigidity is a key impediment as now the Saas provider must evolve and maintain the Saas Offering at run time, on a gradual, per-tenant basis. This in turn causes a reality of multiple co-existing versions of individual components and as such introduces substantial management complexity. To address these challenges, this paper motivates and defines key requirements that allows per-tenant, SLA-aware and gradual upgrades in the context of multi-tenant Saas applications. In addition, we define an approach that allows the involved stakeholders (tenants, Saas operators, Saas developers, etc.) to customize the dynamic enactment of upgrades, and provide a number of alternative software upgrade strategies that represent different service quality trade-offs.status: publishe
Wouter Joosen - One of the best experts on this subject based on the ideXlab platform.
-
ICSOC - Middleware for Dynamic Upgrade Activation and Compensations in Multi-tenant Saas
Service-Oriented Computing, 2017Co-Authors: Dimitri Van Landuyt, Eddy Truyen, Wouter JoosenAbstract:Multi-tenant Software as a Service (Saas) is the cloud computing delivery model that maximizes resource sharing up to the level of a single application instance, servicing many customer organizations (tenants) at once. Due to this scale of delivery, a Saas Offering, once successful, becomes difficult to upgrade and evolve without affecting service continuity, and this in turn limits its capabilities to respond to the reality of changing customer requirements.
-
SEAMS@ICSE - Evolving multi-tenant Saas applications through self-adaptive upgrade enactment and tenant mediation
Proceedings of the 11th International Workshop on Software Engineering for Adaptive and Self-Managing Systems - SEAMS '16, 2016Co-Authors: Dimitri Van Landuyt, Wouter JoosenAbstract:When successful, multi-tenant Saas applications service many customer organizations (tenants) at once, and Saas providers face the challenge of complying to the different SLAs of each of these tenants. As a consequence, evolving a Saas application is in practice done at run time to limit service disruptions, and preferably on a gradual, tenant-per-tenant basis, while taking into account the nature of the upgrade at hand but also these different tenant SLAs. The economic viability and cost-effectiveness of a Saas Offering depends strongly on two principles: (i) maximal automation of its operation, and (ii) self-service: allowing tenant organizations themselves to customize and configure different aspects of the service to their specific needs. In this position paper, we highlight the value of adopting the principles of self-adaptive systems in the design of middleware solutions that support continuous evolution of multi-tenant Saas applications as a means to implement the first principle. Furthermore, we discuss the additional challenges imposed by the second principle, more specifically for supporting tenant mediation, i.e. introducing human stakeholders such as tenant administrators into the inner control-loop of a self-adaptive system. We present the design of our middleware that addresses these challenges for the specific purpose of evolving multi-tenant Saas applications, but also discuss the relevance for self-adaptive systems that support stakeholder mediation in general.
-
UCC - Middleware for customizable multi-staged dynamic upgrades of multi-tenant Saas applications
2015Co-Authors: Dimitri Van Landuyt, Wouter JoosenAbstract:Multi-tenant Software as a Service (Saas) is the cloud computing delivery model that maximizes resource sharing up to the level of a single application instance servicing many customer organizations (tenants) at once. Due to this scale of delivery, a Saas Offering, once successful, becomes difficult to upgrade and evolve without affecting service continuity and tenant businesses profoundly. However, not all tenants are equal, and to some organizations such disruptions are more costly than to others. To account for such tenant-specific requirements, middleware for upgrading Saas applications should support tenant-specific enactment of upgrades that allow for a customizable schedule and type of enactment in accordance to the tenant SLA. In this paper, we present our design and implementation of a Saas middleware that enables run-time adaptation by means of a gradual tenant-by-tenant activation of upgrades. The adaptation mechanism is multi-staged, i.e. supports configuration based on the inputs of the tenant administrator and other stakeholders, and is maximally automated. We have validated the middleware in an OSGi-based prototype implementation and evaluated this prototype, showing negligible performance overhead of the middleware and yet clearly showcasing service continuity improvements in realistic upgrade scenarios.
