arXiv subjects

Michael Luby

Publications and source records attributed to Michael Luby.

11 recordsLinked to original sources

arXiv2026

FLUID: Slack-based Low-latency Delivery

We introduce FLUID (Fountain LiqUId Delivery), a protocol that uses fountain coding and receiver feedback for low-latency delivery of data blocks over lossy networks. Idealized Automatic Repeat reQuest (ARQ) protocols are bandwidth-optimal, but must deliver every packet in a block and therefore can require additional rounds under packet loss. FLUID uses a controlled amount of slack to relax this all-packets requirement, allowing delivery to finish once enough encoded packets have been received. This yields substantially tighter delivery latency while remaining deterministically close to the ARQ bandwidth optimum. FLUID is controlled by a slack parameter $\epsilon$. Under the Loss-Product Rule, delivery finishes once the product of packet loss fractions across transmission rounds falls below $\epsilon$. Thus, FLUID can finish delivery in a small number of rounds even when every round experiences packet loss, while $\epsilon$ controls the gap between FLUID and bandwidth-optimal ARQ.

cs.NI↗

arXiv2025

Local Dominance in Mixed-Strength Populations -- Fast Maximal Independent Set

In many natural and engineered systems, agents interact through local contests that determine which individuals become dominant within their neighborhoods. These interactions are shaped by inherent differences in strength, and they often lead to stable dominance patterns that emerge surprisingly quickly relative to the size of the population. This motivates the search for simple mathematical models that capture both heterogeneous agent strength and rapid convergence to stable local dominance. A widely studied abstraction of local dominance is the Maximal Independent Set (MIS) problem. In the Luby MIS protocol that provably converges quickly to an MIS, each agent repeatedly generates a strength value chosen uniformly and becomes locally dominant if its value is smaller than those of its neighbors. This provides a theoretical explanation for fast dominance convergence in populations of equal-strength agents and naturally raises the question of whether fast convergence also holds in the more realistic setting where agents are inherently mixed-strength. To investigate this question, we introduce the mixed-strength agents model, in which each agent draws its strength from its own distribution. We prove that the extension of the Luby MIS protocol where each agent repeatedly generates a strength value from its own distribution still exhibits fast dominance convergence, providing formal confirmation of the rapid convergence observed in many mixed-strength natural processes. We also show that heterogeneity can significantly change the dynamics of the process. In contrast to the equal-strength setting, a constant fraction of edges need not be eliminated per round. We construct a population and strength profile in which progress per round is asymptotically smaller, illustrating how inherent strength asymmetry produces qualitatively different global behavior.

cs.MA↗

arXiv2025

Singing a MIS

We introduce a broadcast model called the singing model, where agents are oblivious of the size and structure of the communication network, even their immediate neighborhood. Agents can sing multiple notes which are heard by their neighbors. The model is a generalization of the beeping model, where agents can only emit sound at a single frequency. We give a simple and natural protocol where agents compete with their neighbors and their strength is reflected in the number of notes they sing. It converges in $O(log(n))$ time with high probability, where $n$ is the number of agents in the network. The protocol works in an asynchronous model where rounds vary in length and have different start times. It works with completely dynamic networks where agents can be faulty. The protocol is the first to converge to an MIS in logarithmic time for dynamic networks in a network oblivious model.

cs.DS↗

arXiv2025

ARC: Consistent, Low-Latency Delivery via Receiver-Side Scheduling

Applications such as cloud gaming, video streaming, telemetry, ML inference, and data transfer provide a better experience when data is released at the receiver with timing reflecting how the data enters the sender. In practice, network delay variation and recovery dynamics at the receiver distort this timing even when transports deliver all packets correctly, producing visible jitter, stalls, and unstable playback. Many such applications operate best when delivery preserves this timing behavior and its implied order; out-of-order or irregular delivery can significantly degrade performance even when all data eventually arrives. We present a lightweight receiver-side release scheduling protocol, Adaptive Release Control (ARC), that restores this timing at the receiver. ARC releases recovered data in a manner that follows the sender's timing, maintaining ordering and limiting reordering when necessary while producing smooth delivery with minimal added latency given network conditions. It operates entirely on the receiver clock and requires no feedback, synchronization, or changes to the underlying transport. As an example, we integrate ARC into LT3, a network-layer system currently deployed as a software overlay that forwards traffic without altering the transport protocols it carries, where ARC functions as an independent module that regulates release timing for forwarded data. Evaluating LT3 with ARC on a cloud-gaming workload shows that the protocol removes virtually all large jitter excursions and yields release intervals that closely match the sender's timing, translating into improved perceptual smoothness. Broader latency improvements arise from the behavior of the full LT3 system. The benefits of ARC extend to transport protocols carried over LT3, including TCP, QUIC, WebRTC, UDP, and RTP, as preserving sender timing improves their behavior across a wide range of conditions.

