-
Enhancing Network Management Using Code Generated by Large Language Models
Authors:
Sathiya Kumaran Mani,
Yajie Zhou,
Kevin Hsieh,
Santiago Segarra,
Ranveer Chandra,
Srikanth Kandula
Abstract:
Analyzing network topologies and communication graphs plays a crucial role in contemporary network management. However, the absence of a cohesive approach leads to a challenging learning curve, heightened errors, and inefficiencies. In this paper, we introduce a novel approach to facilitate a natural-language-based network management experience, utilizing large language models (LLMs) to generate t…
▽ More
Analyzing network topologies and communication graphs plays a crucial role in contemporary network management. However, the absence of a cohesive approach leads to a challenging learning curve, heightened errors, and inefficiencies. In this paper, we introduce a novel approach to facilitate a natural-language-based network management experience, utilizing large language models (LLMs) to generate task-specific code from natural language queries. This method tackles the challenges of explainability, scalability, and privacy by allowing network operators to inspect the generated code, eliminating the need to share network data with LLMs, and concentrating on application-specific requests combined with general program synthesis techniques. We design and evaluate a prototype system using benchmark applications, showcasing high accuracy, cost-effectiveness, and the potential for further enhancements using complementary program synthesis techniques.
△ Less
Submitted 11 August, 2023;
originally announced August 2023.
-
A Framework for Enabling Safe and Resilient Food Factories for the Public Feeding Programs
Authors:
Nataraj Kuntagod,
Sanjay Podder,
Satya Sai Srinivas Abbabathula,
Venkatesh Subramanian,
Giju Mathew,
Suresh Kumar Mani
Abstract:
Public feeding programs continue to be a major source of nutrition to a large part of the population across the world. Any disruption to these activities, like the one during the Covid-19 pandemic, can lead to adverse health outcomes, especially among children. Policymakers and other stakeholders must balance the need for continuing the feeding programs while ensuring the health and safety of work…
▽ More
Public feeding programs continue to be a major source of nutrition to a large part of the population across the world. Any disruption to these activities, like the one during the Covid-19 pandemic, can lead to adverse health outcomes, especially among children. Policymakers and other stakeholders must balance the need for continuing the feeding programs while ensuring the health and safety of workers engaged in the operations. This has led to several innovations that leverage advanced technologies like AI and IOT to monitor the health and safety of workers and ensure hygienic operations. However, there are practical challenges in its implementation on a large scale. This paper presents an implementation framework to build resilient public feeding programs using a combination of intelligent technologies. The framework is a result of piloting the technology solution at a facility run as part of a large mid-day meal feeding program in India. Using existing resources like CCTV cameras and new technologies like AI and IOT, hygiene and safety compliance anomalies can be detected and reported in a resource-efficient manner. It will guide stakeholders running public feeding programs as they seek to restart suspended operations and build systems that better adapt to future crises.
△ Less
Submitted 8 March, 2021;
originally announced March 2021.
-
iHorology: Lowering the Barrier to Microsecond-level Internet Time
Authors:
Sathiya Kumaran Mani,
Yi Cao,
Paul Barford,
Darryl Veitch
Abstract:
High precision, synchronized clocks are essential to a growing number of Internet applications. Standard protocols and their associated server infrastructure have been shown to typically enable client clocks to synchronize on the order of tens of milliseconds. We address one of the key challenges to high precision Internet timekeeping - the intrinsic contribution to clock error of path asymmetry b…
▽ More
High precision, synchronized clocks are essential to a growing number of Internet applications. Standard protocols and their associated server infrastructure have been shown to typically enable client clocks to synchronize on the order of tens of milliseconds. We address one of the key challenges to high precision Internet timekeeping - the intrinsic contribution to clock error of path asymmetry between client and time server, a fundamental barrier to microsecond level accuracy. We first exploit results of a measurement study to quantify asymmetry and its effect on timing. We then describe three approaches to addressing the path asymmetry problem: LBBE, SBBE and K-SBBE, each based on timestamp exchange with multiple servers, with the goal of tightening bounds on asymmetry for each client. We explore their capabilities and limitations through simulation and argument. We show that substantial improvements are possible, and discuss whether, and how, the goal of microsecond accuracy might be attained.
△ Less
Submitted 12 November, 2020;
originally announced November 2020.
-
A System for Clock Synchronization in an Internet of Things
Authors:
Sathiya Kumaran Mani,
Ramakrishnan Durairajan,
Paul Barford,
Joel Sommers
Abstract:
Synchronizing clocks on Internet of Things (IoT) devices is important for applications such as monitoring and real time control. In this paper, we describe a system for clock synchronization in IoT devices that is designed to be scalable, flexibly accommodate diverse hardware, and maintain tight synchronization over a range of operating conditions. We begin by examining clock drift on two standard…
▽ More
Synchronizing clocks on Internet of Things (IoT) devices is important for applications such as monitoring and real time control. In this paper, we describe a system for clock synchronization in IoT devices that is designed to be scalable, flexibly accommodate diverse hardware, and maintain tight synchronization over a range of operating conditions. We begin by examining clock drift on two standard IoT prototyping platforms. We observe clock drift on the order of seconds over relatively short time periods, as well as poor clock rate stability, each of which make standard synchronization protocols ineffective. To address this problem, we develop a synchronization system, which includes a lightweight client, a new packet exchange protocol called SPoT and a scalable reference server. We evaluate the efficacy of our system over a range of configurations, operating conditions and target platforms. We find that SPoT performs synchronization 22x and 17x more accurately than MQTT and SNTP, respectively, at high noise levels, and maintains a clock accuracy of within ~15ms at various noise levels. Finally, we report on the scalability of our server implementation through microbenchmark and wide area experiments, which show that our system can scale to support large numbers of clients efficiently.
△ Less
Submitted 6 June, 2018;
originally announced June 2018.
-
TimeWeaver: Opportunistic One Way Delay Measurement via NTP
Authors:
Ramakrishnan Durairajan,
Sathiya Kumaran Mani,
Paul Barford,
Rob Nowak,
Joel Sommers
Abstract:
One-way delay (OWD) between end hosts has important implications for Internet applications, protocols, and measurement-based analyses. We describe a new approach for identifying OWDs via passive measurement of Network Time Protocol (NTP) traffic. NTP traffic offers the opportunity to measure OWDs accurately and continuously from hosts throughout the Internet. Based on detailed examina- tion of NTP…
▽ More
One-way delay (OWD) between end hosts has important implications for Internet applications, protocols, and measurement-based analyses. We describe a new approach for identifying OWDs via passive measurement of Network Time Protocol (NTP) traffic. NTP traffic offers the opportunity to measure OWDs accurately and continuously from hosts throughout the Internet. Based on detailed examina- tion of NTP implementations and in-situ behavior, we develop an analysis tool that we call TimeWeaver, which enables assessment of precision and accuracy of OWD measurements from NTP. We apply TimeWeaver to a ~1TB corpus of NTP traffic collected from 19 servers located in the US and report on the characteristics of hosts and their associated OWDs, which we classify in a precision/accuracy hierarchy. To demonstrate the utility of these measurements, we apply iterative hard-threshold singular value decomposition to estimate OWDs between arbitrary hosts from the high- est tier in the hierarchy. We show that this approach results in highly accurate estimates of OWDs, with average error rates on the order of less than 2%. Finally, we outline a number of applications---in particular, IP geolocation, network operations and management---for hosts in lower tiers of the precision hierarchy that can benefit from TimeWeaver, offering directions for future work.
△ Less
Submitted 6 January, 2018;
originally announced January 2018.
