Self-programming networks : measuring and controlling networks from the edge
Abstract/Contents
- Abstract
- The Self-Programming Networks (SPNs) research project aims to build autonomous networks; i.e., networks which sense, monitor, program and control themselves. SPNs use an "edge-centric" architecture, where sensing and controlling are primarily conducted at the network's edge. Prior work on SPNs has yielded two key systems: (i) Huygens, a highly accurate and scalable algorithm for network clock synchronization, and (ii) SIMON, a system for fine-grained network telemetry using observations at the network's edge. These systems have been applied in distributed computing -- e.g., for building faster distributed databases and scalable state-machine replication, and networking -- e.g., for building "jitter-free networks" which underlie financial trading exchanges. In this dissertation, I present my contributions to SPNs. First, by building on SIMON, I present the methods of reconstructing more detailed network states, extending its deployability, and reducing its overhead in computation and storage. Second, I present a novel approach, called On-Ramp, for controlling congestion in SPNs. Akin to on-ramp meters at freeways, On-Ramp "holds" or "pauses" traffic at the network's edge if congestion on the traffic's path is high. Congestion is measured using accurately synchronized clocks at the sender and receiver. On-Ramp operates as a universal underlay: it augments, and greatly improves, the performance of any existing congestion-control protocol during periods of transient overload. It can be deployed by cloud users in their virtual machines (VMs) without any in-network support. I demonstrate the effectiveness of On-Ramp on Google Cloud, CloudLab, ns-3 simulations, and the production clusters in Facebook. Finally, I conclude the dissertation by describing another SPN-enabled application, ClockChain, which scales the transaction ordering service in permissioned distributed ledgers.
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 | 2021; ©2021 |
| Publication date | 2021; 2021 |
| Issuance | monographic |
| Language | English |
Creators/Contributors
| Author | Liu, Shiyu |
|---|---|
| Degree supervisor | Prabhakar, Balaji, 1967- |
| Thesis advisor | Prabhakar, Balaji, 1967- |
| Thesis advisor | Alizadeh, Mohammad, (Professor) |
| Thesis advisor | Rosenblum, Mendel |
| Degree committee member | Alizadeh, Mohammad, (Professor) |
| Degree committee member | Rosenblum, Mendel |
| Associated with | Stanford University, Department of Electrical Engineering |
Subjects
| Genre | Theses |
|---|---|
| Genre | Text |
Bibliographic information
| Statement of responsibility | Shiyu Liu. |
|---|---|
| Note | Submitted to the Department of Electrical Engineering. |
| Thesis | Thesis Ph.D. Stanford University 2021. |
| Location | https://purl.stanford.edu/pf355zz9227 |
Access conditions
- Copyright
- © 2021 by Shiyu Liu
- 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...