-
SPLC - Variability middleware for multi-tenant Saas applications: a research roadmap for service lines
Proceedings of the 19th International Conference on Software Product Line - SPLC '15, 2015Co-Authors: Dimitri Van Landuyt, Stefan Walraven, Wouter JoosenAbstract:Software product line engineering (SPLE) and variability enforcement techniques have been applied to run-time adaptive systems for quite some years, also in the context of multi-tenant Software-as-a-Service (Saas) applications. The focus has been mainly on (1) the pre-deployment phases of the development life cycle and (2) fine-grained (tenant-level), run-time activation of specific variants. However, with upcoming trends such as DevOps and continuous delivery and deployment, operational aspects become increasingly important. In this paper, we present our integrated vision on the positive interplay between SPLE and adaptive middleware for multi-tenant Saas applications, focusing on the operational aspects of running and maintaining a successful Saas Offering. This vision, called Service Lines, is based on and motivated by our experience and frequent interactions with a number of Belgian Saas providers. We concretely highlight and motivate a number of operational use cases that require advanced variability support in middleware and have promising added value for the economic feasibility of Saas Offerings. In addition, we provide a gap analysis of what is currently lacking from the perspectives of variability modeling and management techniques and middleware support, and as such sketch a concrete roadmap for continued research in this area.
-
PESOS@ICSE - Continuous evolution of multi-tenant Saas applications: a customizable dynamic adaptation approach
2015 IEEE ACM 7th International Workshop on Principles of Engineering Service-Oriented and Cloud Systems, 2015Co-Authors: Dimitri Van Landuyt, Wouter Joosen, Viviane JonckersAbstract:Applying application-level multi-tenancy in Software-as-a-Service (Saas) Offerings yields a number of compelling benefits: sharing a single instance of the application between large numbers of customer organizations increases cost efficiency and allows the Saas provider to attain true economies-of-scale benefits. There is however a main downside to this: increased sharing of resources causes the Saas application to be very difficult to modify after initial development and deployment without affecting service continuity: any change potentially affects the service levels promised to all enrolled tenant organizations and their end users. This rigidity is a key impediment as now the Saas provider must evolve and maintain the Saas Offering at run time, on a gradual, per-tenant basis. This in turn causes a reality of multiple co-existing versions of individual components and as such introduces substantial management complexity. To address these challenges, this paper motivates and defines key requirements that allows per-tenant, SLA-aware and gradual upgrades in the context of multi-tenant Saas applications. In addition, we define an approach that allows the involved stakeholders (tenants, Saas operators, Saas developers, etc.) to customize the dynamic enactment of upgrades, and provide a number of alternative software upgrade strategies that represent different service quality trade-offs.
Van Landuyt Dimitri - One of the best experts on this subject based on the ideXlab platform.
-
Middleware for dynamic upgrade activation and compensations in multi-tenant Saas
2017Co-Authors: Van Landuyt Dimitri, Gey Fatih, Truyen Eddy, Joosen WouterAbstract:Multi-tenant Software as a Service (Saas) is the cloud computing delivery model that maximizes resource sharing up to the level of a single application instance, servicing many customer organizations (tenants) at once. Due to this scale of delivery, a Saas Offering, once successful, becomes difficult to upgrade and evolve without affecting service continuity, and this in turn limits its capabilities to respond to the reality of changing customer requirements. However, not all tenants are equal, and to some organizations such disruptions are more costly than to others. Supporting different quality trade-offs for different tenants is often a manual, error-prone task and far from trivial. This short paper outlines our middleware design for fine-grained, gradual and continuous evolution of multi-tenant Saas applications, providing automated and systematic support for (i) tenant-aware upgrade enactment, and (ii) compensations that allow recovering from negative side-effects of the upgrade enactment.status: accepte
-
Evolving multi-tenant Saas applications through self-adaptive upgrade enactment and tenant mediation
'Association for Computing Machinery (ACM)', 2016Co-Authors: Gey Fatih, Van Landuyt Dimitri, Joosen WouterAbstract:When successful, multi-tenant Saas applications service many customer organizations (tenants) at once, and Saas providers face the challenge of complying to the different SLAs of each of these tenants. As a consequence, evolving a Saas application is in practice one at run time to limit service disruptions, and preferably on a gradual, tenant-per-tenant basis, while taking into account the nature of the upgrade at hand but also these different tenant SLAs. The economic viability and cost-effectiveness of a Saas Offering depends strongly on two principles: (i) maximal automation of its operation, and (ii) self-service: allowing tenant organizations themselves to customize and configure different aspects of the service to their specific needs. In this position paper, we highlight the value of adopting the principles of self-adaptive systems in the design of middleware solutions that support continuous evolution of multi-tenant Saas applications as a means to implement the first principle. Furthermore, we discuss the additional challenges imposed by the second principle, more specifically for supporting tenant mediation, i.e. introducing human stakeholders such as tenant administrators into the inner control-loop of a self-adaptive system. We present the design of our middleware that addresses these challenges for the specific purpose of evolving multi-tenant Saas applications, but also discuss the relevance for self-adaptive systems that support stakeholder mediation in general.status: publishe
