Service Accounting (Chargeback) Drivers
In a Real-Time Infrastructure (RTI), supply and demand are balanced via a real time feedback loop of metered consumption, performance monitoring against SLA’s, and dynamic resource allocation. RTI’s have the following attributes:
- Business demand drives dynamic and automated responses from IT, enhancing IT’s alignment with the BU’s while increasing the agility of the business
- Services are delivered as a utility, and IT is responsible for optimizing costs, performance, and utilization subject to Service Level Agreements
- Usage is metered, and accounting is based on actual consumption
The large-scale adoption of RTI-based services leads to increased efficiencies, agility and cost structure. Often, the decision to launch additional new services will only occur when the savings can be passed on directly through chargebacks – with the tipping point being the institution of a utility operating model, where well-defined and measured services are shared throughout the enterprise.
Evident ClearStone: Service Accounting for the Grid Utility
Evident ClearStone provides Service Accounting capabilities that allow customers to implement cost accounting, including full chargeback, for shared services running on a grid utility within a Real-Time Infrastructure. In heterogeneous grid environments, the ability to charge for the time consumed by various grid services eases the adoption a shared utility and helps realize the economies that are associated with highly optimized IT resources.
The Evident ClearStone Service Accounting capabilities currently uses the measurement of “elapsed time” for tasks executed on engines as the metric that reflects how much compute resources were used by a specific service running on the grid. The Evident ClearStone Pipeline Server uses specialized pipelines that automatically aggregate the elapse task time for all tasks associated with a specific service and at the same time apply different rates (costs per unit of time) based on the engines category used to execute the tasks. For example, engine configurations could fall into the following categories:
- Dedicated Servers – physical servers within the data centers that are dedicated to the grid infrastructure. These servers are intended only for the computations needs of grid services. These servers don’t typically run any other applications and may command higher chargeback rates for tasks running within this dedicated environment.
- Scavenged Servers – these are physical servers within the data centers that are normally running other applications, but can also be used as grid engines within the broader compute grid when they have available CPU capacity. This type of resource can be priced at a lower cost than the dedicated servers.
- Harvested Desktops – these are general-purpose desktops that join the grid as engines available to process service tasks when they have free processing capacity to “donate” CPU resources to the grid. These engines tend to be more dynamic and typically join/leave the grid based on user workstation activity and would likely have the lowest rate/cost than either dedicated or scavenged servers within the grid.
The Evident ClearStone Service Accounting solution for the compute grid support the following capabilities:
- Ability to charge based on the elapsed time of task on an engine by service,
- Support for multiple rate classes for different types of engine configurations,
- Support for on-demand bill execution, and
- Generation of cross-chargeback summary and detailed reports
The following are sample summary and detailed chargeback reports:
|