Should your application consider data mesh connectivity?

Cloud delivery has become the norm for enterprise applications. According to reports, the global public cloud services market will grow 17% this year. However, some applications migrating to the cloud are being held back as they need to connect to data that remains in on-premises, distributed data centers.

There are numerous reasons why data may not be suitable for the cloud, whether it is because the data is too large to be moved, changes too often to carry every change, or because it is controlled and controlled by another company secured. Although the cause can be different, the result is the same: when data cannot be moved to the cloud and stored alongside the application, connectivity challenges arise.

Regardless of why the data needs to be localized, application providers must be able to seamlessly integrate these distributed environments in a secure and sustainable manner. This is where the concept of a data mesh architecture comes into play. A data network enables software vendors to use a single connectivity method to bridge the gap between a centralized application and different local environments.

For those unfamiliar with this term, a data network creates a connectivity structure that integrates data sources with the applications based on them.

I have spent most of my career in IT infrastructure and have seen the rise of public cloud environments and the challenges for application providers integrating on-premises data sources. My most recent company was founded to address these growing connectivity challenges by enabling the creation of data networks. I’ve seen firsthand the potential a data network can offer for applications. Here’s what you need to know when considering data mesh connectivity for your business.

What is a data network?

For context, let’s start with a concept that many may already be familiar with. A service mesh is a software infrastructure that controls how parts of an application in a microservices architecture communicate and integrate with one another.

In contrast to a monolithic application architecture, a service mesh breaks down applications into a collection of services that can be independently created and maintained. Service meshes have improved the way applications create and deliver their services. The accessibility of data has not experienced the same level of development

Similar to a service mesh, a data mesh enables data sources to remain distributed and controlled by different organizations, but accessible for a centralized application. With a data mesh architecture, data is guaranteed to be highly available, easy to find, secure and interoperable with the applications that access it.

How does a data network work?

The move to hybrid cloud environments has forced a re-evaluation of application architectures, connectivity models and security positions. Solving these problems requires a holistic approach that leverages expertise from multiple disciplines.

A data network is uniquely positioned to solve these multidisciplinary problems. Building a connectivity layer, which also includes centralized management and automation functions, combines these functions in a way that eliminates the need for highly specialized experts to build and maintain this new data structure.

The structure of a data network combines software-defined networks with provisioning, administration and automation tools. This creates a secure connective tissue between applications and multiple data sources. This layer of connectivity, built into automated management functions, enables users and applications to consume data freely and on demand, eliminating concerns about where the data is stored or how it is transmitted.

Is a data network secure?

As these resources become more diverse and diversified, you may wonder how all of this is secured. Is it as secure as legacy technologies like IPSec VPN?

Since one of the core use cases is the connection of centralized applications with data from several organizations, data networks inherently solve different, more complex problems than an ordinary VPN. Unlike branch-to-branch networking, challenges arise when an organization does not have both sides of a network connection. This obstacle further distinguishes a data network that is designed to secure sensitive data from the least privileged access via encryption to ongoing maintenance of the system.

Since a data network uses software with native automation and security tools, some of the need for tools and know-how from third-party providers is eliminated. In addition, data mesh security features are designed for speed and scalability of deployment, while legacy technologies like VPN require significantly more time to configure and maintain connections with enhanced security.

A data network can configure new connections in hours instead of weeks, while hundreds of connections can be patched simultaneously instead of individually. Designed as a zero trust solution, it provides additional support to prevent data breaches by defaulting to “deny” traffic until it is a known and trusted source.

A data network allows data and applications to stay where they make the most sense and grants centrally applied administrative controls, while engineering or DevOps teams can provision and deactivate connections as needed without having to rely on internally managed infrastructure .

It’s worth noting that while a data network can be ideal for connections between two different organizations (z-branch connections within the same organization.

As software vendors continue to centralize applications, the need to securely connect to a variety of local environments becomes more important. Building a data network that works in the service of the application not only creates a seamless method for data access, but also enables software providers to focus on delivering software – instead of worrying about infrastructure management.