-
A research roadmap for service lines
2015Co-Authors: Van Landuyt Dimitri, Walraven Stefan, Joosen WouterAbstract:Software product line engineering (SPLE) and variability enforcement techniques have been applied to run-time adaptive systems for quite some years, also in the context of multi-tenant Software-as-a-Service (Saas) applications. The focus has been mainly on (1) the pre-deployment phases of the development life cycle and (2) fine-grained (tenant-level), run-time activation of specific variants. However, with upcoming trends such as DevOps and continuous delivery and deployment, operational aspects become increasingly important. In this paper, we present our integrated vision on the positive interplay between SPLE and adaptive middleware for multi-tenant Saas applications, focusing on the operational aspects of running and maintaining a successful Saas Offering. This vision, called Service Lines, is based on and motivated by our experience and frequent interactions with a number of Belgian Saas providers. We concretely highlight and motivate a number of operational use cases that require advanced variability support in middleware and have promising added value for the economic feasibility of Saas Offerings. In addition, we provide a gap analysis of what is currently lacking from the perspectives of variability modeling and management techniques and middleware support, and as such sketch a concrete roadmap for continued research in this area.status: publishe
-
Variability middleware for multi-tenant Saas applications
'Association for Computing Machinery (ACM)', 2015Co-Authors: Van Landuyt Dimitri, Walraven Stefan, Joosen WouterAbstract:Software product line engineering (SPLE) and variability enforcement techniques have been applied to run-time adaptive systems for quite some years, also in the context of multi-tenant Software-as-a-Service (Saas) applications. The focus has been mainly on (1) the pre-deployment phases of the development life cycle and (2) fine-grained (tenant-level), run-time activation of specific variants. However, with upcoming trends such as DevOps and continuous delivery and deployment, operational aspects become increasingly important. In this paper, we present our integrated vision on the positive interplay between SPLE and adaptive middleware for multi-tenant Saas applications, focusing on the operational aspects of running and maintaining a successful Saas Offering. This vision, called Service Lines, is based on and motivated by our experience and frequent interactions with a number of Belgian Saas providers. We concretely highlight and motivate a number of operational use cases that require advanced variability support in middleware and have promising added value for the economic feasibility of Saas Offerings. In addition, we provide a gap analysis of what is currently lacking from the perspectives of variability modeling and management techniques and middleware support, and as such sketch a concrete roadmap for continued research in this area.status: publishe
-
Continuous evolution of multi-tenant Saas applications: A customizable dynamic adaptation approach
'Institute of Electrical and Electronics Engineers (IEEE)', 2015Co-Authors: Gey Fatih, Van Landuyt Dimitri, Joosen Wouter, Jonckers VivianeAbstract:Applying application-level multi-tenancy in Software-as-a-Service (Saas) Offerings yields a number of compelling benefits: sharing a single instance of the application between large numbers of customer organizations increases cost efficiency and allows the Saas provider to attain true economies-of-scale benefits. There is however a main downside to this: increased sharing of resources causes the Saas application to be very difficult to modify after initial development and deployment without affecting service continuity: any change potentially affects the service levels promised to all enrolled tenant organizations and their end users. This rigidity is a key impediment as now the Saas provider must evolve and maintain the Saas Offering at run time, on a gradual, per-tenant basis. This in turn causes a reality of multiple co-existing versions of individual components and as such introduces substantial management complexity. To address these challenges, this paper motivates and defines key requirements that allows per-tenant, SLA-aware and gradual upgrades in the context of multi-tenant Saas applications. In addition, we define an approach that allows the involved stakeholders (tenants, Saas operators, Saas developers, etc.) to customize the dynamic enactment of upgrades, and provide a number of alternative software upgrade strategies that represent different service quality trade-offs.status: publishe
Dimitri Van Landuyt - One of the best experts on this subject based on the ideXlab platform.
