arXiv subjects

Prateesh Goyal

Publications and source records attributed to Prateesh Goyal.

13 recordsLinked to original sources

arXiv2025

m4: A Learned Flow-level Network Simulator

Flow-level simulation is widely used to model large-scale data center networks due to its scalability. Unlike packet-level simulators that model individual packets, flow-level simulators abstract traffic as continuous flows with dynamically assigned transmission rates. While this abstraction enables orders-of-magnitude speedup, it is inaccurate by omitting critical packet-level effects such as queuing, congestion control, and retransmissions. We present m4, an accurate and scalable flow-level simulator that uses machine learning to learn the dynamics of the network of interest. At the core of m4 lies a novel ML architecture that decomposes state transition computations into distinct spatial and temporal components, each represented by a suitable neural network. To efficiently learn the underlying flow-level dynamics, m4 adds dense supervision signals by predicting intermediate network metrics such as remaining flow size and queue length during training. m4 achieves a speedup of up to 104$\times$ over packet-level simulation. Relative to a traditional flow-level simulation, m4 reduces per-flow estimation errors by 45.3% (mean) and 53.0% (p90). For closed-loop applications, m4 accurately predicts network throughput under various congestion control schemes and workloads.

cs.NI↗

arXiv2025

Fast Userspace Networking for the Rest of Us

After a decade of research in userspace network stacks, why do new solutions remain inaccessible to most developers? We argue that this is because they ignored (1) the hardware constraints of public cloud NICs (vNICs) and (2) the flexibility required by applications. Concerning the former, state-of-the-art proposals rely on specific NIC features (e.g., flow steering, deep buffers) that are not broadly available in vNICs. As for the latter, most of these stacks enforce a restrictive execution model that does not align well with cloud application requirements. We propose a new userspace network stack, Machnet, built for public cloud VMs. Central to Machnet is a new ''Least Common Denominator'' model, a conceptual NIC with a minimal feature set supported by all kernel-bypass vNICs. The challenge is to build a new solution with performance comparable to existing stacks while relying only on basic features (e.g., no flow steering, no RSS reconfiguration). Machnet uses a microkernel design to provide higher flexibility in application execution compared to a library OS design; we show that microkernels' inter-process communication overhead is negligible on large cloud networks.

cs.NI↗

arXiv2024

Ethereal: Divide and Conquer Network Load Balancing in Large-Scale Distributed Training

Large-scale distributed training in production datacenters constitutes a challenging workload bottlenecked by network communication. In response, both major industry players (e.g., Ultra Ethernet Consortium) and parts of academia have surprisingly, and almost unanimously, agreed that packet spraying is \emph{necessary} to improve the performance of large-scale distributed training workloads. In this paper, we challenge this prevailing belief and pose the question: \emph{How close can singlepath transport come to matching the performance of packet spraying?} We demonstrate that singlepath transport (from a NIC's perspective) is sufficient and can perform nearly as well as ideal packet spraying, particularly in the context of distributed training in CLOS-based topologies. Our assertion is based on four key observations about workloads driven by collective communication patterns: \emph{(i)} flow sizes are known upon arrival, \emph{(ii)} flow sizes are equal within each step of a collective, \emph{(iii)} the completion time of a collective is more critical than individual flow completion times, and \emph{(iv)} flows can be \emph{split} upon arrival to control load balancing directly from the application layer. We present Ethereal, a simple distributed load balancing algorithm that opportunistically splits flows and assigns paths to each flow in a transparent manner, requiring little to no changes to existing RDMA NICs. Our evaluation, spanning a wide range of collective communication algorithms and GPT models using Astra-Sim, shows that Ethereal significantly reduces the completion times by up to $30\%$ compared to packet spraying and by up to $40\%$ compared to REPS, even under link failures. This paper offers an alternative perspective for developing next-generation transport protocols tailored to large-scale distributed training.

cs.NI↗

arXiv2023

DBO: Response Time Fairness for Cloud-Hosted Financial Exchanges

In this paper, we consider the problem of hosting financial exchanges in the cloud. Financial exchanges require predictable, equal latency to all market participants to ensure fairness for various tasks, such as high speed trading. However, it is extremely difficult to ensure equal latency to all market participants in existing cloud deployments, because of various reasons, such as congestion, and unequal network paths. In this paper, we address the unfairness that stems from lack of determinism in cloud networks. We argue that predictable or bounded latency is not necessary to achieve fairness. Inspired by the use of logical clocks in distributed systems, we present Delivery Based Ordering (DBO), a new approach that ensures fairness by instead correcting for differences in latency to the participants. We evaluate DBO both in our hardware test bed and in a public cloud deployment and demonstrate that it is feasible to achieve guaranteed fairness and sub-100 microsecond latency while operating at high transaction rates.

