For those who’ve been following the evolution of Cisco Modeling Labs (CML), you recognize we’re at all times on the lookout for methods to make community simulation extra accessible, extra highly effective, and albeit, extra enjoyable. In the present day, I’m excited to share one thing that brings all these parts collectively in a approach which may simply change how you consider constructing and testing community topologies.
We’ve simply launched a Mannequin Context Protocol (MCP) server for CML, and in the event you haven’t heard of MCP but, you’re in for a deal with. Consider it as a bridge that lets AI assistants like Claude, VSCode, Cursor, and LM Studio straight work together along with your CML atmosphere. What does that imply in apply? It means you’ll be able to actually discuss to your lab atmosphere and watch it come to life.
What’s MCP?
The Mannequin Context Protocol (MCP) is an open normal developed by Anthropic that allows AI purposes to work together with exterior instruments and providers. It’s like giving your AI assistant a set of fingers to truly do issues quite than simply having them let you know easy methods to do them (take a look at Kareem’s weblog for a primer on MCP). For CML, this implies your AI can create labs, add nodes, configure gadgets, begin topologies, and even run instructions on digital gadgets—all via pure language dialog.
Getting up and working
One of many coolest facets of the CML MCP server is how straightforward it’s to get began. You may be up and working in only a few minutes. Probably the most simple strategy makes use of uvx, which downloads and runs the server in a standalone atmosphere. Simply add just a few strains to your favourite MCP consumer’s configuration, present your CML server credentials, and also you’re able to go.
For individuals who need the complete expertise, together with the power to execute CLI instructions on gadgets working in your CML labs, you’ll be able to set up the PyATS-enabled model as a Docker container.
So, what are you able to truly do?
That is the place issues get attention-grabbing. Let me provide you with an actual instance of what a dialog with an MCP-enabled AI assistant would possibly seem like:
“Create a brand new CML lab referred to as ‘OSPF Examine Lab’. Add two IOL routers, join them via an unmanaged swap, and configure OSPF between them utilizing the 192.0.2.0/24 subnet.”
That’s it. Your AI assistant will work with the MCP server to create the lab, add the nodes, join them, and configure OSPF. Then you definately’ll have a working topology prepared to begin. Need to validate that OSPF is working? Simply ask to examine the neighboring relationships. Want so as to add an annotation to doc the topology? Inform the AI assistant what you need and the place you need it.
The MCP server helps a complete set of operations:
- Creating and managing lab topologies
- Including nodes and connecting them with interfaces and hyperlinks
- Making use of hyperlink conditioning (bandwidth, latency, jitter, packet loss)
- Configuring gadgets with startup configurations
- Beginning, stopping, and wiping labs and particular person nodes
- Executing CLI instructions on working gadgets via PyATS
- Including visible annotations to doc your topologies
- Managing CML customers and teams
Why this issues for studying and testing
For those who’re finding out for CCNA, CCNP, or CCIE certifications, the CML MCP server can dramatically pace up your lab creation course of. As an alternative of clicking via the UI to construct a topology, you’ll be able to describe what you need to study or check, and the AI will construct it for you. Need to check a particular OSPF situation? Describe it. Must validate BGP route choice? Simply ask.
For community engineers testing configurations or troubleshooting situations, the power to shortly spin up advanced topologies via dialog is a game-changer. You may give attention to the networking ideas and configurations quite than the mechanics of constructing the lab atmosphere.
The ability of automation meets conversational AI
What makes the MCP server significantly highly effective is the way it brings collectively CML’s automation capabilities with the pure language understanding of recent AI. Behind the scenes, it’s utilizing the CML API and PyATS to do the heavy lifting, however you don’t have to know Python or YAML (and even have to make use of your mouse!) to learn from it. You simply want to have the ability to describe what you need.
The server is constructed utilizing FastMCP 2.0, which supplies a strong framework for implementing MCP servers. It’s open supply and out there on GitHub, so in the event you’re interested by the way it works or need to contribute enhancements, submit a problem or create a pull request.
Wanting forward
We’re firstly of one thing actually thrilling right here. The intersection of AI brokers and community simulation opens up prospects we’re solely beginning to discover. Think about describing a posh enterprise community topology with a number of websites, VPNs, and redundancy necessities (and having it constructed robotically) or asking your AI assistant to assist troubleshoot why a specific protocol isn’t working as anticipated, with the power to truly examine the configurations and states of your digital gadgets.
The CML MCP server is now out there on PyPI, and you may get began in minutes. Whether or not you’re utilizing the free model of CML or have a subscription, this device can improve your workflow and make community simulation extra accessible than ever.
For those who’re seeking to begin labbing the straightforward approach, the CML MCP server is certainly price testing. Head over to the GitHub repository, comply with the getting began information, and see what you’ll be able to construct with only a dialog.
Blissful labbing!
Join Cisco U. | Be part of the Cisco Studying Community at present free of charge.
Study with Cisco
X | Threads | Fb | LinkedIn | Instagram | YouTube
Use #CiscoU and #CiscoCert to hitch the dialog.
