Efficient provisioning of resources in public infrastructure for electronic design automation (EDA) tasks

Abstract

Provisioning resources in public cloud infrastructure to perform at least part of electronic design automation (EDA) tasks on the public cloud infrastructure. Performance metrics of servers in the public cloud infrastructure and performance history of a user's past EDA tasks are maintained to estimate operation parameters such as runtime of a new EDA task. Based on the estimation, a user can provision appropriate types and amounts of resources in the public cloud infrastructure in a cost-efficient manner. Also, a plurality of EDA tasks are assigned to computing resources in a manner that minimizes the overall cost for performing the EDA tasks.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS[0001]This application claims priority under 35 U.S.C. §119(e) to co-pending U.S. Provisional Patent Application No. 61 / 319,224 entitled “Method or Apparatus for Automated Partitioning and Distribution of EDA Application Workloads between Multiple Hybrid Public / Private Cloud Infrastructure to Increase Workflow Efficiency,” filed on Mar. 30, 2010; co-pending U.S. Provisional Patent Application No. 61 / 319,219 entitled “Algorithm to Provision Compute Capacity on Public Clouds On-demand Based on Historic EDA Job Flow Characteristics to Reduce Cost of Infrastructure,” filed on Mar. 30, 2010; co-pending U.S. Provisional Patent Application No. 61 / 319,216 entitled “Method and System to Implement Least Infrastructure Cost for Executing EDA Simulation Workloads on Public Clouds,” filed on Mar. 30, 2010; co-pending U.S. Provisional Patent Application No. 61 / 319,206 entitled “A Method Or Apparatus For Providing Intelligent Session Licensing / Entitlement In ...

AI Technical Summary

Benefits of technology

[0009]Embodiments relate to provisioning computing resources in a public computing infrastructure to perform electronic design automation (EDA) tasks where the computing resources are provisioned based on performance metrics of the computing resources. An EDA task set is divided into a plurality of EDA tasks at least part of which is performed on the public computing infrastructure. The performance metrics represent performance characteristics of the computing resources associated with executing the EDA tasks. The performance metrics allows estimation of operational parameters associated with the EDA tasks. Based on the estimation, appropriate types and number of computing resources may be provisioned, and thereby avoid over-provisioning or under-provisioning of computing resources for the EDA task sets.

Problems solved by technology

As more components are integrated into an IC device and more functions are accommodated in the IC device, some EDA task sets have come to require a large amount of computing and storage resources.

Some EDA task sets may take days, weeks or even months to accomplish.

Some EDA tasks may take longer to finish while other EDA tasks may finish in relatively short time.

If computer resources are under-provisioned, the EDA task set may not be finished within time.

Hence, in many cases, redundant computing resources are provisioned to ensure that the EDA tasks are finished within time.

However, over-provisioning computing resources increases the cost associated with the EDA tasks.

During times when a new IC device is not being developed, the developer's servers or server farms may remain idle.

If the servers or server farms are under-provisioned, however, there may be insufficient resources when the designer is actively designing the IC devices.

The large fluctuation in the computing demand as a result of the design or product cycle renders performing the EDA tasks entirely in private infrastructure inefficient.

Images