SynfiniWay Usage Scenario
SynfiniWay may be deployed for a wide range of situations, for a number of users ranging from less than 5 to more than 1000. The infrastructure may be a simple as a LAN comprising desktops and server, to multiple WANs connected through any number of firewalls and public networks.
For almost any IT business process SynfiniWay can be used to improve the flow of data and information across local, distributed or grid resources. The result is to increase the speed with which knowledge is produced by these processes, while making your overall IT assets more efficient. Some possible uses of SynfiniWay include the following:
- Pre-installed within a Linux cluster server to provide a simplified and robust user interface from the desktop to applications running across the cluster nodes, allocated by a standard batch system.
- Front-end for ASP (Application Service Provider) usage.
- Application chaining to link dependent tasks by an encoded workflow process. Translation of data between tasks will be done by appropriate tools.
- Automated data flow management to deliver files and results to file servers and desktops when needed. Used to minimise delays within workflow-encoded processes.
- Loop-back workflows for self-repair processes. Failed tasks, or series of tasks, can be checked for completion, and re-rerun if necessary.
- Optimisation loops with different types of parametric variation. Allows the study of a design space, or the stochastic variation of particular inputs.
Outlined below is an implementation of SynfiniWay that can be found at an existing customer. It highlights several features of SynfiniWay that contribute to increasing the efficiency of an existing process – hardware availability, data flow automation – and provide the basis for further innovation – application chaining, optimisation loops.
Chaining applications across distributed resource grid
The business process above requires the temporal sequencing of a series of applications, connected by data translation between outputs and inputs. At run time each tool might execute on a different machine for optimal resource load-balancing. An optional loop mechanism checks the quality of the result and resumes the process if the result is not sufficiently accurate.
This process maps to the SynfiniWay workflow shown on the right. The applications and tools are encapsulated as services, and implicit data flow is managed between each service.
When implemented within a complete SynfiniWay environment the layout of tasks is as shown below. Data movement is not shown as this is completely transparent to the end-user. After the workflow is launched from a client desktop, the process continues until the looping condition is resolved. Several clients are shown as multiple end-user may connect at any time to monitor the progress of the workflow.
