Rome wasn’t built in a day and neither will your next data center. There’s a great deal of complexity and moving parts to consider along with many potential pain points that cause delays. That’s why the design phase is so critical. Whether you are constructing a data center or erecting a building, you need a blueprint to guide your crew throughout the process. A data center blueprint is more than just drawing equipment racks with mounted devices. It needs to include granular details such as IP address pools, VLAN IDs, routing protocols, and security policies – all of which must be pushed out in the form of configuration settings for all interconnected devices.
Sounds like an arduous process, doesn’t it? It can be if you don’t have a system in place to guide you. We've identified Juniper Apstra to be that essential system that provides the needed guidance, however. It’s an intent-based multi-vendor solution that automates and validates the process of designing and deploying your next data center project. Juniper Apstra will also assist you in operating and managing your data center throughout its life cycle, but let’s not get ahead of ourselves just yet.
The Basic Process of Juniper Apstra
There are two fundamental aspects of the intent-based networking platform for creating a blueprint: intent fulfillment and intent assurance. Intent fulfillment means defining what the intent is so that the data center aligns properly with business objectives. Meanwhile, intent assurance involves validating the design elements before any configurations or settings are deployed. The entire process can be segmented into these five steps:
- Define the intent for the project.
- Translate intent into a series of prescriptive network configurations.
- Verify that the changes are valid prior to deployment.
- Deploy the verified configurations.
- Monitor the network state to ensure continuous compliance with the intent and make changes, if necessary.
Juniper Apstra installs on more than just one virtual machine and is accessed using a web browser. You then start the blueprint creation process using preconfigured or custom templates. As the blueprint is created, Apstra creates the necessary configurations for your devices.
Designing Your Switches and Servers
The initial design steps are created using an abstraction layer that eliminates vendor dependencies. Here, you will define your spine and leaf switches as well as your servers as logical devices. For switches, you will specify details such as port count, port speed, and how the ports are to be used. For instance, you can require that the highest port number of every switch should always use 10 GB. Once these specifications are defined, you can choose amongst a list of multiple switch vendor models that comply with your stated specifications. You can then move on to servers. You only need to create one logical server to represent each server type, so you specify the details just once.
Designing Your Racks
Once your logical switches and servers are created you can go about laying out your racks. You can do this using existing templates, which you can modify to your needs. Like servers, you can design a rack once and use that design as many times as needed from then on. You can start inserting your spine and leaf switches followed by your servers. Once a rack has been completed, Apstra validates the number of available switchports to match your connectivity requirements and ensure that all servers have a connection. Once a rack is completed, you can save it as a template to be used again.
Defining Resource Pools
All your devices need their underlying resources pools. Apstra has four types of resource pools: IPv4 addresses, IPv6 addresses, ASN numbers and VLAN IDs. These pools will later be assigned to devices during the deployment stage. You define IPv4 addresses for things such as loopback IPs for BGP Router IDs and link IPs between switches. There are nifty status indicators that will let you know things like how many IP addresses within a subnet are being used. ASN pools are used for your inter-autonomous routing. Once completed, you can then create your routing policies that govern the movement of traffic throughout the data center.
The Final Blueprint
The final blueprint is comprised of the completed data center templates you created throughout the design process. Think of each template as an instance of the final design, with each instance containing the involved configurations that will be deployed. The design process is performed in blueprint staged view, which means that changes are not applied until committed. Only at that point will the incremental configurations be pushed out to your chosen devices using zero-touch-provisioning. The entire process from start to finish is designed for minimal manual intervention.
Designing a future data center is certainly a daunting task, but Juniper Apstra takes the intimidation away by guiding you through the entire process using an inviting GUI. For complex enterprises, cloud providers and service providers deploying more complex EVPN-VXLAN/IP fabric data centers, there is Juniper’s Apstra Automated Data Center Deployment Service. This proven service gives you access to highly experienced data center deployment experts who assist you when needed. Whatever your future data center migration demands, Juniper has the tools to simply and accelerate your project, while ensuring that it meets or exceeds its intended purpose.
Next Steps: Whether you are responding to evolving workloads and performance needs or changing scalability and resiliency requirements, chances are your enterprise will undergo a data center migration at some point. Download our white paper, Simplifying Your Data Center Deployment And Management Strategy and learn:
- The challenges of inevitable data center migrations
- How to incrementally build and validate blueprints
- How Juniper Apstra executes your intent
- Apstra automated data center deployment service