Introducing DriveNets Network Disaggregation Fundamentals
A few weeks back I had the pleasure of being invited to take and review DriveNets Certification Course on Network Disaggregation. In this article, I’m going to discuss what I learned and provide an in-depth review of the course, its benefits, and my plan for the certification exam.
First, who is DriveNets?
If you haven’t heard of DriveNets, they are changing the game when it comes to Networks. Their mission is to “Build Networks like Cloud.” Meaning they are applying the virtualization techniques we’ve come to know and leverage with server resources to network devices. This software-defined approach allows large service and cloud providers to save on the cost of network hardware, scale efficiently and speed up innovation without sacrificing network performance.
What is Network Disaggregation?
Now, let’s get into the cool stuff.
Current networks typically follow the traditional 2- or 3-Tiered Network Design. Hardware is designated for the Core, Distribution, and Access layers. Each layer contains network equipment dedicated to providing purpose-specific connectivity to customers. For an example of this, think about what would be required in terms of an enterprise customer vs. a home user. The enterprise would require significantly more resources, speed, SLAs and security than home customers. However, as service providers expand their network and service offerings, they face challenges around design inefficiencies, operational complexity, scalability, and cost.
A major challenge is the waste of hardware resources. As capacity spikes, there’s no way to shift unused resources around. Service providers invest in large hardware chassis and don’t always fully utilize them, leading to wasting ports and resources. Currently, Service
Providers address this with hyperscalers and spine/leaf architectures, but they are still tied to specific vendors and often run into resource and cost issues.
DriveNets has developed a novel approach that addresses these issues.
DriveNets’ disaggregated network architecture – the DriveNets Network Cloud – separates the hardware and software, removing the dependencies of network functions on the hardware. This architecture also separates the data and control planes, allowing for improved network scale. This is accomplished by adopting the Distributed Disaggregated Chassis (DDC) model from the Open Compute Project (OCP). The OCP community promotes commodity hardware to address new needs for compute and networking infrastructures. The DDC model creates a Linux-based “white box” architecture that allows you to run containers on top. Network functions are no longer tied to the hardware and are now essentially services that are containerized. Containerizing these functions allow them to be shifted as needed without requiring new hardware.
The two key DriveNets’ building blocks of the data plane are the Network Cloud Packet Forwarder (NCP) and the Network Cloud Fabric (NCF). Think of the NCP as your standard line card handling all of the traffic forwarding, routing tables, QoS and policies. NCPs can be used as standalone devices or clustered for more throughput. The NCF is used to cluster and manage multiple NCPs, ensuring optimal network performance. The control plan functions run off of a server. DriveNets’ orchestrator is called DNOR and provides the interface for you to interact, configure and monitor the system.
The NCPs and NCFs connect to each other with 400 Gbps connections and can be clustered together. These clusters can reach speeds up to 691Tbps.
To summarize, the DDC and DriveNets Network Cloud architecture address the need for a new innovative approach to service provider networks. It allows for full and efficient use of all network resources, thus reducing complexity and cost.
Leveraging this information
If you’re in the service provider space, this is a game-changer as it addresses the need for improved scalability, vendor choice, simplified operations, and reduced costs.
While this innovation is mainly targeted to the Service Provider space for now (and the course’s main audience is service and cloud provider professionals at this point), I can see this expanding to the enterprise space and becoming an architecture option in the future. As the world transitions to a hybrid workforce, we will see the need to dynamically shift resources to meet the demands of workloads. I believe DriveNets’ mission of building networks with a cloud approach will definitely cause a shift in our industry. Taking this course will give you a solid introduction to the topic and you can take the certification exam to display your new skillset.
Course Review
The Network Disaggregation Fundamentals course targets network professionals in the service or cloud provider field. The course has ten modules covering the current challenges with existing network architectures, and deep-diving into the details of DriveNets’ solutions. The course breaks down the solution layer by layer, ensuring that you have a clear understanding of the building blocks and use cases.
Overall, I felt the course was straightforward, with no unnecessary fluff, and was easy to understand. While I’m new to the Service Provider world, DriveNets did an excellent job of explaining the problem and its solutions. You will need a solid understanding of networks and network architectures. The material is crystal clear, and the video modules are brief and to the point. I also enjoyed the fact they included lots of graphics and diagrams to better explain the concepts presented. The course is about four hours long so it’s perfect for you to look at over a weekend. The slides make it easy to pause and take notes. Throughout the course, you will have quizzes to test your knowledge of what you’ve just covered. The course also ends with a certification exam that will test your knowledge. I will be discussing the exam later in a video.
You can check out DriveNets’ New Training Portal Here.
Update: Here’s my YouTube Video Detailing my experience.
