Serverless computing is a popular cloud paradigm that frees developers from server management. Function-as-a-Service (FaaS) the most implementation of serverless computing, representing applications as event-driven and stateless functions. However, existing studies report functions FaaS severely suffer cold-start latency. In this article, we propose an approach, namely, FaaSLight , to accelerating cold start for through application-level optimization. We first conduct measurement study...

Serverless computing is an emerging paradigm for cloud computing, gaining traction in a wide range of applications such as video processing and machine learning. This new allows developers to focus on the development logic serverless based (abbreviated serverless-based applications) granularity function, thereby freeing from tedious error-prone infrastructure management. Meanwhile, it also introduces challenges design, implementation, deployment applications, current platforms are far away...

Serverless computing is a popular cloud paradigm that has found widespread adoption across various online workloads. It allows software engineers to develop applications as set of functions (called serverless ). However, accurately measuring the performance (i.e., end-to-end response latency) challenging due highly dynamic nature environment in which they run. To tackle this problem, potential solution apply checks testing techniques determine how many repetitions given function range inputs...

Abstract Serverless computing has become a new trending paradigm in cloud computing, allowing developers to focus on the development of core application logic and rapidly construct prototype via composition independent functions. With prosperity serverless major vendors have successively rolled out their commodity platforms. However, characteristics these platforms not been systematically studied. Measuring can help select most adequate platform develop serverless‐based applications right...

Serverless computing is a popular cloud paradigm that frees developers from server management. Function-as-a-Service (FaaS) the most implementation of serverless computing, representing applications as event-driven and stateless functions. However, existing studies report functions FaaS severely suffer cold-start latency. In this paper, we propose an approach namely FaaSLight to accelerating cold start for through application-level optimization. We first conduct measurement study investigate...

Deep learning (DL) has become a key component of modern software. In the “ big model ” era, rich features DL-based software (i.e., DL software) substantially rely on powerful models, e.g., BERT, GPT-3, and recently emerging GPT-4, which are trained cloud with large datasets. Hence, training effective models vital stage in whole lifecycle. When deep especially those developers need to parallelize distribute computation memory resources amongst multiple devices (e.g., cluster GPUs) process, is...

Major cloud providers increasingly roll out their serverless workflow services to orchestrate functions, making it possible construct complex applications effectively. A comprehensive study is necessary help developers understand the pros and cons, make better choices among these services. However, characteristics of have not been systematically analyzed. To fill knowledge gap, we conduct a measurement on four mainstream services, focusing both features performance. First, review official...

Serverless computing is an emerging cloud paradigm that has been applied to various domains, including machine learning, scientific computing, video processing, etc. To develop serverless computing-based software applications (a.k.a., applications), developers follow the new cloud-based architecture, where they event-driven without need for complex and error-prone server management. The great demand developing poses unique challenges developers. However, Software Engineering (SE) not yet...

Serverless computing is an emerging cloud paradigm that enables developers to build applications at the function level, known as serverless applications. Amazon Web Services (AWS), leading provider in this domain, provides Application Model (AWS SAM), most widely adopted configuration schema for configuring and managing through a specified file. However, misconfigurations pose significant challenge development. Traditional data-driven techniques may struggle with because complexity of...

Serverless computing has become a new trending paradigm in cloud computing, allowing developers to focus on the development of core application logic and rapidly construct prototype via composition independent functions. With prosperity serverless major vendors have successively rolled out their commodity platforms. However, characteristics these platforms not been systematically studied. Measuring can help select most adequate platform develop serverless-based applications right way. To...

Along with the wide-adoption of Serverless Computing, more and applications are developed deployed on cloud platforms. Major providers present their serverless workflow services to orchestrate functions, making it possible perform complex effectively. A comprehensive instruction is necessary help developers understand pros cons, make better choices among these services. However, characteristics have not been systematically analyzed. To fill knowledge gap, we survey four mainstream services,...

Deep learning (DL) has become a key component of modern software. In the "big model" era, rich features DL-based software substantially rely on powerful DL models, e.g., BERT, GPT-3, and recently emerging GPT-4, which are trained cloud with large datasets. Hence, training effective models vital stage in whole lifecycle. When deep especially those big developers need to parallelize distribute computation memory resources amongst multiple devices process, is known as distributed training, or...

Serverless computing is an emerging cloud paradigm, which allows software engineers to develop applications at the granularity of function (called serverless functions). However, multiple identical runs same functions can show different performance (i.e., response latencies) due highly dynamic underlying environment where these are executed. We conduct first work study raise awareness this variance among researchers. investigate 59 related research papers published in top-tier conferences,...

Serverless computing is an emerging cloud paradigm that allows software engineers to develop applications as a set of functions (called serverless functions). However, accurately obtaining the performance (i.e., response latency) challenging due highly dynamic nature environment in which they run. To tackle this problem, possible solution use testing determine how many repetitions function with given input are needed cater fluctuation. end, we conduct empirical study state-of-the-art...

In the current study, a two-dimensional numerical study is carried out to investigate performance of novel Double-chamber Parallel Flexible Valves micropump, which utilized electrowetting-on-dielectrics (EWOD) effect drive microfluid flow. By observing flow fields, internal circulations are seen on both left and right side pump. The generation backflow discussed by tracking movement vortices. Only slight fluctuation in micropump. Based simulation outcomes, structural parameters including...

In the current study, a two-dimensional numerical study is carried out to investigate performance of novel Double-Chamber Parallel Flexible Valve micropump, which utilized electrowetting-on-dielectrics (EWOD) effect drive microfluid flow. By observing flow fields, internal circulations are seen on both left and right sides pump. The generation backflow discussed as well by tracking movement vortices. Only slight fluctuations in micropump. Based simulation results, structural parameters...