Centralized Check Recognition Workflow
Centralized image processing remains a standard methodology for many financial institutions. Even with distributed capture, most financial institutions centralize a wide range of functions to streamline manual processes and retain a consolidated control point for processing. In a centralized environment, check recognition becomes a critical component and must provide a platform that is scalable, redundant, robust, maintainable and fit within the IT constraints of both large and small financial institutions.
Manager/Server Architecture
Orbograph provides several architectures for end users to select from. The OrboCAR Accura XV centralized architecture can be viewed as hierarchical or multilayered in nature. The ability to apply multiple computing resources (CPU’s) at several layers in a configuration allows for virtually unlimited scalability. Additionally, each layer is completely transparent to the check processing application as all CPU optimization is controlled by the workflow of the OrboCAR system. The importance of redundancy cannot be underestimated as well, and such a solution eliminates any single point of failure, insuring virtually 100% uptime. Note that with smaller scale installations, certain layers are not necessarily required to attain adequate system robustness.
The evolution of CPU’s have complemented well the OrboCAR Accura XV architecture. As the number of “cores” have increased for each chipset, the innate nature of the load leveling capabilities in the Orbograph solution fully utilizes the CPU throughput capacities of these machines.
Blade Servers have been benchmarked and are fully supported with this architecture. Typically the Blade chassis will house up to 16 Blade Servers. Internal to the system, each Blade Server can be assigned to a specific Cluster as either an Manger or Server. When using a Blade chassis, the RBox can be an external machine to the Blade Chassis. Note that “virtual environments” or VM’s are fully supported.
SOA (Services Oriented Architecture)
The highly scalable G6 Enterprise Recognition Technology provides the infrastructure for an enterprise recognition resource for multiple applications utilizing any of the recognition-centric applications simultaneously. Additionally, interface is the same, whether the system is activated as a local recognition engine or remotely as an IIS service.
The system uses a flexible configuration methodology for many of the internal components, allowing for recognition services to be run across multiple systems, supporting virtualization. Because most customers run the recognition process as a mission critical application, complete redundancy can be configured.
The G6 Platform achieves high levels of scalability by using an SOA multilayered architecture, exposing Web Services. Using HTTP/HTTPS as its protocol of communication, The G6 Platform is transparently accessible for local client applications over the LAN as well as remote applications over the WAN or Internet. When using IIS as the primary service for thin client processing, proxies are available in C# and Java. Using HTTP makes the interface highly resilient to network noise and disconnections, and facilitates easy reconnection of jobs in the event of network failures. This is particularly important when utilizing the system in a distributed environment where high resiliency is a requirement in a branch.
Relevant Orbograph Products
|