arXiv subjects

H. Jonathan Chao

Publications and source records attributed to H. Jonathan Chao.

16 recordsLinked to original sources

arXiv2026

Dynamic Hierarchical Birkhoff-von Neumann Decomposition for All-to-All GPU Communication

All-to-all GPU communication is a critical bottleneck in large-scale training clusters, where completion time is constrained by per-port bandwidth and can be severely impacted by traffic skew across GPUs and network interface cards (NICs). This issue is amplified by the two-tier structure of modern GPU systems, which combine fast intra-server links with much slower inter-server networks. Motivated by recent system observations that highlight the importance of traffic reshaping and hierarchy awareness, we study all-to-all scheduling from an online switching and queueing-theoretic perspective. We propose a dynamic hierarchical Birkhoff--von Neumann (BvN) decomposition framework tailored to two-tier GPU fabrics. At each frame boundary, traffic is first balanced within each server using simple local operations to mitigate micro-level GPU/NIC skew while preserving aggregate server-to-server demand. A hierarchical BvN decomposition is then applied at the server level and refined into GPU-level matchings, significantly reducing decomposition complexity relative to a flat GPU-level approach. By integrating this construction with the dynamic frame sizing (DFS) principle, we obtain an online scheduler with provable stability under admissible Poisson arrivals. Simulations demonstrate substantial reductions in mean frame length, particularly under server-localized hotspot traffic.

cs.NI↗

arXiv2020

SharpEdge: An Asynchronous and Core-Agnostic Solution to Guarantee Bounded-Delays

What are the key properties that a network should have to provide bounded-delay guarantees for the packets? In this paper, we attempt to answer this question. To that end, we explore the theory of bounded-delay networks and provide the necessary and the sufficient conditions required to have deterministic bounded-delays in the network. We prove that as long as a network is work-conserving, independent of the packet scheduling and queue management algorithms used in the switches, it is sufficient to shape the traffic~\textit{properly} at the edge of the network to meet hard bounded-delays in the network. Using the derived theorems, we present SharpEdge, a novel design to meet deterministic bounded-delays in the network. To the best of our knowledge, SharpEdge is the first scheme that can meet all following key properties: 1) it supports coexistence of different classes of traffic, while it can guarantee their different required bounded-delays 2) it does not require any changes in the core of the network, 3) it supports both periodic and bursty traffic patterns, and 4) it does not require any time synchronization between network devices.

cs.NI↗

arXiv2020

Wanna Make Your TCP Scheme Great for Cellular Networks? Let Machines Do It for You!

Can we instead of designing just another new TCP, design a TCP \textit{plug-in} which can boost the performance of the existing/future TCP designs in cellular networks? To answer this question, we introduce DeepCC plug-in. DeepCC leverages deep reinforcement learning (DRL), a modern decision-making tool, to steer TCP toward achieving applications' desired delay and high throughput in a highly dynamic network such as the cellular network. The fact that DeepCC does not try to reinvent/replace TCP but aims to boost the performance of it differentiates it from the most (if not all) of the existing reinforcement learning (RL) systems where RL systems are considered clean-slate alternative designs replacing the traditional ones. We used DeepCC plug-in to boost the performance of various old and new TCP schemes including TCP Cubic, Google's BBR, TCP Westwood, and TCP Illinois in cellular networks. Through both extensive trace-based evaluations and in-field tests, we show that not only DeepCC can significantly improve the performance of TCP, but also after accompanied by DeepCC, these schemes can outperform state-of-the-art TCP protocols including Aurora, Sprout, Verus, C2TCP, Copa, Indigo, Remy, PCC-Vivace, and LEDBAT in cellular networks.

cs.NI↗

arXiv2020

CFR-RL: Traffic Engineering with Reinforcement Learning in SDN

Traditional Traffic Engineering (TE) solutions can achieve the optimal or near-optimal performance by rerouting as many flows as possible. However, they do not usually consider the negative impact, such as packet out of order, when frequently rerouting flows in the network. To mitigate the impact of network disturbance, one promising TE solution is forwarding the majority of traffic flows using Equal-Cost Multi-Path (ECMP) and selectively rerouting a few critical flows using Software-Defined Networking (SDN) to balance link utilization of the network. However, critical flow rerouting is not trivial because the solution space for critical flow selection is enormous. Moreover, it is impossible to design a heuristic algorithm for this problem based on fixed and simple rules, since rule-based heuristics are unable to adapt to the changes of the traffic matrix and network dynamics. In this paper, we propose CFR-RL (Critical Flow Rerouting-Reinforcement Learning), a Reinforcement Learning-based scheme that learns a policy to select critical flows for each given traffic matrix automatically. CFR-RL then reroutes these selected critical flows to balance link utilization of the network by formulating and solving a simple Linear Programming (LP) problem. Extensive evaluations show that CFR-RL achieves near-optimal performance by rerouting only 10%-21.3% of total traffic.

