Cloud data processing and storage technology ensures more agility, high scalability, and security in product testing, reducing research and development time for cosmetic products by one third
The Google Cloud Platform, Google's cloud computing platform, is helping Natura accelerate its product innovation process. The Brazilian multinational cosmetics company has migrated to the cloud the technical data resulting from laboratory tests it conducts to evaluate the effectiveness of active ingredients that make up a cosmetic formulation. Thus, since August 2017, it has been utilizing the high scalability and storage capacity of Google's platform in the development of new technologies. In practice, this has meant a 30% gain in the time taken to conduct some tests, thereby reducing the development time of a new item in the portfolio.
The tests are important for Natura to optimize the results of its research and development efforts in the medium and long term. The company simulates the reactions of cosmetic ingredients on human skin models reproduced in a controlled environment to measure whether a particular active ingredient is effective for a specific use. Additionally, with the help of the cloud, it is able to predict through computational simulation how mixtures of these actives in a formulation may behave in contact with the skin. The company has not conducted animal testing since 2006 and has been actively working on developing alternative methodologies to the use of animals. Currently, Natura has 67 alternative methodologies to assess the safety and efficacy of its products, in partnership with Brazilian and international universities and research institutes.
“The volume of data generated by the tests is very large, so the robustness of the Google Cloud Platform is essential for the success of the experiments,” says Daniel Gonzaga, Director of Product Innovation at Natura. “In addition to storing information about the cosmetic active and the study model, the cloud quickly cross-references them and records the generated evidence in the history, contributing even to future studies.”
The Google Cloud Platform also allows Natura to optimize laboratory tests by simulating protocols with small variations on a large scale in the virtual environment. This way, it reduces the time and cost for conducting the tests in the laboratory. Another advantage found was the ability to research and correlate data from already completed studies in the system to assist in decision-making regarding the need for new tests. Technically, the algorithm runs on GKE (Google Kubernetes Engine), an open-source container orchestration system that automates the deployment, scaling, and management of applications in data repositories.
“We reduced the researcher’s rework, as we now have confirmation of the active ingredient's effectiveness in a quick and accessible way,” explains Gonzaga. “On many occasions, the researcher will even skip a certain test because they already know it doesn’t work, proposing a new methodology with a higher chance of success.”
A simple experiment demonstrates the time and scale gains of the testing project. What previously took 48 hours for processing is now done in just 13 hours, a significant reduction when 200 to 300 processes like this are done simultaneously, something made possible by the computational scale in the cloud.
In the cloud with Google
Natura chose Google's solution after a proof of concept, in which the service and delivery capacity were tested. “The platform delivered efficient results and the investment was lower than we projected,” says Luiz Mussini, Director of Digital Technology Infrastructure at Natura. “Scalability and security were relevant attributes in choosing this solution.”
The technical foundation of the data centers explains the high performance of Google’s cloud, says Antonio Chaddad, Google Cloud account manager for Brazil. Google’s platform works as if it were a large machine, using the processing power of all the machines together.
“In this solution, the computer's own operating system separates it into several machines through a virtualizer, allowing Google to use the same number of machines to process more data,” explains Chaddad.
Also, see the article published in the Revista Exame.
