ask for help

Question

ask for help

SkyShiS opened this issue 4 years ago · 7 comments

Hello! This is a great program and I download it, but it misses "request.json". Can you share the file with me? My Email address is shishang@bupt.edu.cn. Thank you very much for your consideration. I am very pleased to hear from you.

Yours sincerely

Answer 1 · 2021-01-23T21:55:38.000Z

Hi, requests.json configures the kind of traffic/requests you want to simulate.

As example requests, see the ones in the results directory, eg, https://github.com/CN-UPB/NFVdeep/blob/main/results/abilene/static/default/requests.json

Or in the data directory.

Does this help?

Answer 2 · 2021-01-24T02:47:24.000Z

On, I see the file. Thanks for help!

Answer 3 · 2021-01-25T04:50:30.000Z

Hello! I get a question that I have already installed tensorflow-gpu2.2.0 and cuda, but when I run the script.py, it said "Using cpu device". How do I set up gpu device to run the program?

Yours sincerely

Answer 4 · 2021-01-25T09:19:38.000Z

We only tested our NFVdeep installation on CPU, not GPU. So I'm not sure if I can help here.

Either way, my understading is that the reinforcement learning part uses stable-baselines3, which in turn builds on PyTorch, not TensorFlow.
Not sure why we have the TF dependency. @NilsLuca @stwerner97 Do you know?

Answer 5 · 2021-01-26T08:49:14.000Z

Hi :) As @stefanbschneider mentioned, we primarily used the CPU installation of PyTorch for stable-baselines3. The TF dependencies could either be deprecated since we migrated from stable-baselines (Tensorflow) to stable-baselines3 (PyTorch) during development. Also, we used TensorBoard for monitoring and testing which requires an installation of TF (as far as I am aware of it).

Answer 6 · 2021-02-03T08:02:00.000Z

Hi, thanks for answering! I have already installed the environment and successfully run the program. You have done a wonderful job! For further study, I want to learn more about what the differences between static and dynamic in requests.json are. And I want to learn about the Abilene network topology. Is there any introductions or graphs about abilene.gpickle? Thanks!

Answer 7 · 2021-02-03T08:27:30.000Z

Thanks! The Abilene network topology is from Topology Zoo. There's more documentation on their website. You'll also find more information about Abilene in general on Google or Wikipedia.

Regarding static/dynamic in requests.json, I think it has to do something with how the requests are generated, but I'm not sure about the details. @stwerner97 @NilsLuca Maybe put a short documentation/description of all fields in requests.json into the Readme?