FinGRID: Grid Computing in the Financial Service Industry
Keywords: Grid Computing, Service-Oriented Computing, Virtualization, Dynamic Provisioning, Financial Grid Services
Today’s trend away from static preplanning of IT resources with a fixed allocation of IT resources towards an efficiently operating IT infrastructure that is working automatically through sharing and dynamic assignment of resources will increase. With IT being a key driver of economic and financial processes, sophisticated service-oriented Grid applications can be the answer to the increasing demand for cutting costs of operations management while making existing or even new business processes more effective.
In general, Grid computing can be defined as the pooling of IT resources into a set of shared services. Within this context, the demand for Grid resources has to be evaluated continuously to be supplied as required. From a user’s perspective, it is only important that the appropriate compute power is delivered on-demand. There is no concern for where data resides or what specific server processes the respective request. This also holds for the financial service industry. For example, the IT department of a bank serves many internal business departments. For these business departments, there is no need to understand the underlying technology of the IT resources they require. The research clusters of the FinGrid project are shown in the following diagram.
Fig. 1 Clusters of the FinGRID Project
The FinGRID consortium has selected three applications as use cases for investigating different issues associated with adopting Grid computing technology in the financial service domain:
- Grid resource pricing for the financial service industry
- Customer portfolio performance measurements
- Asset-backed security Grid factory
The goal of our work in FinGrid is to realize virtualization and dynamic provisioning as two core principles that distinguish Grid computing from other forms of computing.Virtualization means that individual resources (servers, storage, etc.) are pooled by type and then made available as services. The association of computing requirement to physical server is flexible, allowing workloads to change their profile over their usage lifecycle. Provisioning implies that when virtualized services are requested, the appropriate resources can be identified and made available automatically. Equally, when the resource is no longer needed, the physical server can be re-provisioned for a different purpose, allowing vastly better usage of the asset over its whole useful lifecycle.
Furthermore, our working group has the twofold goal of integrating standard Grid security mechanisms into FinGRID and of investigating the unique security challenges posed by the on-demand utilization of Grid resources in the financial service sector and developing prototypical solutions to handle these new threats. The main objective is the development of a secure container, into which financial service applications can be deployed on-demand without interrupting running applications. The containers should constitute a safe working environment, in which sensitive data can be stored or computed without any risk; no user is allowed to gain knowledge about another party operating on the same resource. To achieve this requirement, the operating system virtualization techniques will be employed. Another issue is to integrate the login and authentication methods for the virtual instances into the existing certificate infrastructure, to limit access to known and allowed persons. For this purpose, the current login/certification schemes and the solutions for managing virtual organizations provided by the D-Grid Initiative must be analyzed, and it must be investigated how they can be adapted to suit the virtualization needs.
Partners: University of Frankfurt, University of Siegen, University of Stuttgart, Deutsche Bank, Dresdner Bank, IBM, PA Consulting, FinanzIT, Data Synapse
Funding: Bundesministerium für Bildung und Forschung, D-Grid Inititiative