AI Native Networking

The blending of network connectivity and advanced computing capabilities, both in the cloud as well as at the network edge, paves the way to the advent of self-driving networks, thanks to a comprehensive and data-rich view of the underlying network components.

The first wave (+AI)

In the first wave of network AI research, the focus has been on how to leverage advances in AI technologies to carry out networking task (AI4NET), or how to evolve network technologies to facilitate execution of AI applications such as training (NET4AI).

In this context, the usage of AI has been limited to an addendum to the existing network architecture, i.e., where specific isolated tasks were executed with the addition of AI (aka +AI). As graphically illustrated above, this is the time where since the early 2000, and with significany acceleration on the last decade, AI has been increasedly adopted in IP networking.

The upcoming wave (AI+)

We posit that to fully harness the power of AI, the network can evolve to embrace a larger synergy with AI technologies, that become a fundamental building block of the AI-Native network architecture, where AI is no longer an afterthough (as in +AI) but is rather the starting point of the equation (aka AI+) leading to the confluence of networking and AI and a more interwined evolution path.

Resources

This website contains documentation resources on AI Native networking. Notably,

the vision [TNSM-22] document analyzes the integration of AI from an evolutionary+AI viewpoint and starts outlininig the necessary properties for a radically deeper AI+integration.
the keynote address at [NetSoft-22] further outline this vision in a more graphical manner, and the [CoNEXT-NNI-22b] is a writeup of the AI native agenda
the [ComCom-22] analyzes Natural Language Processing (NLP) techniques applied to the domain of computer programming, to project the expected technical benefits and challenges for NLP applications to network configuration
the preliminary work in [HotNets-22] instantiate a part of the above vision, by tackling multimodal data representation learning

The full documents are accessible here:

