publications

2023

1. ACM HotNets

   Boosting Application Performance Using Heterogeneous Virtual Channels: Challenges and Opportunities

   Talal Touseef, William Sentosa, Milind Kumar Vaddiraju, and 4 more authors

   In ACM HotNets, 2023

   Abs PDF

   Interactive networked applications require high throughput, low latency, and high reliability from the network to provide a seamless user experience. While meeting these three requirements simultaneously is difficult, there has been an emergence of heterogeneous virtual channels (HVCs) which support some subset of them at the expense of the others. For instance, URLLC sacrifices throughput to achieve low latency and reliability in 5G NR, and Wi-Fi 7 and other novel Internet architectures provide similar disparate types of service. Prior work either focuses on aggregating the bandwidth of these channels whilst neglecting their unique properties or fails to generalize in the sense of achieving high performance across different applications and channels. To utilize HVCs to their fullest, we argue that there are challenges and opportunities across the network, transport and application layers, and the application-transport interface of the network stack. In this work, we explore the trade-offs of these architectural choices in the context of web browsing and real-time video, and identify the constituting principles of a design that is general, performant, and deployable.

2. ArXiv

   T3P: Demystifying Low-Earth Orbit Satellite Broadband

   Shubham Tiwari, Saksham Bhushan, Aryan Taneja, and 8 more authors

   In ArXiv, 2023

   Abs PDF

   The Internet is going through a massive infrastructural revolution with the advent of low-flying satellite networks, 5/6G, WiFi7, and hollow-core fiber deployments. While these networks could unleash enhanced connectivity and new capabilities, it is critical to understand the performance characteristics to efficiently drive applications over them. Low-Earth orbit (LEO) satellite mega-constellations like SpaceX Starlink aim to offer broad coverage and low latencies at the expense of high orbital dynamics leading to continuous latency changes and frequent satellite hand-offs. This paper aims to quantify Starlink’s latency and its variations and components using a real testbed spanning multiple latitudes from the North to the South of Europe. We identify tail latencies as a problem. We develop predictors for latency and throughput and show their utility in improving application performance by up to 25%. We also explore how transport protocols can be optimized for LEO networks and show that this can improve throughput by up to 115% (with only a 5% increase in latency). Also, our measurement testbed with a footprint across multiple locations offers unique trigger-based scheduling capabilities that are necessary to quantify the impact of LEO dynamics.

3. AAAI

   Simulating Network Paths with Recurrent Buffering Units

   Divyam Anshumaan, Sriram Balasubramanian, Shubham Tiwari, and 3 more authors

   In AAAI, 2023

   Abs PDF

   Simulating physical network paths (e.g., Internet) is a cornerstone research problem in the emerging sub-field of AI-for-networking. We seek a model that generates end-to-end packet delay values in response to the time-varying load offered by a sender, which is typically a function of the previously output delays. The problem setting is unique, and renders the state-of-the-art text and time-series generative models inapplicable or ineffective. We formulate an ML problem at the intersection of dynamical systems, sequential decision making, and time-series modeling. We propose a novel grey-box approach to network simulation that embeds the semantics of physical network path in a new RNN-style model called Recurrent Buffering Unit, providing the interpretability of standard network simulator tools, the power of neural models, the efficiency of SGD-based techniques for learning, and yielding promising results on synthetic and real-world network traces.

2022

1. JCN

   qMon: A method to monitor queueing delay in OpenFlow networks

   Sandhya Rathee, Shubham Tiwari, K. Haribabu, and 1 more author

   Journal of Communications and Networks, 2022

   Abs PDF

   In software-defined networking (SDN), the decoupled architecture provides opportunities for efficiently measuring critical quality of service (QoS) parameters, such as delay. Existing approaches, to dynamically obtain delay, are based around calculating the transit time of a probe packet that travels through the data links. These approaches are not efficient as the probe packet injected into the data plane incurs considerable overhead. Additionally, a separate probe packet is required to measure the delay of each queue if more than one queue is present on the egress port of a switch. Thus, these approaches are not scalable. In this paper, we propose an efficient passive delay estimation method, queueing delay monitoring (qMon), to monitor queueing delay in SDN networks. qMon leverages the OpenFlow protocol to obtain queue statistics from switches at regular intervals, which are further employed to estimate the mean queueing delay for each interval. Thus, the proposed approach differs from the existing approaches as no packet is injected into the data plane to measure delay. The results show that for Poisson traffic and for bursty traffic with large ON intervals, round trip time (RTT) values estimated using qMon and ping utility demonstrate high correlation when the measured RTT value is considered as time-series data.

2. ACM SIGMETRICS

   Data-Driven Network Path Simulation with iBox

   Sachin Ashok, Shubham Tiwari, Nagarajan Natarajan, and 2 more authors

   In ACM SIGMETRICS, 2022

   Abs PDF

   While network simulation is widely used for evaluating network protocols and applications, ensuring realism remains a key challenge. There has been much work on simulating network mechanisms faithfully (e.g., links, buffers, etc.), but less attention on the critical task of configuring the simulator to reflect reality. We present iBox ("Internet in a Box"), which enables data-driven network path simulation, using input/output packet traces gathered at the sender/receiver in the target network to create a model of the end-to-end behaviour of a network path. Our work builds on recent work in this direction and makes three contributions: (1) estimation of a lightweight non reactive cross-traffic model, (2) estimation of a more powerful reactive cross-traffic model based on Bayesian optimization, and (3) evaluation of iBox in the context of congestion control variants in an Internet research testbed and also controlled experiments with known ground truth.