-
ICSOC - Middleware for Dynamic Upgrade Activation and Compensations in Multi-tenant Saas
Service-Oriented Computing, 2017Co-Authors: Dimitri Van Landuyt, Eddy Truyen, Wouter JoosenAbstract:Multi-tenant Software as a Service (Saas) is the cloud computing delivery model that maximizes resource sharing up to the level of a single application instance, servicing many customer organizations (tenants) at once. Due to this scale of delivery, a Saas Offering, once successful, becomes difficult to upgrade and evolve without affecting service continuity, and this in turn limits its capabilities to respond to the reality of changing customer requirements.
-
SEAMS@ICSE - Evolving multi-tenant Saas applications through self-adaptive upgrade enactment and tenant mediation
Proceedings of the 11th International Workshop on Software Engineering for Adaptive and Self-Managing Systems - SEAMS '16, 2016Co-Authors: Dimitri Van Landuyt, Wouter JoosenAbstract:When successful, multi-tenant Saas applications service many customer organizations (tenants) at once, and Saas providers face the challenge of complying to the different SLAs of each of these tenants. As a consequence, evolving a Saas application is in practice done at run time to limit service disruptions, and preferably on a gradual, tenant-per-tenant basis, while taking into account the nature of the upgrade at hand but also these different tenant SLAs. The economic viability and cost-effectiveness of a Saas Offering depends strongly on two principles: (i) maximal automation of its operation, and (ii) self-service: allowing tenant organizations themselves to customize and configure different aspects of the service to their specific needs. In this position paper, we highlight the value of adopting the principles of self-adaptive systems in the design of middleware solutions that support continuous evolution of multi-tenant Saas applications as a means to implement the first principle. Furthermore, we discuss the additional challenges imposed by the second principle, more specifically for supporting tenant mediation, i.e. introducing human stakeholders such as tenant administrators into the inner control-loop of a self-adaptive system. We present the design of our middleware that addresses these challenges for the specific purpose of evolving multi-tenant Saas applications, but also discuss the relevance for self-adaptive systems that support stakeholder mediation in general.
-
UCC - Middleware for customizable multi-staged dynamic upgrades of multi-tenant Saas applications
2015Co-Authors: Dimitri Van Landuyt, Wouter JoosenAbstract:Multi-tenant Software as a Service (Saas) is the cloud computing delivery model that maximizes resource sharing up to the level of a single application instance servicing many customer organizations (tenants) at once. Due to this scale of delivery, a Saas Offering, once successful, becomes difficult to upgrade and evolve without affecting service continuity and tenant businesses profoundly. However, not all tenants are equal, and to some organizations such disruptions are more costly than to others. To account for such tenant-specific requirements, middleware for upgrading Saas applications should support tenant-specific enactment of upgrades that allow for a customizable schedule and type of enactment in accordance to the tenant SLA. In this paper, we present our design and implementation of a Saas middleware that enables run-time adaptation by means of a gradual tenant-by-tenant activation of upgrades. The adaptation mechanism is multi-staged, i.e. supports configuration based on the inputs of the tenant administrator and other stakeholders, and is maximally automated. We have validated the middleware in an OSGi-based prototype implementation and evaluated this prototype, showing negligible performance overhead of the middleware and yet clearly showcasing service continuity improvements in realistic upgrade scenarios.
