NetBackup is an interesting piece of software. Designed to handle large-scale data protection needs, it was built for scale. As a result, it can manage vast amounts of data efficiently, even across disparate infrastructures. Originally developed by supercomputer engineers, it was built from the ground up to handle massive data environments with speed, efficiency, and resilience. Its architecture supports high levels of parallelization, allowing thousands of operators to run simultaneously across hybrid and multi-cloud infrastructures. In short, it’s designed to scale intelligently and powerfully. This results in a highly scalable platform that allows us to handle thousands of operations simultaneously, making it a powerful tool for data protection.

What happened when we started to think about cloud

When considering cloud solutions, we realized that yes, Cloud isn’t that different, from on-prem, but… there are unique challenges and opportunities and new ways of thinking. First off, we quickly realized that above all, we needed to optimize for cost to ensure that our application leverages the core cloud services in a way that is affordable and sustainable. For example, getting the best dedup rate isn’t going to mean anything if the cost it too high to run it. Additionally, it's essential to prioritize automation, resiliency, and scale to make the most of cloud technology and ensure reliable performance.

What did we figure out

Through our exploration, we discovered that ideal cloud native is built with a true cloud-native scalable architecture. By using the cost-effective and widely available K8s services built into the CSPs. This delivers the most cloud-native way to manage containerized applications, providing flexibility and scalability. The NBU ephemeral nature fits K8s well, allowing for efficient and dynamic resource management. To move from a server-based architecture to a true service-based architecture required a targeted investment over a nearly 3-year period and has resulted in the most powerful, scalable and efficient next-next-gen technology for enterprise  grade cloud data protection.

How did we accomplish true Cloud Scale

To create Cloud Scale, we focused on container foundations, which provide a robust and flexible infrastructure. We moved from a server-based architecture to service-based, enhancing scalability and efficiency. Solid, rolling updates for web, scheduling, and database services ensure continuous improvement and reliability. Each job gets a dedicated data mover pod that scales elastically, adapting to workload demands. K8s nodes will scale up and down automatically based on pod activity, ensuring optimal resource utilization.

Ephemerality in the core architecture

It may look complex, but don’t worry, it's all automated  behind the scenes. This all you need to worry about:

In this architecture, central, persistent services for orchestration at scale play a crucial role in managing operations. We’re talking things like a prescient scheduler that can launch more than one job per second per client.  These services coordinate various tasks to ensure smooth functioning. On the other hand, the actual job work is done by distributed, ephemeral services designed to do the heavy lifting (Data Movers).  These services can focus on their specific tasks that require flexibility and scalability while the cloud-native k8s platform handles scaling out the cluster infrastructure automatically. A pioneering distributed storage architecture is deduped, compressed and encrypted at source, with a dedicated pod and moves data directly to object storage, avoiding a storage bottleneck while still being centrally tracked to ensure optimal, global deduplication. The recovery points are portable across infrastructures to ensure that data can be efficiently managed and transferred between different infrastructures or clouds. Not to mention this is all done with granular immutability to guarantee your data is secure.

What does this get the customer

For customers, this approach delivers a next-next gen, cloud-native distributed architecture that is highly efficient and scalable. It allows them to move on from the constant sizing challenges, providing flexibility in resource management. Granular price controls for infrastructure enable cost-effective operations. Additionally, data protection as a service cloud, not just an application ensures comprehensive and reliable data management.

AWS, Azure, the architecture is the same regardless of where it runs.

A good cloud data protection platform needs to have broad cloud workload protection to manage various types of data and applications. Storage needs to be optimized for cost above all, ensuring affordability. Global management SaaS ties all the up into a single pane of glass for all deployments on all infrastructures, providing a unified and efficient management solution. Want to see it in action? Take a look at the attached video. 

Look for upcoming posts to hear about: 

How does distributed global deduplication deliver uniquely dedicated resources for the best performance?

What good is a platform that doesn't do anything? How we delivered support for over 64 cloud-native workloads with truly cloud native design in record time.

Here's a preview:

Global management SaaS that puts everything under a single pane of glass, no matter where it’s running