English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Serverless computing has gained a strong traction in the cloud computingcommunity in recent years. Among the many benefits of this novel computingmodel, the rapid auto-scaling capability of user applications takes prominence.However, the offer of adhoc scaling of user deployments at function levelintroduces many complications to serverless systems. The added delay andfailures in function request executions caused by the time consumed fordynamically creating new resources to suit function workloads, known as thecold-start delay, is one such very prevalent shortcoming. Maintaining idleresource pools to alleviate this issue often results in wasted resources fromthe cloud provider perspective. Existing solutions to address this limitationmostly focus on predicting and understanding function load levels in order toproactively create required resources. Although these solutions improvefunction performance, the lack of understanding on the overall systemcharacteristics in making these scaling decisions often leads to thesub-optimal usage of system resources. Further, the multi-tenant nature ofserverless systems requires a scalable solution adaptable for multipleco-existing applications, a limitation seen in most current solutions. In thispaper, we introduce a novel multi-agent Deep Reinforcement Learning basedintelligent solution for both horizontal and vertical scaling of functionresources, based on a comprehensive understanding on both function and systemrequirements. Our solution elevates function performance reducing cold starts,while also offering the flexibility for optimizing resource maintenance cost tothe service providers. Experiments conducted considering varying workloadscenarios show improvements of up to 23% and 34% in terms of applicationlatency and request failures, while also saving up to 45% in infrastructurecost for the service providers.

A Deep Reinforcement Learning based Algorithm for Time and Cost Optimized Scaling of Serverless Applications