0

I'm seeking guidance on simulating a network infrastructure using GNS3, specifically focusing on the challenges related to switch stacks. The network environment consists of multiple physical switches organized in stacks to achieve scalability and streamlined management.

I would like to replicate this network infrastructure within GNS3 for testing and configuration purposes. However, GNS3 lacks native support for switch stacking technology. I'm looking for advice on the best approach to simulate switch stacks in GNS3 or find alternatives to achieve similar outcomes.

If I won't work with stacks, I can understand that I will need different unique IP addresses for each one of the switches, something that will not be representative for the existing network infrastructure that I need to simulate. Here are my specific questions and concerns:

  1. Is it possible to simulate individual switches within a stack in GNS3? How can I configure and test their functionalities independently?
  2. Are there any virtual switch appliances compatible with GNS3 that can emulate switch stacking behavior?
  3. Is there a way to emulate the interconnections between switches in a stack using virtual interfaces and network topologies within GNS3?
  4. What are the limitations or considerations I should be aware of when simulating switch stacks in GNS3 compared to a physical environment?
  5. Are there any recommended tutorials that address simulating switch stacks in GNS3?

I would definitely appreciate your help, support and guidance on the above mentioned questions and concerns.

Improve this question
4
  • Sorry, resource recommendations are explicitly off-topic here, see the help center. Please edit your question accordingly. Also, if you need to test performance then a simulator is the wrong approach - you'll need to build a physical lab for your use cases.
    – Zac67 ♦
    Jul 14 at 11:58
  • @Zac67 Thank you for your immediate reply. There is no option for a physical lab, since the implementation has to do with a large company's network. That is why I chose gns3.
    – wajaap
    Jul 14 at 12:12
  • 1
    A large company's network is precisely when you should have a capable lab environment. Cold spare equipment can be used for lab testing, or you should have a budget for a lab anyway. A large company should consider spares and testing equipment as part of a standard deployment.
    – FrameHowitzer
    Jul 14 at 16:45
  • Possible duplicate of networkengineering.stackexchange.com/questions/71126/…
    – Zac67 ♦
    Jul 15 at 12:39

2 Answers 2

Reset to default
1

GNS3 is really a training/education tool. It's not a network simulator. Moreover, hardware features like stacking is very dependent on the exact hardware, and GNS3 only provides a rather generic simulation of that -- just enough to make the basic software functions work.

If there are specific questions or issues you're investigating regarding switch stacks, consider posting them here and perhaps we can provide you with answers.

Improve this answer
7
  • Thank you for your prompt response, much appreciated. May I ask you if you have a good network simulator to suggest? Moreover, what do you suggest as the best practice in order to simulate a large company's network? Maybe a more general view can be represented by GNS3, for example only the routers, stacks/switches and firewall?
    – wajaap
    Jul 14 at 12:28
  • 1
    The only way to simulate hardware . . . is with hardware.
    – Ron Trunk
    Jul 14 at 12:30
  • Or maybe I can simulate the infrastructure as it is and wherever a stack is needed, I can use only one switch. This means less switches -> less ports -> less users, but still a more general view of the infrastructure.
    – wajaap
    Jul 14 at 12:47
  • 2
    In my experience, trying to 'model' an entire network is not worth the time and effort for the information you obtain. You can simulate specific parts where operations or outcomes are in question
    – Ron Trunk
    Jul 14 at 14:21
  • 2
    Instead of a switch stack, just use a larger model switch. Many companies offer higher port count switches that are similar in features from a 24 to 48 to 96 port switch in a single model line. Alternately, just use one switch and forget about the extra ports. You're not going to simulate 400 PCs in the model anyway right? So just use a single 48 port model.
    – FrameHowitzer
    Jul 14 at 16:48
1

Your choice for GNS3 seems to be not the best one, as others pointed out. It's just an emulator with a limited feature set. I wouldn't trust any results you'd get from that.

If you cannot do this with actual hardware, running virtualised or containerized switches (like for example Arista's vEOS and cEOS or Nokia's SR linux) may be your best option.

Improve this answer
2
  • There is already a hardware since it is a company's infrastructure. The point is to have all that equipment on an emulator to try new ideas before implementing
    – wajaap
    Jul 14 at 13:22
  • 2
    Many people have now tried to explain that what you want to test is not something you test on an emulator. You try to use GNS3 for things it wasn't built for.
    – Teun Vink ♦
    Jul 14 at 14:18

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.