VALUEGRIDS

ValueGrids opens the D-Grid-infrastructure for a fundamentally new community, the users and providers of software-based services (Saas, software-as-a-service providers). The market researchers from Gartner estimate the market share of this community for business software in 4 years being at 30%:

"With software as service (SaaS), the user organization pays for software services in proportion ton use. This is fundamentally different from the fixed price perpetual license of the traditional on-premises technology. Endorsed and promoted by all leading business applications vendors (Oracle, SAP, Microsoft) and many Web technology leaders (Google, Amazon), the SaaS model of deployment and distribution of software services wants enjoy steady growth into mainstream use during the next five years."

Already today, SaaS providers offer complex software solutions, which can be adapted comprehensively and individually to customer needs. In order to be able to react flexibly to changing markets and customer's requests, a service-oriented solution must rely on modular services of different providers. Equally, operation of such services implies a high need of computing and storage capacities. Grid technologies promise flexible and cost-efficient adjustment towards those changes in demand.

The project ValueGrids is focused on so-called value chains, which are traditionally characterized by a well-defined succession of process steps by which a superordinate goal – a creation of value – is to be achieved. With the increasing globalization, this well-defined succession nowadays more frequently is composed of rapidly changing providers. The chain thus becomes a dynamic network (value network) composed of decentralized participants.

Service Level Agreements (SLAs) form the basis for a transparent specification of the services offered by the providers as well as for the pricing. Vertically, SLAs cover all levels of an enterprise (for infrastructure providers addressed by the SLA project of the D-Grid), while horizontally, SLAs involve all participants of a value network. Therefore, a concept is to be provided allowing for seamless integration of interconnected providers into business processes as well as allows for flexible management of value networks based on service level agreements.

The challenges for SLA management result from the cooperation of different distributed enterprises, and are demonstrated in the following scenario (see figure 1): An SaaS provider (e.g. SAP, Conemis AG) offers a service-oriented application for the management of business processes of small and medium-size enterprises (SME). In order to be able to offer to customers an individually cut solution, the provider integrates additional functionality from suitable offers of decentralized ISVs (Independent Software Vendors).

The software of the SaaS provider is run in a hosted environment. The requirements of the customers (e.g. regarding quality, penalties in case of SLA breach) are defined in mutual agreements as part of the SLAs. For deployment and operation of the application, the provider takes advantage of computing and storage capacities from a Computing Cloud, which is managed and made available over a D-Grid infrastructure. Provisioning of distributed hardware resources by a D-Grid infrastructure is extended by a component, which facilitates   the management of service levels in value networks.




Figure 1: Service-Level-Management in Value Networks

Infrastructure requirements (e.g. computing capacity, memory) are dependent on various factors like e.g. number of systems, number of users and complexity of the implemented processes. In order to enable seamless operation, management of the solution has to rely on appropriately defined SLAs which are in close interdependency to SaaS providers and their respective customers.

Already this simple constellation of business relations shows the complexity of service level agreement management. Hard- and software providers assure a defined service level to SaaS providers and those, finally, provide their SLAs to the application end user. At the end of the day, those service levels represent the basis of the daily business for the customer.

The final goal of the ValueGrids is thus i) the holistic management of value networks, which are composed and orchestrated by Grid based services, and ii) the transparent coupling between the respective service agreements of value networks and on Grid infrastructure level.

It has to be identified, which kinds of services play a vital role in service-oriented economics, which kinds of IT-services may be able to support them and how these can be provided in a Grid infrastructure. For different services, different kinds of service level agreements are needed – SLA concepts for such services are developed in ValueGrids. Specifically, for complex business relations, it is also to be clarified, how SLAs and their attributes can be propagated consistently.

Besides conceptual analysis, a component is to be developed which facilitates the management of value networks in the context of Grid-based services and the D-Grid infrastructure. The so-called management cockpit will analyze and process runtime-data concerning the interweaved contract situation in the value network and will allow for the SaaS provider (by means of certain visualization techniques) to assess risks and dependencies within SLAs and between SLAs that are difficult to recognize otherwise.

The developed concepts as well as the developed management component will be evaluated in the context of an integrated industrial case study with representatives of the SaaS Community and five end customers. One result of the evaluation will be an assessment, how the developed concept can be applied in other industrial contexts and from the point of view of the customer.

With the project results of ValueGrids, the D-Grid infrastructure can be integrated into complex enterprise value networks and will contribute to a sustainable D-Grid infrastructure.