cs.NI↗

arXiv2020

To schedule or not to schedule: when no-scheduling can beat the best-known flow scheduling algorithm in datacenter networks

Conventional wisdom for minimizing the average flow completion time (AFCT) in the datacenter network (DCN), where flow sizes are highly variable, would suggest scheduling every individual flow. However, we show that considering scheduling delay (including scheduler's computational and communication delays), serving most of the flows without any scheduling and only in first-come-first-served (FCFS) manner significantly improves their performance even when it is compared to the shortest remaining processing time (SRPT)-known as optimum algorithm when scheduling delay is zero. To do so, we only require to have two coarse classes of flows categorized based on flows' sizes (1st-class including flows smaller than a threshold, H, and 2nd-class including others) and serve 1st-class flows always before serving 2nd-class ones. To show that, we take SRPT scheduling algorithm accompanied by the global knowledge of flows, formulate impact of scheduling delay on its performance, and prove that for any flow size distribution and network load (<1), there is always a threshold, H, which guarantees 1st-class flows achieve lower AFCT under FCFS compared to SRPT. Our numerically calculated results and extensive flow-level simulations show that on average, more than 90% of flows could be in 1st-class and consequently do not require any scheduling.

cs.NI↗

arXiv2019

Bounding Queue Delay in Cellular Networks to Support Ultra-Low Latency Applications

Most of the current active queue management (AQM) designs have major issues including severe hardship of being tuned for highly fluctuated cellular access link bandwidths. Consequently, most of the cellular network providers either give up using AQMs or use conservative offline configurations for them. However, these choices will significantly impact the performance of the emerging interactive and highly delay sensitive applications such as virtual reality and vehicle-to-vehicle communications. Therefore, in this paper, we investigate the problems of existing AQM schemes and show that they are not suitable options to support ultra-low latency applications in a highly dynamic network such as current and future cellular networks. Moreover, we believe that achieving good performance does not necessarily come from complex drop rate calculation algorithms or complicated AQM techniques. Consequently, we propose BoDe an extremely simple and deployment friendly AQM scheme to bound the queuing delay of served packets and support ultra-low latency applications. We have evaluated BoDe in extensive trace-based evaluations using cellular traces from 3 different service providers in the US and compared its performance with state-of-the-art AQM designs including CoDel and PIE under a variety of streaming applications, video conferencing applications, and various recently proposed TCP protocols. Results show that despite BoDe's simple design, it outperforms other schemes and achieves significantly lower queuing delay in all tested scenarios.

cs.NI↗

arXiv2019

Intelligent Anomaly Detection and Mitigation in Data Centers

Data centers play a key role in today's Internet. Cloud applications are mainly hosted on multi-tenant warehouse-scale data centers. Anomalies pose a serious threat to data centers' operations. If not controlled properly, a simple anomaly can spread throughout the data center, resulting in a cascading failure. Amazon AWS had been affected by such incidents recently. Although some solutions are proposed to detect anomalies and prevent cascading failures, they mainly rely on application-specific metrics and case-based diagnosis to detect the anomalies. Given the variety of applications on a multi-tenant data center, proposed solutions are not capable of detecting anomalies in a timely manner. In this paper we design an application-agnostic anomaly detection scheme. More specifically, our design uses a highly distributed data mining scheme over network-level traffic metrics to detect anomalies. Once anomalies are detected, simple actions are taken to mitigate the damage. This ensures that errors are confined and prevents cascading failures before administrators intervene.

cs.NI↗

arXiv2019

In-network Congestion-aware Load Balancing at Transport Layer

Load balancing at transport layer is an important function in data centers, content delivery networks, and mobile networks, where per-connection consistency (PCC) has to be met for optimal performance. Cloud-native L4 load balancers are commonly deployed as virtual network functions (VNFs) and are a critical forwarding element in modern cloud infrastructure. We identify load imbalance among service instances as the main cause of additional processing delay caused by transport-layer load balancers. Existing transport-layer load balancers rely on one of two methods: host-level traffic redirection, which may add as much as 12.48% additional traffic to underlying networks, or connection tracking, which consumes a considerable amount of memory in load balancers. Both of these methods result in inefficient usage of network resources. We propose the in-network congestion-aware load Balancer (INCAB) to achieve even load distribution among service instances and optimal network resources usage in addition to meeting the PCC requirement. We show that INCAB is capable of identifying and monitoring each instance's most-utilized resource and can improve the load distribution among all service instances. INCAB utilizes a Bloom filter and an ultra-compact connection table for in-network flow distribution. Furthermore, it does not rely on end hosts for traffic redirection. Our flow level simulations show that INCAB improves flows' average completion time by 31.97% compared to stateless solutions.

