Google LLC’s cloud business today debuted a new capability, dubbed startup CPU boost, that will enable customers’ applications to perform certain computing tasks with less latency.

The capability is currently in preview. It’s available for Google Cloud’s Cloud Run and Cloud Functions services.

Cloud Run is a service that enables companies to run software container applications on Google Cloud’s infrastructure. The service automates many of the maintenance tasks involved in running applications, thereby saving time for developers. 

Cloud Functions is likewise designed to save time for developers. It’s a serverless computing service that, similarly to Cloud Run, automates many of the tasks historically involved in running software on Google Cloud’s infrastructure. Google Cloud last month released an upgraded version of the service with several performance improvements.

The new startup CPU boost capability that Google debuted today can help companies improve the performance of the workloads they deploy on Cloud Functions and Cloud Run. It does so by reducing applications’ startup time, a metric that influences the speed at which a program carries out certain computing tasks.

Enterprise applications are usually implemented not as a single file, but rather as a collection of standalone software modules. Each module runs in a separate software container. Companies often configure their applications to launch modules only when they are needed and then automatically delete them once they’re no longer used, an approach that helps reduce hardware usage. 

Launching an application module can take a significant amount of time and, until the task is complete, requests to the application may be processed slower. As a result, users may experience increased latency. Google Cloud’s new CPU boost capability is designed to help companies reduce their application modules’ startup time, or the time that the modules take to launch.

The feature is available as an opt-in setting through the Google Cloud management console. When enabled, the feature detects when a new application module is about to launch and assigns additional central processing units to the cloud instance that runs the module. The additional CPUs enable the module to load faster, thereby reducing latency for users.

The number of additional CPUs that Google Cloud deploys depends on the configuration of a company’s cloud environment. According to Google Cloud, its platform provisions between two and eight CPUs to speed up application loading times.

The company measured the performance of the new startup CPU boost capability in a series of internal benchmark tests. During the tests, the capability reduced the startup time of Java applications by as much as 50%. It can also deliver significant performance improvements for applications written in other programming languages, according to Google.

“Customers testing the feature in private preview with Node.js have observed startup time reductions of up to 30%, a significant improvement, a bit less than Java due to the single-threaded nature of Node.js,” Google Cloud group project manager Steren Giannini detailed in a blog post. “Each language, framework, and code base will see different levels of benefit.”

Image: Google