How to solve non-standard tasks in the Backend and not lose. Speakers of the DUMP conference will tell

Usually, participants expect useful reports from conferences with figures, facts and only on the case. We took this into account, took trendy topics and invited experienced speakers. This is a clear and concentrated program for the Backend section. Let’s talk about technologies and methods that are currently relevant. Let’s understand server-side development without binding to programming languages. And also we will look for the optimal solution to the problems faced by most back-end developers.

Just look at which speakers will help us with this!

Mikhail Belyaev from Prosoft-System with a report “Embedded problems or how we left sqlite”

There is a very big problem with embedded development – the very limited resources of the device on which your code is executed. During development, when customer requests are growing, and the resources of embedded devices are running out, very often standard and proven solutions are not suitable. The developer is faced with a choice: either to say that our device cannot cope with the required tasks, or to make a non-standard move and win. Michael will tell other developers about this process of transition from using SQLite to his own data warehouse, which allows reading and writing data much faster.

Andrey Tsvettsikh from EPAM will present a talk “ Clean Architecture in Practice ”… Here’s what he told us about his performance:

– Enough time has passed since the release of Uncle Bob’s book Clean Architecture. Someone read it, and someone just watched the reports on youtube. But all these reports are ideological. Their authors usually have no practical experience in creating large projects using this architecture (as well as launching these projects in production). And all the examples are too simple! In practice, many questions still remain. Over a year ago, we started 2 new projects, in which we applied the principles described in the book. These are enterprise applications in C # (API, backend). Enterprise, which has not yet managed to become bloody 🙂 But this is quite enough to get the first results and share experience.

Roman Nevolin iz Kontur will make a presentation on “ Functional languages ​​for business development ”

Functional languages ​​are often perceived as beautiful and fashionable toys – it’s fun to look and turn around, but they don’t belong in a harsh enterprise. It is generally accepted that simple and reliable languages ​​such as Java are best suited here. Roman cannot agree with this in any way and is constantly trying to apply his favorite functional language – F # – to the next business problem. At the speech, we will discuss how Roman does it, why he (and everyone else) needs it, and what kind of rake he managed to collect on the way.

By the way, Roman can be heard in our free stream, about which we are in detail told here

Eugene Peshkov from JetBrains with a talk “ Client HTTP in .NET: a rake road from WebRequest to SocketsHttpHandler ”

At first glance, it seems that sending an HTTP request is very easy. Nevertheless, even HTTP / 1.1 is quite nontrivial – the RFC for it contains more than 150 pages, in addition, browsers already support HTTP / 2 and HTTP / 3. This leaves no choice: the standard client in the platform must be implemented at a high level.

From the .NET Framework 1.0 to .NET 5, the client APIs for working with HTTP and its implementation have undergone many changes. In some versions they were successful, in others they were failed and clearly temporary.

In his speech, Eugene will talk about the history of the development of the client-side HTTP API in .NET, its features, and the migration of applications from Framework to Core, taking them into account. It will also break down some hacks that are useful when working with it. Let’s take a look at NuGet and consider the wrappers over the HTTP API presented in it in terms of efficiency and cross-platform.

Alexander Polyakov from Yandex will tell you “ How Yandex.Afisha moved to GraphQL 2 times ”

This talk is about how we rewrote the Ya.Afisha API from REST to GraphQL in node.js + Python. And then, as part of the optimization, we got rid of node.js + Python and rewrote all GraphQL in Java.

Let’s deal with the following questions:

– why we chose GraphQL technology

– what problems and tasks were solved with its help

– we will tell and show how our architecture has evolved

– for which teams and projects our solution is suitable, and for which not and why

We’ll also give you some practical tips on how best to get started with GraphQL.

Fagim Sadykov from SpectrumData will present us his talk “ Kotlin as the main language for backend development and data processing ”

Kotlin is a controversial language, on the one hand, it has already occupied a respectable niche of “popular language”, on the other hand, it is perceived by many as some kind of exotic language for Android or just as a curious “try” toy.

At SpectrumData, Kotlin is used as the main (and in fact the only) product development language on the JVM platform for a wide range of tasks: working with data, implementing a micro-service architecture, and implementing a backend. The history of the successful use of Kotlin in the company has been going on for 4 years since the language version 1.1. SpectrumData is characterized by full-fledged use in its codebase of all the features and recommended practices for this language.

After listening to the report, listeners will be able to draw a conclusion about the real applicability and options for migrating to Kotlin from Java. Get a full understanding of the most complex and controversial aspects of using Kotlin.

Denis Tsvettsikh from Invent will talk about “Aspect-oriented programming in C # and .NET yesterday, today and tomorrow”

Aspect-oriented programming (AOP) allows you to add infrastructure functionality like caching and logging to different layers of your application without duplicating code. And all this – without changing the already written code! It’s a very powerful, handy … and rarely used technique. This is partly justified, 10 years ago, the tools for implementing aspects were not developed, so AOP has a reputation for an approach used only by enthusiasts. But since then the world has changed and today AOP can be seen even in web frameworks, it is only important to be able to recognize it.

In his talk, Denis will share 10 years of experience in using AOP in C # and .NET. He will talk about approaches to the implementation of AOP, and also show how the tools for developing aspects have changed along with the programming language and platform. Naturally, he will offer the most optimal implementation of the aspects for today. And together we will think about what we would like to see tools for developing aspects in the future. Examples will be in C # and .NET, but the ideas of the talk will be relevant for any platform.

And at the end of the day, the program committee put Anton Shishkina from SKB LAB with a report ” Recommendations and features of the first freshness ”. Here Anton will talk about his experience in building a recommendation system from offline calculations to online. And separately about how the relevance of features for online calculations is ensured. The “assumptions” due to which the features remain “fresh” will be analyzed, while ensuring the high availability of the feature store.

What will happen in the rest of the sections, see in the main program

And here you can choose tickets in online and offline format.

Similar Posts

Leave a Reply

Your email address will not be published. Required fields are marked *