cs.NI↗

arXiv2019

Enabling Mobility in LTE-Compatible Mobile-edge Computing with Programmable Switches

Network softwarization triggered a new wave of innovation in modern network design. The next generation of mobile networks embraces this trend. Mobile-edge computing (MEC) is a key part of emerging mobile networks that enables ultra-low latency mission-critical application such as vehicle-to vehicle communication. MEC aims at bringing delay-sensitive applications closer to the radio access network to enable ultra-low latency for users and decrease the back-haul pressure on mobile service providers. However, there are no practical solutions to enable mobility at MEC where connections are no longer anchored to the core network and serving applications are supposed to move as their users move. We propose the mobile-edge gateway (MEGW) to address this gap. MEGW enables mobility for MEC applications transparently and without requiring any modifications to existing protocols and applications. MEGW supports mobility by reconstructing mobile users' location via listening to LTE control plane in addition to using two-stage location-dependent traffic steering for edge connections. Networks can incrementally upgrade to support MEC by upgrading some IP router to programmable switches that run MEGW. We have implemented MEGW using P4 language and verified its compatibility with existing LTE networks in a testbed running reference LTE protocol stack. Furthermore, using packet-level simulations we show that the two-stage traffic steering algorithm reduces the number of application migrations and simplifies service provisioning.

cs.NI↗

arXiv2019

Spotlight: Scalable Transport Layer Load Balancing for Data Center Networks

Load Balancing plays a vital role in modern data centers to distribute traffic among instances of network functions or services. State-of-the-art load balancers such as Silkroad dispatch traffic obliviously without considering the real-time utilization of service instances and therefore can lead to uneven load distribution and suboptimal performance. In this paper, we design and implement Spotlight, a scalable and distributed load balancing architecture that maintains connection-to-instance mapping consistency at the edge of data center networks. Spotlight uses a new stateful flow dispatcher which periodically polls instances' load and dispatches incoming connections to instances in proportion to their available capacity. Our design utilizes distributed control plane and in-band flow dispatching and thus scales horizontally in data center networks. Through extensive flow-level simulation and packet-level experiments on a testbed, we demonstrate that compared to existing methods Spotlight distributes the traffic more efficiently and has near-optimum performance in terms of overall service utilization. Moreover, Spotlight is not sensitive to utilization polling interval and therefore can be implemented with low polling frequency to reduce the amount of control traffic. Indeed, Spotlight achieves the mentioned performance improvements using O(100ms) polling interval.

cs.NI↗

arXiv2018

C2TCP: A Flexible Cellular TCP to Meet Stringent Delay Requirements

Since current widely available network protocols/systems are mainly throughput-oriented designs, meeting stringent delay requirements of new applications such as virtual reality and vehicle-to-vehicle communications on cellular network requires new network protocol/system designs. C2TCP is an effort toward that new design direction. C2TCP is inspired by in-network active queue management (AQM) designs such as RED and CoDel and motivated by lack of a flexible end-to-end (e2e) approach which can adapt itself to different applications' QoS requirements without modifying any network devices. It copes with unique challenges in cellular networks for achieving ultra-low latency (including highly variable channels, deep per-user buffers, self-inflicted queuing delays, radio uplink/downlink scheduling delays) and intends to satisfy stringent delay requirements of different applications while maximizing the throughput. C2TCP works on top of classic throughput-oriented TCP and accommodates various target delays without requiring any channel prediction, network state profiling, or complicated rate adjustment mechanisms. We have evaluated C2TCP in both real-world environment and extensive trace-based emulations and compared its performance with different TCP variants and state-of-the-art schemes including PCC-Vivace, Google's BBR, Verus, Sprout, TCP Westwood, and Cubic. Results show that C2TCP outperforms all these schemes and achieves lower average delay, jitter, and 95th percentile delay for packets.

