Toote Unified Logging üleviimine serverivabale mudelile
Date
2019
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Mobiiltelefonide ja Interneti kasutamine igapäevasteks tegevusteks kasvad päev-päevalt ja seetõttu otsivad veebipõhiseid rakendusi ja teenuseid arendavad ettevõtted lahendusi, kuidas nende tooted ja teenused oleksid miljonitele kasutajatele pidevalt kättesaadavad. Seetõttu on järjest enam kasvanud pilveteenuste kasutamine nii suurte kui väikeste ettevõtete poolt, kes veebipõhiseid rakendusi arendavad. Üha suurenev pilveteenuste nõudlus on viimastel aastatel kaasa toonud pilveteenuste kiire arengu, mille tulemusena on välja arenenud kolm peamist pilveressursi tarbimise mudelit: infrastruktuur kui teenus (IaaS), platvorm kui teenus (PaaS) and tarkvara kui teenus (SaaS). Nimetatud mudeleid kasutades saab klient valida, kui suures ulatuses vastutab pilveteenuse pakkuja infrastruktuuri ehitamise ja juhtimise eest. Pilveteenused on pidevas arengus ning IaaS-i, PaaS-i ja SaaS-i kõrval on viimasel ajal järjest enam tähelepanu ja kasutust saamas kontseptsioon “funktsioonid kui teenus (FaaS)”, kus arendatav rakendus põhineb ühel või mitmel funktsioonil, mille lõplik paigaldus toimub pilveteenuse pakkuja poolt. Funktsioonid kui teenus, mida sageli nimetatakse ka serverivabaks mudeliks, võimaldab ettevõtetel hoida kokku nii kulusid kui aega, kuna pilveteenuse pakkuja vastutab kõigi tarkvara arendamiseks vajalike ressursside haldamise eest ning arendusressursside jaotamine on automaatne. Erinevad uurimised on näidanud, et nö. serverivabale arhitektuurile üleminek võimaldab ettevõtetel kulusid infrastruktuurile vähendada kuni 70% ulatuses. Microsoft, üks suurimaid pilveteenuste pakkujaid turul, on ka oma korporatsioonisiseseid rakendusi ja teenuseid üha enam pilveteenustele üle viimas, kuna Microsofti pilveteenuste portfoolio Azure-i poolt pakutavad teenused on tõestanud kõrget konkurentsivõimet. Käesoleva magistritöö eesmärk on analüüsida, mis on pilveteenuste olemus, millised on pilveteenuste eelised ja puudused ning turul pakutavad teenused. Töö empiiriline osa keskendub Microsoftis sisemiseks kasutamiseks mõeldud tööriista pilveteenustele üle viimise protsessi analüüsile ja selle mõju hindamisele.
With millions of users connected to the Internet and using mobile phones for their daily ac-tivities, companies developing web applications and services are facing a situation where their products need to be highly scalable and reliable to meet the increasing usage. This has led to a greater adoption of cloud computing services by enterprises of different sizes, from small start-ups to large corporations. The emergence of cloud computing has resulted in sev-eral cloud deployment and service models becoming available for companies developing online services and applications. Three main service models have developed over the recent years: Infrastructure as a Service (IaaS), Platform as a Service (PaaS) and Software as a Service (SaaS), allowing companies to optimize resources and control the level of infrastruc-ture they need to manage. As cloud computing technologies are relatively new, they are evolving rapidly, and new cloud computing concepts are developing based on IaaS, PaaS and FaaS. One of the fastest growing paradigms in cloud computing is serverless compu-ting. Through adoption of a serverless computing model, companies are building their appli-cations in separate functions deployed to a serverless service managed by the service provid-er. A serverless approach allows companies to reduce infrastructure costs significantly, in some cases even up to 70%, as the pricing model of serverless is based on the CPU con-sumption of the functions only while they are actively running. Microsoft, one of the market leaders in terms of serverless offerings, is also moving their own products and solutions to cloud-based services. The goal of the given thesis is to present a review of cloud computing models and services, and the evolution and essence of serverless computing. The empirical part of the paper analyzes the process of moving an internal debugging tool at Microsoft to a cloud computing model.
With millions of users connected to the Internet and using mobile phones for their daily ac-tivities, companies developing web applications and services are facing a situation where their products need to be highly scalable and reliable to meet the increasing usage. This has led to a greater adoption of cloud computing services by enterprises of different sizes, from small start-ups to large corporations. The emergence of cloud computing has resulted in sev-eral cloud deployment and service models becoming available for companies developing online services and applications. Three main service models have developed over the recent years: Infrastructure as a Service (IaaS), Platform as a Service (PaaS) and Software as a Service (SaaS), allowing companies to optimize resources and control the level of infrastruc-ture they need to manage. As cloud computing technologies are relatively new, they are evolving rapidly, and new cloud computing concepts are developing based on IaaS, PaaS and FaaS. One of the fastest growing paradigms in cloud computing is serverless compu-ting. Through adoption of a serverless computing model, companies are building their appli-cations in separate functions deployed to a serverless service managed by the service provid-er. A serverless approach allows companies to reduce infrastructure costs significantly, in some cases even up to 70%, as the pricing model of serverless is based on the CPU con-sumption of the functions only while they are actively running. Microsoft, one of the market leaders in terms of serverless offerings, is also moving their own products and solutions to cloud-based services. The goal of the given thesis is to present a review of cloud computing models and services, and the evolution and essence of serverless computing. The empirical part of the paper analyzes the process of moving an internal debugging tool at Microsoft to a cloud computing model.