cs.NI↗

arXiv2022

Scalable Tail Latency Estimation for Data Center Networks

In this paper, we consider how to provide fast estimates of flow-level tail latency performance for very large scale data center networks. Network tail latency is often a crucial metric for cloud application performance that can be affected by a wide variety of factors, including network load, inter-rack traffic skew, traffic burstiness, flow size distributions, oversubscription, and topology asymmetry. Network simulators such as ns-3 and OMNeT++ can provide accurate answers, but are very hard to parallelize, taking hours or days to answer what if questions for a single configuration at even moderate scale. Recent work with MimicNet has shown how to use machine learning to improve simulation performance, but at a cost of including a long training step per configuration, and with assumptions about workload and topology uniformity that typically do not hold in practice. We address this gap by developing a set of techniques to provide fast performance estimates for large scale networks with general traffic matrices and topologies. A key step is to decompose the problem into a large number of parallel independent single-link simulations; we carefully combine these link-level simulations to produce accurate estimates of end-to-end flow level performance distributions for the entire network. Like MimicNet, we exploit symmetry where possible to gain additional speedups, but without relying on machine learning, so there is no training delay. On large-scale networks where ns-3 takes 11 to 27 hours to simulate five seconds of network behavior, our techniques run in one to two minutes with 99th percentile accuracy within 9% for flow completion times.

cs.NI↗

arXiv2022

Optimal Congestion Control for Time-varying Wireless Links

Modern networks exhibit a high degree of variability in link rates. Cellular network bandwidth inherently varies with receiver motion and orientation, while class-based packet scheduling in datacenter and service provider networks induces high variability in available capacity for network tenants. Recent work has proposed numerous congestion control protocols to cope with this variability, offering different tradeoffs between link utilization and queuing delay. In this paper, we develop a formal model of congestion control over time-varying links, and we use this model to derive a bound on the performance of any congestion control protocol running over a time-varying link with a given distribution of rate variation. Using the insights from this analysis, we derive an optimal control law that offers a smooth tradeoff between link utilization and queuing delay. We compare the performance of this control law to several existing control algorithms on cellular link traces to show that there is significant room for optimization.

cs.NI↗

arXiv2021

SWP: Microsecond Network SLOs Without Priorities

The increasing use of cloud computing for latency-sensitive applications has sparked renewed interest in providing tight bounds on network tail latency. Achieving this in practice at reasonable network utilization has proved elusive, due to a combination of highly bursty application demand, faster link speeds, and heavy-tailed message sizes. While priority scheduling can be used to reduce tail latency for some traffic, this comes at a cost of much worse delay behavior for all other traffic on the network. Most operators choose to run their networks at very low average utilization, despite the added cost, and yet still suffer poor tail behavior. This paper takes a different approach. We build a system, swp, to help operators (and network designers) to understand and control tail latency without relying on priority scheduling. As network workload changes, swp is designed to give real-time advice on the network switch configurations needed to maintain tail latency objectives for each traffic class. The core of swp is an efficient model for simulating the combined effect of traffic characteristics, end-to-end congestion control, and switch scheduling on service-level objectives (SLOs), along with an optimizer that adjusts switch-level scheduling weights assigned to each class. Using simulation across a diverse set of workloads with different SLOs, we show that to meet the same SLOs as swp provides, FIFO would require 65% greater link capacity, and 79% more for scenarios with tight SLOs on bursty traffic classes.

cs.NI↗

arXiv2020

Site-to-Site Internet Traffic Control

Queues allow network operators to control traffic: where queues build, they can enforce scheduling and shaping policies. In the Internet today, however, there is a mismatch between where queues build and where control is most effectively enforced; queues build at bottleneck links that are often not under the control of the data sender. To resolve this mismatch, we propose a new kind of middlebox, called Bundler. Bundler uses a novel inner control loop between a sendbox (in the sender's site) and a receivebox (in the receiver's site) to determine the aggregate rate for the bundle, leaving the end-to-end connections and their control loops intact. Enforcing this sending rate ensures that bottleneck queues that would have built up from the bundle's packets now shift from the bottleneck to the sendbox. The sendbox then exercises control over its traffic by scheduling packets to achieve higher-level objectives. We have implemented Bundler in Linux and evaluated it with real-world and emulation experiments. We find that Bundler allows the sender-chosen policy to be effective: when configured to implement Stochastic Fairness Queueing (SFQ), it improves median flow completion time (FCT) by between 28% and 97% across various scenarios.

cs.NI↗

arXiv2020

Elasticity Detection: A Building Block for Internet Congestion Control

This paper introduces Nimbus, a robust technique to detect whether the cross traffic competing with a flow is "elastic", and shows that this elasticity detector improves congestion control. If cross traffic is inelastic, then a sender can control queueing delays while achieving high throughput, but in the presence of elastic traffic, it may lose throughput if it attempts to control packet delay. To estimate elasticity, Nimbus modulates the flow's sending rate with sinusoidal pulses that create small traffic fluctuations at the bottleneck link, and measures the frequency response of the rate of the cross traffic. Our results on emulated and real-world paths show that congestion control using elasticity detection achieves throughput comparable to Cubic, but with delays that are 50-70 ms lower when cross traffic is inelastic. Nimbus detects the nature of the cross traffic more accurately than Copa, and is usable as a building block by other end-to-end algorithms.

