arXiv subjects

Yiting Xia

Publications and source records attributed to Yiting Xia.

5 recordsLinked to original sources

arXiv2025

OpenOptics: An Open Research Framework for Optical Data Center Networks

Optical data center networks (DCNs) are emerging as a promising design for cloud infrastructure. However, existing optical DCN architectures operate as closed ecosystems, tying software solutions to specific optical hardware. We introduce OpenOptics, an open research framework that decouples software from hardware, allowing them to evolve independently. OpenOptics features: (1) a time-flow table abstraction as a common interface between optical hardware and software, (2) a unified workflow and user-friendly API for implementing various optical DCNs with simple Python scripts, and (3) a backend system that re-architects queue management to support the time-flow tables and provides rich infrastructure services for diverse applications. Built on programmable switches, OpenOptics achieves a record-breaking minimum optical circuit duration of 2 $μ$s using commodity devices. We validate OpenOptics' generality by implementing six optical architectures and seven routing schemes on an optical testbed and conducting benchmarks on a 108-ToR setup, showcasing its efficiency. Additionally, case studies highlight novel research opportunities enabled by OpenOptics.

cs.NI↗

arXiv2024

Unlocking Diversity of Fast-Switched Optical Data Center Networks with Unified Routing

Optical data center networks (DCNs) are emerging as a promising solution for cloud infrastructure in the post-Moore's Law era, particularly with the advent of 'fast-switched' optical architectures capable of circuit reconfiguration at microsecond or even nanosecond scales. However, frequent reconfiguration of optical circuits introduces a unique challenge: in-flight packets risk loss during these transitions, hindering the deployment of many mature optical hardware designs due to the lack of suitable routing solutions. In this paper, we present Unified Routing for Optical networks (URO), a general routing framework designed to support fast-switched optical DCNs across various hardware architectures. URO combines theoretical modeling of this novel routing problem with practical implementation on programmable switches, enabling precise, time-based packet transmission. Our prototype on Intel Tofino2 switches achieves a minimum circuit duration of 2us, ensuring end-to-end, loss-free application performance. Large-scale simulations using production DCN traffic validate URO's generality across different hardware configurations, demonstrating its effectiveness and efficient system resource utilization.

cs.NI↗

arXiv2024

Nanosecond Precision Time Synchronization for Optical Data Center Networks

Optical data center networks (DCNs) are renovating the infrastructure design for the cloud in the post Moore's law era. The fact that optical DCNs rely on optical circuits of microsecond-scale durations makes nanosecond-precision time synchronization essential for the correct functioning of routing on the network fabric. However, current studies on optical DCNs neglect the fundamental need for accurate time synchronization. In this paper, we bridge the gap by developing Nanosecond Optical Synchronization (NOS), the first nanosecond-precision synchronization solution for optical DCNs general to various optical hardware. NOS builds clock propagation trees on top of the dynamically reconfigured circuits in optical DCNs, allowing switches to seek better sync parents throughout time. It predicts drifts in the tree-building process, which enables minimization of sync errors. We also tailor today's sync protocols to the needs of optical DCNs, including reducing the number of sync messages to fit into short circuit durations and correcting timestamp errors for higher sync accuracy. Our implementation on programmable switches shows 28ns sync accuracy in a 192-ToR setting.

cs.NI↗

arXiv2024

Optimizing Mixture-of-Experts Inference Time Combining Model Deployment and Communication Scheduling

As machine learning models scale in size and complexity, their computational requirements become a significant barrier. Mixture-of-Experts (MoE) models alleviate this issue by selectively activating relevant experts. Despite this, MoE models are hindered by high communication overhead from all-to-all operations, low GPU utilization due to the synchronous communication constraint, and complications from heterogeneous GPU environments. This paper presents Aurora, which optimizes both model deployment and all-to-all communication scheduling to address these challenges in MoE inference. Aurora achieves minimal communication times by strategically ordering token transmissions in all-to-all communications. It improves GPU utilization by colocating experts from different models on the same device, avoiding the limitations of synchronous all-to-all communication. We analyze Aurora's optimization strategies theoretically across four common GPU cluster settings: exclusive vs. colocated models on GPUs, and homogeneous vs. heterogeneous GPUs. Aurora provides optimal solutions for three cases, and for the remaining NP-hard scenario, it offers a polynomial-time sub-optimal solution with only a 1.07x degradation from the optimal. Aurora is the first approach to minimize MoE inference time via optimal model deployment and communication scheduling across various scenarios. Evaluations demonstrate that Aurora significantly accelerates inference, achieving speedups of up to 2.38x in homogeneous clusters and 3.54x in heterogeneous environments. Moreover, Aurora enhances GPU utilization by up to 1.5x compared to existing methods.

cs.LG↗

arXiv2020

Music Classification in MIDI Format based on LSTM Mdel

Music classification between music made by AI or human composers can be done by deep learning networks. We first transformed music samples in midi format to natural language sequences, then classified these samples by mLSTM (multiplicative Long Short Term Memory) + logistic regression. The accuracy of the result evaluated by 10-fold cross validation can reach 90%. Our work indicates that music generated by AI and human composers do have different characteristics, which can be learned by deep learning networks.

cs.SD↗