A multi-vendor network lab is a single test topology where devices from more than one vendor, such as Cisco routers, Juniper switches and a Palo Alto firewall, run and talk to each other. You need one when a change, a migration rehearsal or a vendor POC crosses vendor lines. Where the lab runs decides most of what can go wrong, so ask each provider the nine questions below before you commit. CloudMyLab sells hosted emulators and Lab as a Service, so read our examples with that in mind.
Many hosted labs run the images you bring. A good answer says which images you upload, which the provider supplies, and what help with uploading and booting them is included, in writing.
Every vendor sets its own terms, and some licence agreements restrict hosting the software for a third party. Cisco, for example, says the images that come with CML "are only licensed for use within Cisco Modeling Labs" (Cisco CML image documentation). Read each vendor's hosting clause yourself. A provider cannot settle it for you.
An image that boots in one emulator can fail in another. Ask for a named platform per image, at your major versions, and which platform the provider recommends for your image list.
Some images refuse to run nested inside another virtual machine. Juniper states that vJunos-switch "cannot be used in any deployments that launch the instances from within a VM" (Juniper vJunos-switch KVM guide), so ask which one your plan uses.
Router and firewall VMs carry far more weight than containers, and a topology that boots one node at a time can still stall when everything starts together. The strongest answer is an offer to boot your full topology before you commit.
Lab configs often carry production credentials, keys and addresses. Ask where images, configs, snapshots and logs are stored, who on the provider's side can access them, and what is deleted at teardown.
A provider can keep the host running around the clock and still treat an image that will not boot as your problem. Ask whether support covers your imported images and in-lab routing, or only the platform.
Cisco says CML is "not generally suitable for testing throughput or timing" (Cisco CML image documentation), and hardware behaviour such as ASICs, optics and failover timing stays on physical equipment. A good provider tells you what still needs hardware. For how to structure the test itself, see our guide to safer changes in a managed virtual lab.
If the provider keeps the files when the engagement ends, you cannot rerun the test or show an auditor what was tested. Ask for formats you can rerun yourself.
Copy this multi-vendor network lab checklist into your vendor-call notes.
| Check | Question to ask | Good answer | Red flag |
| Images | Who supplies each vendor's image? | You bring them; upload and boot help stated in writing | "We include all vendor images" |
| Licences | Whose licence covers each image, and does it permit hosting here? | A named licence holder per image, and the hosting clause checked | "Licences are covered", with no detail per vendor |
| Emulator fit | Which emulator runs my topology, and has it run each of my images? | A named platform per image, with known limits | "Any image runs on any emulator" |
| Bare metal | Does the emulator run on bare metal or inside a VM? | A clear answer per plan | No answer, or "it does not matter" |
| Resources | Can the server run my heaviest nodes at the same time? | An offer to boot the full topology first | "Unlimited nodes" |
| Isolation | Where do configs and images live, and who can reach them? | Named storage, named access path, deletion at teardown | "It is private" with no detail |
| Support | Which SLA applies to my plan, and does support cover my images? | Written scope, including what is excluded | "Support is 24/7" and nothing more |
| Test validity | What will a passing result not prove? | A list of what still needs hardware | "If it works here, it works in production" |
| Handover | Can I export the topology, configs and test evidence? | Yes, in formats you can rerun | "The topology stays with us" |
If your lab serves recurring training courses instead, our provider questions for recurring training labs cover cohorts, resets and live-class support.
Nothing above requires a provider. If you have the hardware and the time, you can run a multi-vendor network lab on a self-hosted emulator and use the same checks on your own setup.
If you would rather not run the lab host, CloudMyLab covers it in four ways. With Hosted Emulators, you build and we host: EVE-NG Pro, GNS3 or CML with your own images (CML runs on your own Cisco licence), on virtual machines or bare metal servers. We are an official EVE-NG hosting partner, tested and approved by EVE-NG Ltd. With Lab as a Service, we build and change the topology with you, and your team connects through Cisco AnyConnect with DUO MFA. Learning Labs are "highly customizable Physical Topologies" for certification practice. For a vendor bake-off, see our Proof of Concept service.
Ask us the same nine questions. The free 24-hour trial covers hosted EVE-NG Pro and CML, which is enough to try your topology on either platform. For a managed topology or a POC, contact the team with your vendor list and the change you need to test.
A multi-vendor network lab is a test topology where devices from more than one vendor, such as Cisco, Juniper, Arista and Palo Alto, run together. Teams use one to test changes, rehearse migrations and run vendor POCs before touching production.
Yes, on an emulator that supports each image, such as EVE-NG, GNS3, CML with custom images, or Containerlab. Each vendor's image still needs its own licence, and some images, such as vJunos-switch, need the emulator to run on bare metal, outside any VM.
Usually, yes. Hosted emulator providers generally run the images you bring, and the licence stays with you. Each vendor sets its own terms, and some restrict hosting by a third party, so check each vendor's licence agreement.
It depends on the heaviest images running at the same time. Router and firewall VMs need far more CPU and memory than container-based images, so size the lab by the nodes that run together and confirm each vendor's own requirements.