Hamish Ostride Labs

Hamish Lister

Senior Content Manager

Hamish is a senior content manager at Ostride Labs. Hamish’s diverse background in technical research, analysis and market demand are the main drivers behind the topics he enjoys exploring and writing about.

Read more

Configuring customer data platform when doing cloud native application development

Updated 25 Jan 2022


To succeed in evolving, software-driven markets, organizations must optimize the way they design, build, and use applications and data platforms. Along with increasingly popular cloud-native applications, data platforms are a big part of companies’ cloud infrastructure as a whole and, therefore, are an integral component of their cloud native application development cycle.


Faced with immense and ever-growing amounts of data being generated at quicker and quicker rates, software developers need to pay particular attention to the scalability of their data platforms and applications. This is one reason why K3s has become so popular, as it improves the flexibility and scalability of cloud native applications due to its lightweight container properties and minimal resource requirements, differentiating it from K8s. They must also design and configure platforms and cloud native applications that can handle an increasing number of concurrent users. It’s not easy and is a constant challenge, but developing for scalability is an indisputable necessity.


What is a data platform?

While we have looked at cloud native application development in previous blog posts, we haven’t considered the data platform which some consider to be the backbone of the modern, data-focused, cloud infrastructure.


A data platform is a comprehensive and necessary solution for consuming, processing, analyzing, and presenting data created by the many systems, applications, processes, and infrastructures of the contemporary digital enterprise. While there are a plethora of solutions and tailor-made applications for managing various aspects of the data lifecycle effectively, a true data platform ensures end-to-end data management.


A data platform goes much further than providing simple business intelligence statistics. While it does deliver relevant data to enhance an enterprise’s decision-making, a true data platform collects and organizes many more types and configurations of data across the company, including not only integral data used for security and privacy, but also technical IT operations data. Essentially, a complete data platform has the ability to manage ALL the data a company ingests or generates.


Data platforms are composed of data storage, servers, and data architecture. Aside from that, there are data consumption needs, data consolidation, and the ETL procedure.

Businesses routinely face challenges with data management as a whole, including the consolidation of diverse data types stored in various silos, cloud and on-premise servers. This is where effective data platforms show their worth.

The purpose of a data platform is to provide real-time insights through detailed analytics in a scalable, cost-efficient, and secure way. The most efficient data platforms can be held across distant geographies and in hybrid-cloud environments to strengthen business continuity plans. 


For a definition of cloud native applications, see this article


Why data platforms need to be built and configured for scale

Today, the majority of the fastest-growing and most successful businesses are data-driven in some way or another. From more online visitors to an increased desire for analytics, one way or another, data is always being generated that needs to be securely stored. While data frequently appears to be the answer to many business problems, its formidable nature, from its assortment of technologies, skill sets, tools, and platforms, can be complicated and hard to manage. 

Data’s complexity and businesses’ raised demand for it is an added challenge as it becomes more challenging to prioritize, grow the team, recruit leading talent, keep costs down, and satisfy clients and stakeholders. 


Whatever the reason for scaling a data platform, whether that be for an increased user or data volume, there are two many strategies: Vertical and Horizontal.


Vertical Scaling

Possibly the most straightforward way to scale is to do so vertically – deploy on a more advanced cloud server with more CPU power and memory. 

However, there are functional limitations to what can be accomplished through vertical scaling alone. Firstly, even the best machines and cloud servers may not be able to tolerate the immense data volumes and workloads required by modern cloud native applications. Secondly, the power and capacity required to effectively operate the necessary data platform will probably not be too cost-effective. 

Capacity management for single-server architectures can also be challenging, particularly for data platform solutions that will have inconsistent workloads. Having the capacity to manage peak loads could result in wasteful underutilization throughout off-hours. In contrast, having too little server capacity may cause performance to slow significantly during high usage. Moreover, expanding the capability of a single server architecture implies buying an entirely new machine or expanding cloud server storage.

In short, while it is crucial for cloud native applications and data platforms to utilize the full potential of the hardware or server on which it was deployed, vertical scalability on its own isn’t enough to achieve anything above the most stationary workloads.


Horizontal Scaling

For the reasons provided above, most institutions pursuing considerable scalability for their data platform will deploy on hybrid cloud architectures, scaling workloads and data volumes “horizontally” by spreading the load across multiple servers. Utilizing K8s, developers have found this to be most effective with the added benefit of more secure cloud storage.

Software developers will understand that no two workloads are identical. Some modern cloud native applications may be used by millions of users concurrently, with a lot of small transactions per second. Others may have only hundreds of users, but with petabytes worth of data. Both scenarios are very taxing workloads, but they demand different scalability strategies.




With constantly increasing amounts of data being generated, from more data-intensive apps and at quicker rates, developers are now required to pay particular attention to scalability when designing and building their data platforms and cloud native applications. Developers are primarily utilizing vertical and horizontal scaling methods to achieve this scalability by moving to a more advanced cloud architecture or spreading the load with a hybrid model.




Git Kubernetes, or git revert, is now one of the most popular commands among developers for its universal undo function. It doesn’t completely reverse the original command, however, as it is related to a specific commit. It doesn’t take away a commit from the versions history, alternatively, it creates a new commit with inverted content, reverting the project to the former state before the commit.


Soon, we will begin posting content on the ever-popular GitHub.




Our newsletter (you’ll love it):

    Let's talk!