cs.NI↗

arXiv2018

HyLine: a Simple and Practical Flow Scheduling for Commodity Datacenters

Today's datacenter networks (DCNs) have been built upon multipath topologies where each path contains multiple links. However, flow scheduling schemes proposed to minimize flow completion times (FCT) in DCNs are based on algorithms which are optimum or close-to-optimum only over single link. Moreover, most of these scheduling schemes seek either fully centralized approaches having overhead of communicating to a central entity or fully distributed approaches requiring changes in the fabric. Motivated by these shortcomings, we present HyLine a simple scheduling design for commodity DCNs which is equipped with a joint load-balancing and flow scheduling (path-aware) design exploiting the multipath nature of DCNs. HyLine takes a hybrid approach and uses the global-awareness of centralized and agility of distributed techniques without requiring any changes in the fabric. To that end, it determines a threshold margin identifying flows for which using centralized approach is beneficial. We have shown through extensive ns2 simulations that despite HyLine's simplicity, it significantly outperforms existing schemes and achieves lower average and 99th percentile FCTs. For instance, compared to Qjump-state-of-the-art practical scheme-and pFabric-one of the best performing flow scheduling schemes-HyLine reduces average FCT up to 68% and 31%, respectively, under a production datacenter workload.

cs.NI↗

arXiv2018

Cellular Controlled Delay TCP (C2TCP)

Cellular networks have special characteristics including highly variable channels, fast fluctuating capacities, deep per user buffers, self-inflicted queuing delays, radio uplink/downlink scheduling delays, etc. These distinguishing properties make the problem of achieving low latency and high throughput in cellular networks more challenging than in wired networks. That's why in this environment, TCP and its flavors, which are generally designed for wired networks, perform poorly. To cope with these challenges, we present C2TCP, a flexible end-to-end solution targeting interactive applications requiring high throughput and low delay in cellular networks. C2TCP stands on top of loss-based TCP and brings it delay sensitivity without requiring any network state profiling, channel prediction, or complicated rate adjustment mechanisms. The key idea behind C2TCP is to absorb dynamics of unpredictable cellular channels by investigating local minimum delay of packets in a moving time window and react to the cellular network's capacity changes very fast. Through extensive trace-based evaluations using traces from five commercial LTE and 3G networks, we have compared performance of C2TCP with various TCP variants, and state-of-the-art schemes including BBR, Verus, and Sprout. Results show that on average, C2TCP outperforms these schemes and achieves lower average and 95th percentile delay for packets.

cs.NI↗

arXiv2017

Design of a Hybrid Modular Switch

Network Function Virtualization (NFV) shed new light for the design, deployment, and management of cloud networks. Many network functions such as firewalls, load balancers, and intrusion detection systems can be virtualized by servers. However, network operators often have to sacrifice programmability in order to achieve high throughput, especially at networks' edge where complex network functions are required. Here, we design, implement, and evaluate Hybrid Modular Switch (HyMoS). The hybrid hardware/software switch is designed to meet requirements for modern-day NFV applications in providing high-throughput, with a high degree of programmability. HyMoS utilizes P4-compatible Network Interface Cards (NICs), PCI Express interface and CPU to act as line cards, switch fabric, and fabric controller respectively. In our implementation of HyMos, PCI Express interface is turned into a non-blocking switch fabric with a throughput of hundreds of Gigabits per second. Compared to existing NFV infrastructure, HyMoS offers modularity in hardware and software as well as a higher degree of programmability by supporting a superset of P4 language.

cs.NI↗

arXiv2011

Trimming the Multipath for Efficient Dynamic Routing

Multipath routing is a trivial way to exploit the path diversity to leverage the network throughput. Technologies such as OSPF ECMP use all the available paths in the network to forward traffic, however, we argue that is not necessary to do so to load balance the network. In this paper, we consider multipath routing with only a limited number of end-to-end paths for each source and destination, and found that this can still load balance the traffic. We devised an algorithm to select a few paths for each source-destination pair so that when all traffic are forwarded over these paths, we can achieve a balanced load in the sense that the maximum link utilization is comparable to that of ECMP forwarding. When the constraint of only shortest paths (i.e. equal paths) are relaxed, we can even outperform ECMP in certain cases. As a result, we can use a few end-to-end tunnels between each source and destination nodes to achieve the load balancing of traffic.

cs.NI↗

arXiv2011

Use of Devolved Controllers in Data Center Networks

In a data center network, for example, it is quite often to use controllers to manage resources in a centralized man- ner. Centralized control, however, imposes a scalability problem. In this paper, we investigate the use of multiple independent controllers instead of a single omniscient controller to manage resources. Each controller looks after a portion of the network only, but they together cover the whole network. This therefore solves the scalability problem. We use flow allocation as an example to see how this approach can manage the bandwidth use in a distributed manner. The focus is on how to assign components of a network to the controllers so that (1) each controller only need to look after a small part of the network but (2) there is at least one controller that can answer any request. We outline a way to configure the controllers to fulfill these requirements as a proof that the use of devolved controllers is possible. We also discuss several issues related to such implementation.

cs.NI↗