arXiv subjects

JooYoung Park

Publications and source records attributed to JooYoung Park.

5 recordsLinked to original sources

arXiv2026

CodecSight: Leveraging Video Codec Signals for Efficient Streaming VLM Inference

Continuous inference over concurrent video streams imposes substantial compute and memory demands on vision-language model (VLM) serving. Streaming inference uses sliding windows to maintain a bounded context of recent video, but processing each window independently repeats visual encoding and large language model (LLM) prefilling for similar and overlapping content. Existing optimizations provide limited coordination across these stages and often rely on model-specific training, profiling, or model-generated signals. We present CodecSight, a streaming VLM serving system that uses codec metadata as shared runtime guidance across visual encoding and LLM prefilling, without model-specific training or offline profiling. Codec-derived change signals guide patch pruning before visual encoding, reducing both visual computation and the number of downstream visual tokens. Codec-defined frame types guide selective key-value (KV) refresh across windows, while positional correction enables reuse of the remaining cached keys. Across three VLMs and four video workloads, our vLLM-based implementation supports up to $3.3\times$ as many concurrent streams and achieves up to a $5.3\times$ speedup in average time-to-first-token relative to the state-of-the-art baselines. It also reduces executed FLOPs by up to 93%, with a maximum task-quality decrease of 4.64 percentage points.

cs.DC↗

arXiv2026

Rethinking AI Cloud Infrastructure for Agentic Serving Systems with the Aries Experimentation Framework

Autonomous agents challenge conventional LLM serving by coupling repeated inference with persistent context and sandboxed tool execution. We present Aries, a full-stack experimentation framework that separates task semantics from execution configurations, reconstructs cross-component agent trajectories with correlated system telemetry, and exposes stateful tool execution through a consistent interface across heterogeneous sandbox substrates. We use Aries to conduct reproducible experiments on open agent harnesses and benchmarks. We complement these experiments with production traces from a commercial platform, grounding low-level systems research in observed production behavior. Our results show that (1) token-centric metrics miss non-inference bottlenecks, (2) retaining additional context yields diminishing accuracy benefits while reducing serving capacity, and (3) tool sandboxes alternate between long idle periods and short resource bursts, while current snapshot-based state management makes aggressive suspension costly. A complementary security analysis further highlights the need to reduce the sandbox attack surface. We then discuss the vision for agent-native serving systems designed around trajectory-level metrics, adaptive context management, elastic sandbox resource management, and sandboxes with minimized attack surface.

cs.DC↗

arXiv2026

Nexus: Transparent I/O Offloading for High-Density Serverless Computing

Serverless computing relies on extreme multi-tenancy to remain economically viable, driving providers to rely on virtual machines (VMs) that ensure strong isolation and seamless ecosystem compatibility with the FaaS programming model. However, current architectures tightly couple application processing logic with I/O processing, forcing every VM to duplicate a heavy communication fabric (cloud SDK, RPC, and TCP/IP). Our analysis reveals this duplication consumes over 25% of a function's memory footprint, and may double the CPU cycles in VMs compared to bare-metal execution. While prior systems attempt to solve this using WebAssembly or library OSes, they naively sacrifice ecosystem compatibility, forcing developers to migrate code and dependencies to new languages. We introduce Nexus, a serverless-native KVM-based hypervisor that transparently decouples compute from I/O. Nexus shifts the execution model by intercepting communication fabric at the API boundary and offloading it to an always-on host shared backend via zero-copy shared memory. This removes the heavyweight communication fabric from the guest VM, while preserving the conventional serverless programming model. By structurally separating these domains, Nexus unlocks asynchronous I/O optimizations: overlapping input payload prefetching with VM restoration from a snapshot and writing output payloads back to storage off the critical path. Compared to the production baseline, Nexus reduces overall node-level CPU and memory consumption by up to 44% and 31%, respectively, thus increasing deployment density by 37%. Also, Nexus reduces warm- and cold-start latency by 39% and 10%, respectively, bringing the response time within 20% of that of a WASM-based, ecosystem-incompatible hypervisor.

cs.DC↗

arXiv2024

Enabling Large Batch Size Training for DNN Models Beyond the Memory Limit While Maintaining Performance

Recent deep learning models are difficult to train using a large batch size, because commodity machines may not have enough memory to accommodate both the model and a large data batch size. The batch size is one of the hyper-parameters used in the training model, and it is dependent on and is limited by the target machine memory capacity because the batch size can only fit into the remaining memory after the model is uploaded. Moreover, the data item size is also an important factor because if each data item size is larger then the batch size that can fit into the remaining memory becomes smaller. This paper proposes a method called Micro-Batch Processing (MBP) to address this problem. This method helps deep learning models to train by providing a batch processing method that splits a batch into a size that can fit in the remaining memory and processes them sequentially. After processing the small batches individually, a loss normalization algorithm based on the gradient accumulation is used to maintain the performance. The purpose of our method is to allow deep learning models to train using larger batch sizes that exceed the memory capacity of a system without increasing the memory size or using multiple devices (GPUs).

cs.LG↗

arXiv2022

Dataloader Parameter Tuner: An Automated Dataloader Parameter Tuner for Deep Learning Models

Deep learning has recently become one of the most compute/data-intensive methods and is widely used in many research areas and businesses. One of the critical challenges of deep learning is that it has many parameters that can be adjusted, and the optimal value may need to be determined for faster operation and high accuracy. The focus of this paper is the adjustable parameters of the dataloader. The dataloader in a system mainly groups the data appropriately and loads it to the main memory for the deep learning model to use. We introduce an automated framework called Dataloader Parameter Tuner (DPT) that determines the optimal value for the parameters required for the dataloader. This framework discovers the optimal values for the number of dataloader's subprocesses (i.e., worker) and prefetch factor through grid search to accelerate the data transfer for machine learning systems.

cs.DC↗