Tikal 2022 Radar, Survey of Technologies - results

Published in

Israeli Tech Radar

5 min readAug 24, 2022

Introduction:

Every year Tikal publishes its Tech Radar. The radar, amongst other things, tries to forecast where the tech world is headed, what the trends are, and which software, libraries, and systems will most likely conquer the software development arena.
The full story about the Radar can be found in this article.

The Radar for 2022 can be found here.

Getting the final recommendations about technologies involves many actions. This year, in addition to all the research, debates, etc... We have implemented a short technologies survey.

This article will focus on the survey and its results.

The Survey:

The survey is based on the 2021 Tikal Radar.
The questionnaire starts with:

Name
Company
Domain
where Domain can be one of:
Backend and Machine Learning
DevOps
Frontend
Mobile

After selecting Domain, a relevant list of technologies is displayed. In this stage, the participant can check every technology it is being used in the organization.

The list of technologies is taken from last year’s Radar.

Disclaimers:

As mentioned above, the list of technologies is taken from last year’s Radar.
As a result of the above bullet, some major technologies might not be part of the list.
We are focused on Open Source tools. There might be some non-free technologies on the list, but these are rare. E.g. DataDog is not in the list.
The request to fill in the survey was sent mainly to customers of Tikal. All of these companies are Israelis or companies holding an Israeli branch.

Conclusions per Domain:

Backend & ML (151 answers):

Backend

In the battle of the languages, Python (68%) and node.js (46%) are leading over Java (33%) while Go is closing the list with 12%. This trend was predicted by the Radar of last year. When it comes to Node.js more than 66% mentioned they use Typescript. It turns out that there are still 10% using the deprecated Python 2!
When it comes to the data engineering arena, Elasticsearch is being used the most (47%). The Kafka ecosystem continues to flourish with 30% using the message broker, 16% with Kafka Streams and 13% with Kafka Connect. RabbitMQ is still here with an impressive 23% use. BigQuery is used by 9% and MongoDB by merely 2%. Apache Spark is the strongest data engine with 20%. Only 2% reported to use Apache Beam.
The favorite infrastructure for backend systems is Kubernetes, with 50%. 30% reported to use sidecar containers in their pods. AWS Lambda is being used in 30% of the companies. Airflow is also here with 16% usage.
Protocols: gRPC protocol is playing a role at 20%, GraphQL with 14% and OpenAPI (21%) is being used to visualize the API.

Machine Learning

Tensorflow and Scikit-Learn are each being used at 15% of the companies. Torch is lagging behind with 6%. AutoML is also here with 5%.
When it comes to processing data, it is no surprise that 27% of the companies report using Pandas. 19% are using Jupyter Notebooks and 13% reported using the newer Jupyter Labs.

DevOps (90 answers):

Kubernetes is the mighty king of all orchestration platforms, used in 81% of organizations. We can also add Docker Compose in 50% usage, but we believe that it is not one on top of the other. In many organizations, Kubernetes and Docker Compose are living side by side, whereas Docker Compose is more of a quick start rather than solid prod.
The popularity of Kubernetes makes other related tools popular as well, like:
Helm with 64%
Nginx Ingress Controller with 48%
Lens with 32%
Istio: 20%
Jenkins is being used in 64% of the organizations — still the Sultan of CI/CD tools. This brings Groovy to get a high score as well with 42%.
Other tools like Gitlab CI and another git repo-oriented CI/CD tools are gaining power as well (e.g. Gitlab CI with 26%).
The interesting thing to see is GitOps with 48%.
The main conclusion is the changed concept. Organizations are starting to go for GitOps in Hybrid mode. Slowly but surely we see this trend over the last 2–3 years — Transitioning from legacy CI/CD tools into GitOps.
Loki looks like a decent competitor to the legacy logging tools with 24%.
Pulumi is not taking off with only 6% of usage.
Grafana is being used in 67% of the organizations. This is by far the most popular UI for metrics.
AWS Lambda — with 43%. This shows that Serverless, and particularly the AWS product is very popular.

Frontend (100 answers):

React is still the most popular frontend library (24%).
Typescript is the most used language in frontend and BFF (82%).
Node is used for every frontend framework and library, as a runner and builder, and also for BFF. It mostly uses Typescript for BFF (72%).
Cypress is the top framework for E2E (30%).
Vite rises, it still has a way to go but it might gain its share by migration from traditional CLIs (8%).
Angular share is getting smaller, however slowly (10%).

Mobile (30 answers):

Many companies have ditched the Android Support library and moved to Android X & Jetpack (62%).
Kotlin is still the main language for Android development and we can see less using Java (64% using Kotlin).
Kotlin Flow is starting to be adopted by more developers introducing an easier way to reactive programming (28%).
Firebase is still the quick go-to/POC when wanting to use a serverless backend on mobile (67%).
Lottie is being used by more companies and developers, as it allows designers to create easy animations which look and work great on mobile apps (23%).
Jetpack Compose has not yet been used by many companies and developers, a new declarative approach for Android UI development (8%).
We can see more apps and companies giving Flutter a shot and starting to understand the development power and speed (22%).

Rough plan for next year:

The survey for next year is going to be a bit more detailed and wider in terms of population.
We plan to add some questions about the company like R&D size, Area (Finance, Gaming, Security, Transportation, etc…), and more.

The list of technologies is going to be presented by categories and detailed below. Some categories might be Languages, Orchestration tools, Provisioning tools, etc…

The main additional goal that we would like to achieve is analyzing usage, also by a company profile. For example: How often Corporates are using Open Source tools vs small companies?