cs.NI↗

arXiv2019

Backpressure Flow Control

Effective congestion control for data center networks is becoming increasingly challenging with a growing amount of latency sensitive traffic, much fatter links, and extremely bursty traffic. Widely deployed algorithms, such as DCTCP and DCQCN, are still far from optimal in many plausible scenarios, particularly for tail latency. Many operators compensate by running their networks at low average utilization, dramatically increasing costs. In this paper, we argue that we have reached the practical limits of end-to-end congestion control. Instead, we propose, implement, and evaluate a new congestion control architecture called Backpressure Flow Control (BFC). BFC provides per-hop per-flow flow control, but with bounded state, constant-time switch operations, and careful use of buffers. We demonstrate BFC's feasibility by implementing it on Tofino2, a state-of-the-art P4-based programmable hardware switch. In simulation, we show that BFC achieves near optimal throughput and tail latency behavior even under challenging conditions such as high network load and incast cross traffic. Compared to existing end-to-end schemes, BFC achieves 2.3 - 60 X lower tail latency for short flows and 1.6 - 5 X better average completion time for long flows.

cs.NI↗

arXiv2019

ABC: A Simple Explicit Congestion Controller for Wireless Networks

We propose Accel-Brake Control (ABC), a simple and deployable explicit congestion control protocol for network paths with time-varying wireless links. ABC routers mark each packet with an "accelerate" or "brake", which causes senders to slightly increase or decrease their congestion windows. Routers use this feedback to quickly guide senders towards a desired target rate. ABC requires no changes to header formats or user devices, but achieves better performance than XCP. ABC is also incrementally deployable; it operates correctly when the bottleneck is a non-ABC router, and can coexist with non-ABC traffic sharing the same bottleneck link. We evaluate ABC using a Wi-Fi implementation and trace-driven emulation of cellular links. ABC achieves 30-40% higher throughput than Cubic+Codel for similar delays, and 2.2X lower delays than BBR on a Wi-Fi path. On cellular network paths, ABC achieves 50% higher throughput than Cubic+Codel.

cs.NI↗

arXiv2018

Secure Incentivization for Decentralized Content Delivery

Prior research has proposed technical solutions to use peer-to-peer (P2P) content delivery to serve Internet video, showing that it can reduce costs to content providers. Yet, such methods have not become widespread except for a few niche instances. An important challenge is incentivization: what tangible benefits does P2P content delivery offer users who bring resources to the table? In this paper, we ask whether monetary incentives can help attract peers in P2P content delivery systems. We commissioned a professional survey of people around theUnited States to answer several relevant questions. We found that 51% of the 876 respondents--substantially larger than our expectations--answered "yes" to whether they would participate for suitable financial incentives. Encouraged by the results of the survey, we propose Gringotts, a system to structure incentives and securely incorporate P2P delivery into content delivery systems. Gringotts provides a novel Proof of Delivery mechanism that allows content providers to verify correct delivery of their files, and shows how to use cryptocurrency to pay peers while guarding against liars and Sybil attacks.

cs.NI↗

arXiv2014

Faster Algorithms for Algebraic Path Properties in RSMs with Constant Treewidth

Interprocedural analysis is at the heart of numerous applications in programming languages, such as alias analysis, constant propagation, etc. Recursive state machines (RSMs) are standard models for interprocedural analysis. We consider a general framework with RSMs where the transitions are labeled from a semiring, and path properties are algebraic with semiring operations. RSMs with algebraic path properties can model interprocedural dataflow analysis problems, the shortest path problem, the most probable path problem, etc. The traditional algorithms for interprocedural analysis focus on path properties where the starting point is \emph{fixed} as the entry point of a specific method. In this work, we consider possible multiple queries as required in many applications such as in alias analysis. The study of multiple queries allows us to bring in a very important algorithmic distinction between the resource usage of the \emph{one-time} preprocessing vs for \emph{each individual} query. The second aspect that we consider is that the control flow graphs for most programs have constant treewidth. Our main contributions are simple and implementable algorithms that support multiple queries for algebraic path properties for RSMs that have constant treewidth. Our theoretical results show that our algorithms have small additional one-time preprocessing, but can answer subsequent queries significantly faster as compared to the current best-known solutions for several important problems, such as interprocedural reachability and shortest path. We provide a prototype implementation for interprocedural reachability and intraprocedural shortest path that gives a significant speed-up on several benchmarks.

cs.PL↗