In harell/microservices: Expose the Functionality of a Software Program to the Outside via 'HTTP' Requests

source(file.path(usethis::proj_get(), "vignettes",  "_common.R"))

microservices

r read.dcf("DESCRIPTION", "Title")[1]

knitr::include_graphics("https://i.imgur.com/nZ5jw2g.png")

Introduction

r read.dcf("DESCRIPTION", "Description")[1]

Should I use microservices?

As a start, ask yourself if a microservice architecture is a good choice for the system you're working on?

stop("If you do not plan to deploy a system into production, then you do not need microservices.")

The microservice architecture entails costs and provides benefits. Making lots of independent parts work together incurs complexities. Management, maintenance, support, and testing costs add up in the system. Many software development efforts would be better off if they don't use it.

If you plan to deploy a system into production, then consider the following:

• Favour a monolith over microservices for simple applications. Monolith can get you to market quicker than microservices.
• Monolith-first Strategy: If possible, start with a monolith and design it with clear bounded contexts. Then, when deemed necessary, gradually peel off microservices at the edges.

After mentioning the disadvantages and dangers of implementing microservices, why should someone consider using them?

[@Newman2015] suggests seven key benefits of using microservices:

• Technology Heterogeneity. With a system composed of multiple collaborating services, we can decide to use different technologies/programming-languages inside each one. This allows us to pick the right tool for each job rather than to select a more standardised, one-size-fits-all approach that often ends up being the lowest common denominator.

knitr::include_graphics("https://i.imgur.com/vX1u9Po.png")

• Resilience. A key concept in resilience engineering is the bulkhead which originates in ship design (see illustration). If one component of a system fails, but that failure doesn't cascade, you can isolate the problem and the rest of the system can carry on working.

knitr::include_graphics("https://i.imgur.com/qelkZ9P.png")

• Scalability. With a large, monolithic service, we have to scale everything together. One small part of our overall system is constrained in performance, but if that behaviour is locked up in a giant monolithic application, we have to handle scaling everything as a piece. With smaller services, we can scale those services that need scaling, allowing us to run other parts of the system on smaller, less powerful hardware.

knitr::include_graphics("https://i.imgur.com/1Tf9Hrh.png")

• Ease of Deployment. With microservices, we can make a change to a single service and deploy it independently of the rest of the system. This allows us to get our code deployed faster. If a problem does occur, it can be isolated quickly to an individual service, making fast rollback easy to achieve. It also means we can get our new functionality out to customers faster.

knitr::include_graphics("https://i.imgur.com/U60xp1V.png")

• Organizational Alignment. Microservices allow us to better align our architecture to our organisation, helping us minimise the number of people working on anyone codebase to hit the sweet spot of team size and productivity.

knitr::include_graphics("https://i.imgur.com/8mk0BlZ.png")

• Composability. Similarly, to the tidyverse packages, where different compositions of packages are used in different analytic projects in R, with microservices, we allow for our functionality to be consumed in different ways for different purposes. For example, a demand forecasting service can be consumed by several different dashboards and a logging system.

knitr::include_graphics("https://i.imgur.com/gHMkhtV.png")

• Optimizing for Replaceability. With our services being small in size, the cost to replace them with a better implementation, or even delete them altogether, is much easier to manage than in a monolithic app. For example, during the football world cup, a news website may offer football-related analysis. Rather than making the analysis part of the website codebase, we can create a microservice to deliver the football analytic service and remove it when the tournament is over.

knitr::include_graphics("https://i.imgur.com/SNQAINt.png")

To conclude, not every application needs to be built as a microservice. In some cases, such as in a system that is an amalgam of programming languages and technologies, microservices architecture is advised or even necessary. However, seldom it is a good choice to start building an application as a microservice. Instead, a better option is to design a system in a modular way and implement it as a monolith. If done well, shifting to microservices would be possible with reasonable refactoring effort.

Installation

You can install microservices by using:

install.packages("microservices")

Further Reading

harell/microservices documentation built on March 2, 2021, 3:15 a.m.