Self-programming networks : architecture and algorithms
Abstract/Contents
- Abstract
- Networking has been a transformative technology over the past 3 decades, providing "connectivity" to people and businesses, enabling massive information networks and eCommerce, and bringing devices online through IoT technology. By providing "comprehension", Artificial Intelligence (AI) and Machine Learning (ML) promise to be similarly transformative: from image and speech recognition, to chatbots and machine translation, to self-driving cars. Working together, Networking and AI/ML are significantly impacting the business world and every aspect of our daily lives. This thesis describes our research on Self-Programming Networks (SPNs). SPNs aim to bring the power of AI/ML to the sensing, understanding and controlling of the operation of networks. Just as self-driving cars are built from traditional cars by adding sensors, controllers (steering and speed), and a lot of intelligence in the form of AI/ML algorithms, so too we envision SPNs to be built from traditional networks. We restrict our attention to data center networks and cloud computing environments. Our goal is to transform such networks into SPNs by adding suitable sensing and control capabilities, as well as algorithms and systems for sensing, inferring, learning and controlling. An SPN will observe the data emitted by a network during the course of its operation, reconstruct the network's evolution, infer key performance metrics, continually learn the best responses to rapidly-changing load and operating conditions, and help the network adapt to them in real time. We describe the architecture, algorithms and systems we have developed for SPNs. Specifically, we present a NIC-centric architecture for SPNs, describe systems and algorithms for (i) fine-grained network measurement using packet and probe timestamps taken at the edge, and (ii) nanosecond-level clock synchronization in real time and at scale. We also describe how these enhancements can enable new applications on top of SPNs. While our initial work has been on developing the sensing and inference capabilities of SPNs, we also describe future work on learning and control.
Description
Type of resource | text |
---|---|
Form | electronic resource; remote; computer; online resource |
Extent | 1 online resource. |
Place | California |
Place | [Stanford, California] |
Publisher | [Stanford University] |
Copyright date | 2018; ©2018 |
Publication date | 2018; 2018 |
Issuance | monographic |
Language | English |
Creators/Contributors
Author | Geng, Yilong |
---|---|
Degree supervisor | Prabhakar, Balaji, 1967- |
Thesis advisor | Prabhakar, Balaji, 1967- |
Thesis advisor | McKeown, Nick |
Thesis advisor | Rosenblum, Mendel |
Degree committee member | McKeown, Nick |
Degree committee member | Rosenblum, Mendel |
Associated with | Stanford University, Department of Electrical Engineering. |
Subjects
Genre | Theses |
---|---|
Genre | Text |
Bibliographic information
Statement of responsibility | Yilong Geng. |
---|---|
Note | Submitted to the Department of Electrical Engineering. |
Thesis | Thesis Ph.D. Stanford University 2018. |
Location | electronic resource |
Access conditions
- Copyright
- © 2018 by Yilong Geng
- License
- This work is licensed under a Creative Commons Attribution Non Commercial 3.0 Unported license (CC BY-NC).
Also listed in
Loading usage metrics...