-
SPLC - Variability middleware for multi-tenant Saas applications: a research roadmap for service lines
Proceedings of the 19th International Conference on Software Product Line - SPLC '15, 2015Co-Authors: Dimitri Van Landuyt, Stefan Walraven, Wouter JoosenAbstract:Software product line engineering (SPLE) and variability enforcement techniques have been applied to run-time adaptive systems for quite some years, also in the context of multi-tenant Software-as-a-Service (Saas) applications. The focus has been mainly on (1) the pre-deployment phases of the development life cycle and (2) fine-grained (tenant-level), run-time activation of specific variants. However, with upcoming trends such as DevOps and continuous delivery and deployment, operational aspects become increasingly important. In this paper, we present our integrated vision on the positive interplay between SPLE and adaptive middleware for multi-tenant Saas applications, focusing on the operational aspects of running and maintaining a successful Saas Offering. This vision, called Service Lines, is based on and motivated by our experience and frequent interactions with a number of Belgian Saas providers. We concretely highlight and motivate a number of operational use cases that require advanced variability support in middleware and have promising added value for the economic feasibility of Saas Offerings. In addition, we provide a gap analysis of what is currently lacking from the perspectives of variability modeling and management techniques and middleware support, and as such sketch a concrete roadmap for continued research in this area.
-
PESOS@ICSE - Continuous evolution of multi-tenant Saas applications: a customizable dynamic adaptation approach
2015 IEEE ACM 7th International Workshop on Principles of Engineering Service-Oriented and Cloud Systems, 2015Co-Authors: Dimitri Van Landuyt, Wouter Joosen, Viviane JonckersAbstract:Applying application-level multi-tenancy in Software-as-a-Service (Saas) Offerings yields a number of compelling benefits: sharing a single instance of the application between large numbers of customer organizations increases cost efficiency and allows the Saas provider to attain true economies-of-scale benefits. There is however a main downside to this: increased sharing of resources causes the Saas application to be very difficult to modify after initial development and deployment without affecting service continuity: any change potentially affects the service levels promised to all enrolled tenant organizations and their end users. This rigidity is a key impediment as now the Saas provider must evolve and maintain the Saas Offering at run time, on a gradual, per-tenant basis. This in turn causes a reality of multiple co-existing versions of individual components and as such introduces substantial management complexity. To address these challenges, this paper motivates and defines key requirements that allows per-tenant, SLA-aware and gradual upgrades in the context of multi-tenant Saas applications. In addition, we define an approach that allows the involved stakeholders (tenants, Saas operators, Saas developers, etc.) to customize the dynamic enactment of upgrades, and provide a number of alternative software upgrade strategies that represent different service quality trade-offs.
Gey Fatih - One of the best experts on this subject based on the ideXlab platform.
