Optimizing Throughput on Guaranteed-Bandwidth WAN Networks for the Large Synoptic Survey Telescope (LSST)
D. Michael Freemon

TL;DR
This paper explores how to maximize data transfer throughput on guaranteed-bandwidth WANs for the LSST project, addressing latency constraints and proposing an approach to achieve near wire-speed data rates.
Contribution
It introduces a novel method for optimizing throughput on guaranteed-bandwidth WANs, specifically tailored for the LSST's stringent latency and throughput requirements.
Findings
Achieves near wire-speed data transfer rates.
Identifies conditions for optimal throughput on long-distance networks.
Addresses limitations of prior TCP congestion control methods.
Abstract
The Large Synoptic Survey Telescope (LSST) is a proposed 8.4-meter telescope that will be located in the Andes mountains in Chile. Every 17 seconds, a 6.4 GB image is transferred to Illinois for immediate processing. That transfer needs to complete within approximately five seconds. LSST is provisioning an international WAN with a 10Gbps bandwidth guarantee for this and other project-related data transfers. The stringent latency requirement drives a re-examination of TCP congestion control for this use case. Specifically, prior work on dedicated Long Fat Networks (LFNs) does not go far enough in fully leveraging the opportunity provided by the bandwidth guarantee. This paper presents an approach for how optimal network throughput can be obtained for the LSST use case, and the conditions under which any project can achieve data throughput rates on long-distance networks approaching wire…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsInterconnection Networks and Systems · Advanced Optical Network Technologies · Network Traffic and Congestion Control
