Zixi Chen

Yifu Ding, Zixi Chen, Thomas Magnanti

Energy demands from data centers have surged and stressed the grid in recent years. Electric grids require balancing supply and demand every second, motivating demand response (reduction) from large loads, including data centers. This can be achieved by rescheduling jobs on physical machines. Its real-time implementation is uncertain due to fluctuating resource utilization, and rescheduling incurs quality-of-service (QoS) losses that providers are unwilling to disclose. We propose a restless multi-arm bandit (RMAB) framework in which the grid operator requests load reductions without access to detailed job-rescheduling procedures. Using the open-source virtual machine (VM) datasets, we model job arrivals and rescheduling at each data center as a restless arm in a Markov decision process (MDP), and derive Whittle-index-based policies based on the learned transition function via Thompson sampling. To overcome the weakness of an increasingly long learning process due to an enlarged state space, we used a mixed strategy that included a global upper confidence bound (UCB) encoded with trust indices to enhance robustness and accelerate learning. Results show that the proposed mixed-strategy algorithm remains robust across varying state-space sizes and consistently outperforms the pure Thompson-Whittle (TW) algorithm, especially when contextual information is noisy. It also demonstrates superior performance compared to the state-of-the-art EXP4 framework. We provided an open-sourced code for reproducibility.

Yifu Ding, Zixi Chen, Ruicheng Ao et al.

The rapid rise in energy consumption from large-scale AI workloads in data centers placed the increasing pressures on power grids in recent years. Since grids must maintain real-time bal- ance between supply and demand, there has been heightened interest in job flexibility services from data centers. We propose an adaptive, learning-based contextual restless multi-armed bandit (RMAB) framework in which the grid requests load reductions without requiring visibility into job (re)scheduling decisions. With multiple virtual machine (VM) datasets, the core strategy models a cyclic job queue and batch-level (re)scheduling within each data center as a Markov decision process (MDP), and derives Whittle-index-based policies computed from the learned transition and reward functions via Thompson sampling. To address the challenges of enlarging state space and sparse state visits, we introduce domain-knowledge-enriched refined strategies, including adaptive mixed strategies, gated priors, low-rank smoothing, and offline posterior support. Extensive simulations under both baseline and stress scenarios demonstrate that the cellwise best refined variant exceeds TW in every reported cell of the stress sweep at lower computational cost than EXP4. These results demonstrate the economic value of data center flexibility services and underscore the importance of high-quality, open-source AI workload datasets for developing and evaluating adaptive scheduling algorithms.

Chendong Song, Meixuan Wang, Hang Zhou et al.

Attention-FFN disaggregation (AFD) is an emerging architecture for LLM decoding that separates state-heavy, KV-cache-dominated Attention computation from stateless, compute-intensive FFN computation, connected by per-step communication. While AFD enables independent scaling of memory and compute resources, its performance is highly sensitive to the Attention/FFN provisioning ratio: mis-sizing induces step-level blocking and costly device idle time. We develop a tractable analytical framework for sizing AFD bundles in an $r$A-$1$F topology, where the key difficulty is that Attention-side work is nonstationary-token context grows and requests are continuously replenished with random lengths-while FFN work is stable given the aggregated batch. Using a probabilistic workload model, we derive closed-form rules for the optimal A/F ratio that maximize average throughput per instance across the system. A trace-calibrated AFD simulator validates the theory: across workloads, the theoretical optimal A/F ratio matches the simulation-optimal within 10%, and consistently reduces idle time.

Zixi Chen, Yinyu Ye, Zijie Zhou

We study the online routing problem in large language model serving, where requests arrive sequentially and must be dispatched to parallel decode workers under tight batch-size and KV-cache constraints. Unlike widely used routing heuristics that are not tied to explicit service-level objectives (SLOs) and offer limited control over latency-throughput trade-offs, we introduce a multi-objective optimization framework that formulates routing as an online linear programming with interpretable decision rewards. We apply an efficient bid-price control policy based on the online linear programming that admits requests when their SLO-weighted benefit exceeds their shadow prices. To meet millisecond decision requirements, we develop a warm-started, projected first-order updates that track the evolving dual shadow prices online with predictable runtime. We integrate our router into the Vidur simulator and demonstrate substantial improvements over standard baselines across multiple SLO regimes, including end-to-end latency, time-to-first-token, throughput, and tail performance. A big picture from our result: a science-based approach outperforms others based on heuristics.

Tianci Bu, Yuan Lyu, Zixi Chen et al.

Data-parallel (DP) load balancing has emerged as a first-order bottleneck in large-scale LLM serving. When a model is sharded across devices via tensor parallelism (TP) or expert parallelism (EP) and replicated across many DP workers, every decode step ends in a synchronization barrier whose latency is set by the most heavily loaded worker; even modest persistent imbalance across DP workers compounds, step after step, into a substantial fraction of wasted compute. The problem is hard for reasons specific to LLM decoding: assignments are sticky (migrating KV caches has a high cost), per-request loads grow over time, arrivals are non-stationary, and the router must decide within a sub-100\,ms decode budget over hundreds of waiting requests and tens of workers. We present \textbf{BalanceRoute}, a family of practical online routing algorithms that target this bottleneck. The first, \textbf{BR-0}, requires no prediction infrastructure and uses a piecewise-linear F-score that captures the sharp asymmetry between admissions that fill safe margin and those that overflow into the envelope; a two-stage decomposition keeps per-step cost compatible with millisecond-scale scheduling. The second, \textbf{BR-H}, generalizes BR-0 with a short, constant lookahead $H$ and a lightweight termination-classifier interface, extending the F-score to a horizon-discounted form. We deploy BalanceRoute on a 144-NPU cluster and evaluate against vLLM baselines on both a proprietary production trace and the public Azure-2024 trace. Across both workloads, BalanceRoute substantially reduces average DP imbalance and improves end-to-end serving throughput.