-
Middleware for dynamic upgrade activation and compensations in multi-tenant Saas
2017Co-Authors: Van Landuyt Dimitri, Gey Fatih, Truyen Eddy, Joosen WouterAbstract:Multi-tenant Software as a Service (Saas) is the cloud computing delivery model that maximizes resource sharing up to the level of a single application instance, servicing many customer organizations (tenants) at once. Due to this scale of delivery, a Saas Offering, once successful, becomes difficult to upgrade and evolve without affecting service continuity, and this in turn limits its capabilities to respond to the reality of changing customer requirements. However, not all tenants are equal, and to some organizations such disruptions are more costly than to others. Supporting different quality trade-offs for different tenants is often a manual, error-prone task and far from trivial. This short paper outlines our middleware design for fine-grained, gradual and continuous evolution of multi-tenant Saas applications, providing automated and systematic support for (i) tenant-aware upgrade enactment, and (ii) compensations that allow recovering from negative side-effects of the upgrade enactment.status: accepte
-
Evolving multi-tenant Saas applications through self-adaptive upgrade enactment and tenant mediation
'Association for Computing Machinery (ACM)', 2016Co-Authors: Gey Fatih, Van Landuyt Dimitri, Joosen WouterAbstract:When successful, multi-tenant Saas applications service many customer organizations (tenants) at once, and Saas providers face the challenge of complying to the different SLAs of each of these tenants. As a consequence, evolving a Saas application is in practice one at run time to limit service disruptions, and preferably on a gradual, tenant-per-tenant basis, while taking into account the nature of the upgrade at hand but also these different tenant SLAs. The economic viability and cost-effectiveness of a Saas Offering depends strongly on two principles: (i) maximal automation of its operation, and (ii) self-service: allowing tenant organizations themselves to customize and configure different aspects of the service to their specific needs. In this position paper, we highlight the value of adopting the principles of self-adaptive systems in the design of middleware solutions that support continuous evolution of multi-tenant Saas applications as a means to implement the first principle. Furthermore, we discuss the additional challenges imposed by the second principle, more specifically for supporting tenant mediation, i.e. introducing human stakeholders such as tenant administrators into the inner control-loop of a self-adaptive system. We present the design of our middleware that addresses these challenges for the specific purpose of evolving multi-tenant Saas applications, but also discuss the relevance for self-adaptive systems that support stakeholder mediation in general.status: publishe
-
Continuous evolution of multi-tenant Saas applications: A customizable dynamic adaptation approach
'Institute of Electrical and Electronics Engineers (IEEE)', 2015Co-Authors: Gey Fatih, Van Landuyt Dimitri, Joosen Wouter, Jonckers VivianeAbstract:Applying application-level multi-tenancy in Software-as-a-Service (Saas) Offerings yields a number of compelling benefits: sharing a single instance of the application between large numbers of customer organizations increases cost efficiency and allows the Saas provider to attain true economies-of-scale benefits. There is however a main downside to this: increased sharing of resources causes the Saas application to be very difficult to modify after initial development and deployment without affecting service continuity: any change potentially affects the service levels promised to all enrolled tenant organizations and their end users. This rigidity is a key impediment as now the Saas provider must evolve and maintain the Saas Offering at run time, on a gradual, per-tenant basis. This in turn causes a reality of multiple co-existing versions of individual components and as such introduces substantial management complexity. To address these challenges, this paper motivates and defines key requirements that allows per-tenant, SLA-aware and gradual upgrades in the context of multi-tenant Saas applications. In addition, we define an approach that allows the involved stakeholders (tenants, Saas operators, Saas developers, etc.) to customize the dynamic enactment of upgrades, and provide a number of alternative software upgrade strategies that represent different service quality trade-offs.status: publishe
-
Middleware for customizable multi-staged dynamic upgrades for multi-tenant Saas applications
'Institute of Electrical and Electronics Engineers (IEEE)', 2015Co-Authors: Gey Fatih, Van Landuyt Dimitri, Joosen WouterAbstract:© 2015 IEEE. Multi-tenant Software as a Service (Saas) is the cloud computing delivery model that maximizes resource sharing up to the level of a single application instance servicing many customer organizations (tenants) at once. Due to this scale of delivery, a Saas Offering, once successful, becomes difficult to upgrade and evolve without affecting service continuity and tenant businesses profoundly. However, not all tenants are equal, and to some organizations such disruptions are more costly than to others. To account for such tenant-specific requirements, middleware for upgrading Saas applications should support tenant-specific enactment of upgrades that allow for a customizable schedule and type of enactment in accordance to the tenant SLA. In this paper, we present our design and implementation of a Saas middleware that enables run-time adaptation by means of a gradual tenant-by-tenant activation of upgrades. The adaptation mechanism is multi-staged, i.e. supports configuration based on the inputs of the tenant administrator and other stakeholders, and is maximally automated. We have validated the middleware in an OSGi-based prototype implementation and evaluated this prototype, showing negligible performance overhead of the middleware and yet clearly showcasing service continuity improvements in realistic upgrade scenarios.status: publishe