[CoNEXT-NNI-22b] Rossi, Dario and Liang, Zhang, "Native Network Intelligence, Fast and Slow" ACM CoNext workshop on Native Network Intelligence (NNI) dec. 2022, Conference
```
@inproceedings{DR:CoNEXT-NNI-22b,
  title = {{Native Network Intelligence, Fast and Slow}},
  author = {Rossi, Dario and Liang, Zhang},
  year = {2022},
  month = dec,
  topic = {ai-native},
  booktitle = {ACM CoNext workshop on Native Network Intelligence (NNI)},
  howpublished = {https://nonsns.github.io/paper/rossi22conext-nni-b.pdf},
  note = {project=huawei}
}
```
As networks have historically been built around connectivity, architectural features concerning quality of service, mobility, security and privacy have been added as afterthoughts – with consequent well known architectural headaches for their later integration. Despite Artificial Intelligence (AI) is more a means to an end, that an architectural feature itself, this is not completely different from what concerns its integration: in particular, while Cloud and Edge computing paradigms made it possible to use AI techniques to relieve part of network operation, however AI is currently little more than an additional tool. This paper describes a vision of future networks, where AI becomes a first class commodity: its founding principle lays around the concept of “fast and slow” type of AI reasoning, each of which offers different types of AI capabilities to process network data. We next outline how these building blocks naturally maps to different network segments, and discuss emerging AI-to-AI communication patterns as we move to more intelligent networks.
[HotNets-22] Houidi, Zied Ben and Azorin, Raphael and Gallo, Massimo and Finamore, Alessandro and Rossi, Dario, "Towards a systematic multi-modal representation learning for network data" ACM HotNets nov. 2022, Conference
```
@inproceedings{DR:HotNets-22,
  title = {{Towards a systematic multi-modal representation learning for network data}},
  author = {Houidi, Zied Ben and Azorin, Raphael and Gallo, Massimo and Finamore, Alessandro and Rossi, Dario},
  year = {2022},
  month = nov,
  booktitle = {ACM HotNets},
  howpublished = {https://nonsns.github.io/paper/rossi22hotnets.pdf},
  note = {project=huawei},
  topic = {ai-native}
}
```
Learning the right representations from complex input data is the key ability of successful machine learning (ML) models. The latter are often tailored to a specific data modality. For example, recurrent neural networks (RNNs) were designed having the processing of sequential data in mind, while convolutional neural networks (CNNs) were designed to exploit spatial correlation naturally present in images. Unlike computer vision (CV) and natural language processing (NLP), each of which targets a single well-defined modality, network ML problems often have a mixture of data modalities as input. Yet, instead of exploiting such abundance, practitioners tend to rely on sub-features thereof, reducing the problem on single modality for the sake of simplicity. In this paper, we advocate for exploiting all the modalities naturally present in network data. As a first step, we observe that network data systematically exhibits a mixture of quantities (e.g., measurements), and entities (e.g., IP addresses, names, etc.). Whereas the former are generally well exploited, the latter are often underused or poorly represented (e.g., with one-hot encoding). We propose to systematically leverage state of the art embedding techniques to learn entity representations, whenever significant sequences of such entities are historically observed. Through two diverse usecases, we show that such entity encoding can benefit and naturally augment classic quantity-based features.
[TNSM-22] Rossi, Dario and Zhang, Liang, "Landing AI on Networks: An equipment vendor viewpoint on Autonomous Driving Networks" In IEEE Transactions on Network and Service Management (TNSM), Vol. 19, sep. 2022, DOI 10.1109/TNSM.2022.3169988 arXiv Journal
```
@article{DR:TNSM-22,
  title = {Landing AI on Networks: An equipment vendor
  viewpoint on Autonomous Driving Networks},
  author = {Rossi, Dario and Zhang, Liang},
  month = sep,
  volume = {19},
  issue = {3},
  year = {2022},
  journal = {IEEE Transactions on Network and Service Management (TNSM)},
  doi = {10.1109/TNSM.2022.3169988},
  howpublished = {https://nonsns.github.io/paper/rossi22tnsm.pdf},
  arxiv = {https://arxiv.org/abs/2205.08347},
  note = {project=huawei},
  topic = {ai-native}
}
```
The tremendous achievements of Artificial Intelligence (AI) in computer vision, natural language processing, games and robotics, has extended the reach of the AI hype to other fields: in telecommunication networks, the long term vision is to let AI fully manage, and autonomously drive, all aspects of network operation. In this industry vision paper, we discuss challenges and opportunities of Autonomous Driving Network (ADN) driven by AI technologies. To understand how AI can be successfully landed in current and future networks, we start by outlining challenges that are specific to the networking domain, putting them in perspective with advances that AI has achieved in other fields. We then present a system view, clarifying how AI can be fitted in the network architecture. We finally discuss current achievements as well as future promises of AI in networks, mentioning roadmap to avoid bumps in the road that leads to true large-scale deployment of AI technologies in network
[ComCom-22] Houidi, Zied Ben and Rossi, Dario, " Neural language models for network configuration: Opportunities and reality check" In Elsevier Computer Communication, Vol. (to appear), pp.Pages 118-125, sep. 2022, DOI https://doi.org/10.1016/j.comcom.2022.06.035 arXiv Journal
```
@article{DR:ComCom-22,
  title = { Neural language models for network configuration: Opportunities and reality check},
  author = {Houidi, Zied Ben and Rossi, Dario},
  month = sep,
  issue = {193},
  pages = {Pages 118-125},
  year = {2022},
  journal = {Elsevier Computer Communication},
  volume = {(to appear)},
  howpublished = {https://nonsns.github.io/paper/rossi22comcom.pdf},
  doi = {https://doi.org/10.1016/j.comcom.2022.06.035},
  arxiv = {https://arxiv.org/abs/2205.01398},
  note = {project=huawei},
  topic = {ai-native}
}
```
Boosted by deep learning, natural language processing (NLP) techniques have recently seen spectacular progress, mainly fueled by breakthroughs both in representation learning with word embeddings (e.g. word2vec) as well as novel architectures (e.g. transformers).This success quickly invited researchers to explore the use of NLP techniques to other field, such as computer programming languages, with the promise to automate tasks in software programming (bug detection, code synthesis, code repair, cross language translation etc.). By extension, NLP has potential for application to network configuration languages as well, for instance considering tasks such as network configuration verification, synthesis, and cross-vendor translation. In this paper, we survey recent advances in deep learning applied to programming languages, for the purpose of code verification, synthesis and translation: in particularly, we review their training requirements and expected performance, and qualitatively assess whether similar techniques can benefit corresponding use-cases in networking.