cs.NI↗

arXiv2025

Whack-a-Mole: Deterministic Packet Spraying Across Multiple Network Paths

We present Whack-a-Mole, a deterministic packet spraying algorithm for distributing packets across multiple network paths with provably tight discrepancy bounds. The algorithm is motivated by large-scale distributed AI/ML training and inference workloads, where collective completion time (CCT) and effective training time ratio (ETTR) are highly sensitive to tail latency and transport imbalance. Whack-a-Mole represents the path profile as a discrete allocation of $m$ selection units across $n$ paths and uses a bit-reversal counter to choose a path for each packet. We prove that the discrepancy between expected and actual packet counts per path is bounded by $O(\log m)$ over any contiguous packet sequence. The algorithm responds quickly to congestion feedback by reducing allocations to degraded paths and redistributing load to healthier ones. This combination of deterministic distribution, low per-packet overhead, and compatibility with erasure-coded transport makes Whack-a-Mole an effective building block for multipath transport protocols that aim to minimize CCT and maximize GPU utilization.

cs.NI↗

arXiv2023

Enabling immersive experiences in challenging network conditions

Immersive experiences, such as remote collaboration and augmented and virtual reality, require delivery of large volumes of data with consistent ultra-low latency across wireless networks in fluctuating network conditions. We describe the high-level design behind a data delivery solution that meets these requirements and provide synthetic simulations and test results running in network conditions based on real-world measurements demonstrating the efficacy of the solution.

cs.NI↗

arXiv2021

Distributed storage algorithms with optimal tradeoffs

One of the primary objectives of a distributed storage system is to reliably store large amounts of source data for long durations using a large number $N$ of unreliable storage nodes, each with $c$ bits of storage capacity. Storage nodes fail randomly over time and are replaced with nodes of equal capacity initialized to zeroes, and thus bits are erased at some rate $e$. To maintain recoverability of the source data, a repairer continually reads data over a network from nodes at an average rate $r$, and generates and writes data to nodes based on the read data. The distributed storage source capacity is the maximum amount of source that can be reliably stored for long periods of time. Previous research shows that asymptotically the distributed storage source capacity is at most $\left(1-\frac{e}{2 \cdot r}\right) \cdot N \cdot c$ as $N$ and $r$ grow. In this work we introduce and analyze algorithms such that asymptotically the distributed storage source data capacity is at least the above equation. Thus, the above equation expresses a fundamental trade-off between network traffic and storage overhead to reliably store source data.

cs.IT↗

arXiv2020

SOPI design and analysis for LDN

Liquid Data Networking (LDN) is an ICN architecture that is designed to enable the benefits of erasure-code enabled object delivery. A primary contribution of LDN is the introduction of SOPIs, which enables client s to concurrently download encoded data for the same object from multiple edge nodes, optimizes caching efficiency, and enables seamless mobility. This paper provides an enhanced design and analysis of SOPI s.

cs.NI↗

arXiv2020

Repair rate lower bounds for distributed storage

One of the primary objectives of a distributed storage system is to reliably store a large amount $dsize$ of source data for a long duration using a large number $N$ of unreliable storage nodes, each with capacity $nsize$. The storage overhead $β$ is the fraction of system capacity available beyond $dsize$, i.e., $β= 1- \frac{dsize}{N \cdot nsize}$. Storage nodes fail randomly over time and are replaced with initially empty nodes, and thus data is erased from the system at an average rate $erate = λ\cdot N \cdot nsize$, where $1/λ$ is the average lifetime of a node before failure. To maintain recoverability of the source data, a repairer continually reads data over a network from nodes at some average rate $rrate$, and generates and writes data to nodes based on the read data. The main result is that, for any repairer, if the source data is recoverable at each point in time then it must be the case that $rrate \ge \frac{erate}{2 \cdot β}$ asymptotically as $N$ goes to infinity and beta goes to zero. This inequality provides a fundamental lower bound on the average rate that any repairer needs to read data from the system in order to maintain recoverability of the source data.

cs.IT↗

arXiv2018

Capacity bounds for distributed storage

One of the primary objectives of a distributed storage system is to reliably store large amounts of source data for long durations using a large number $N$ of unreliable storage nodes, each with $c$ bits of storage capacity. Storage nodes fail randomly over time and are replaced with nodes of equal capacity initialized to zeroes, and thus bits are erased at some rate $E$. To maintain recoverability of the source data, a repairer continually reads data over a network from nodes at a rate $R$, and generates and writes data to nodes based on the read data. The distributed storage source data capacity is the maximum amount of source data that can be reliably stored for long periods of time. We prove the distributed storage source data capacity asymptotically approaches $\left(1-\frac{E}{2 \cdot R}\right) \cdot N \cdot c$ as $N$ and $R$ grow. This equation expresses a fundamental trade-off between network traffic and storage overhead to reliably store source data.

cs.IT↗

arXiv2018

Learning from partial correction

We introduce a new model of interactive learning in which an expert examines the predictions of a learner and partially fixes them if they are wrong. Although this kind of feedback is not i.i.d., we show statistical generalization bounds on the quality of the learned model.

cs.LG↗