An official website of the United States government
Existing machine learning inference-serving systems largely rely on hardware scaling by adding more devices or using more powerful accelerators to handle increasing query demands. However, hardware scaling might not be feasible for fixed-size edge clusters or private clouds due to their limited hardware resources. A viable alternate solution is accuracy scaling, which adapts the accuracy of ML models instead of hardware resources to handle varying query demands. This work studies the design of a high-throughput inferenceserving system with accuracy scaling that can meet throughput requirements while maximizing accuracy. To achieve the goal, this work proposes to identify the right amount of accuracy scaling by jointly optimizing three sub-problems: how to select model variants, how to place them on heterogeneous devices, and how to assign query workloads to each device. It also proposes a new adaptive batching algorithm to handle variations in query arrival times and minimize SLO violations. Based on the proposed techniques, we build an inference-serving system called Proteus and empirically evaluate it on real-world and synthetic traces. We show that Proteus reduces accuracy drop by up to 3× and latency timeouts by 2-10× with respect to baseline schemes, while meeting throughput requirements.
more »
« less
This content will become publicly available on July 7, 2026
Convergo: Multi-SLO-Aware Scheduling for Heterogeneous AI Accelerators on Edge Devices
With the growing prevalence of edge AI, systems are increasingly required to meet stringent and diverse service level objectives (SLOs), such as maintaining specific accuracy levels, ensuring sufficient inference throughput, and meeting deadlines, often simultaneously. However, concurrently achieving these varied and complex SLOs is particularly challenging due to the resource constraints of edge devices and the heterogeneity of AI accelerators. To address this gap, we present a novel AI scheduling framework, Convergo, which uniquely integrates heterogeneous accelerator management, multi-tenancy, and multi-SLO prioritization into one scheduling solution. Convergo not only leverages heterogeneous AI accelerators and supports AI multi-tenancy, but also integrates scheduling heuristics to meet multiple SLOs concurrently. Convergo enables the simultaneous satisfaction of multiple/complex SLO requirements (e.g., accuracy, throughput, and deadline constraints). The scheduling algorithm prioritizes inference requests, imposes critical constraints, and selects the best model combinations for current inferencing. We evaluated Convergo on the Jetson Xavier platform with portable TPU accelerators across various AI workloads, demonstrating its effectiveness. The evaluation results show that Convergo outper- forms state-of-the-art baselines, achieving over 90% satisfaction of all three distinct SLO requirements simultaneously while maintaining approximately 95% satisfaction for individual SLOs. Furthermore, Convergo achieves these results with negligible overhead, making it a promising solution for edge AI systems.
more »
« less
- Award ID(s):
- 2416214
- PAR ID:
- 10634817
- Publisher / Repository:
- IEEE
- Date Published:
- ISBN:
- 979-8-3315-5559-7
- Page Range / eLocation ID:
- 115 to 125
- Subject(s) / Keyword(s):
- Scheduling Edge Computing Edge AI AI Acceleration
- Format(s):
- Medium: X
- Location:
- Helsinki, Finland
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Existing machine learning inference-serving systems largely rely on hardware scaling by adding more devices or using more powerful accelerators to handle increasing query demands. However, hardware scaling might not be feasible for fixed-size edge clusters or private clouds due to their limited hardware resources. A viable alternate solution is accuracy scaling, which adapts the accuracy of ML models instead of hardware resources to handle varying query demands. This work studies the design of a high-throughput inferenceserving system with accuracy scaling that can meet throughput requirements while maximizing accuracy. To achieve the goal, this work proposes to identify the right amount of accuracy scaling by jointly optimizing three sub-problems: how to select model variants, how to place them on heterogeneous devices, and how to assign query workloads to each device. It also proposes a new adaptive batching algorithm to handle variations in query arrival times and minimize SLO violations. Based on the proposed techniques, we build an inference-serving system called Proteus and empirically evaluate it on real-world and synthetic traces. We show that Proteus reduces accuracy drop by up to 3× and latency timeouts by 2-10× with respect to baseline schemes, while meeting throughput requirements.more » « less
-
Description / Abstract: In order to effectively provide INaaS (Inference-as-a-Service) for AI applications in resource-limited cloud environments, two major challenges must be overcome: achieving low latency and providing multi-tenancy. This paper presents EIF (Efficient INaaS Framework), which uses a heterogeneous CPU-FPGA architecture to provide three methods to address these challenges (1) spatial multiplexing via software-hardware co-design virtualization techniques, (2) temporal multiplexing that exploits the sparsity of neural-net models, and (3) streaming-mode inference which overlaps data transfer and computation. The prototype EIF is implemented on an Intel PAC (shared-memory CPU-FPGA) platform. For evaluation, 12 types of DNN models were used as benchmarks, with different size and sparsity. Based on these experiments, we show that in EIF, the temporal multiplexing technique can improve the user density of an AI Accelerator Unit from 2$$\times$$ to 6$$\times$$, with marginal performance degradation. In the prototype system, the spatial multiplexing technique supports eight AI Accelerators Unit on one FPGA. By using a streaming mode based on a Mediated Pass-Through architecture, EIF can overcome the FPGA on-chip memory limitation to improve multi-tenancy and optimize the latency of INaaS. To further enhance INaaS, EIF utilizes the MapReduce function to provide a more flexible QoS. Together with the temporal/spatial multiplexing techniques, EIF can support 48 users simultaneously on a single FPGA board in our prototype system. In all tested benchmarks, cold-start latency accounts for only approximately 5\% of the total response time.more » « less
-
Model-serving systems expose machine learning (ML) models to applications programmatically via a high-level API. Cloud plat- forms use these systems to mask the complexities of optimally managing resources and servicing inference requests across multi- ple applications. Model serving at the edge is now also becoming increasingly important to support inference workloads with tight latency requirements. However, edge model serving differs substan- tially from cloud model serving in its latency, energy, and accuracy constraints: these systems must support multiple applications with widely different latency and accuracy requirements on embedded edge accelerators with limited computational and energy resources. To address the problem, this paper presents Dělen,1 a flexible and adaptive model-serving system for multi-tenant edge AI. Dělen exposes a high-level API that enables individual edge applications to specify a bound at runtime on the latency, accuracy, or energy of their inference requests. We efficiently implement Dělen using conditional execution in multi-exit deep neural networks (DNNs), which enables granular control over inference requests, and evalu- ate it on a resource-constrained Jetson Nano edge accelerator. We evaluate Dělen flexibility by implementing state-of-the-art adapta- tion policies using Dělen’s API, and evaluate its adaptability under different workload dynamics and goals when running single and multiple applications.more » « less
-
Since emerging edge applications such as Internet of Things (IoT) analytics and augmented reality have tight latency constraints, hardware AI accelerators have been recently proposed to speed up deep neural network (DNN) inference run by these applications. Resource-constrained edge servers and accelerators tend to be multiplexed across multiple IoT applications, introducing the potential for performance interference between latency-sensitive workloads. In this article, we design analytic models to capture the performance of DNN inference workloads on shared edge accelerators, such as GPU and edgeTPU, under different multiplexing and concurrency behaviors. After validating our models using extensive experiments, we use them to design various cluster resource management algorithms to intelligently manage multiple applications on edge accelerators while respecting their latency constraints. We implement a prototype of our system in Kubernetes and show that our system can host 2.3× more DNN applications in heterogeneous multi-tenant edge clusters with no latency violations when compared to traditional knapsack hosting algorithms.more » « less
