An official website of the United States government
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
AI on the Edge: Characterizing AI-based IoT Applications Using Specialized Edge Architectures
Edge computing has emerged as a popular paradigm for supporting mobile and IoT applications with low latency or high bandwidth needs. The attractiveness of edge computing has been further enhanced due to the recent availability of special-purpose hardware to accelerate specific compute tasks, such as deep learning inference, on edge nodes. In this paper, we experimentally compare the benefits and limitations of using specialized edge systems, built using edge accelerators, to more traditional forms of edge and cloud computing. Our experimental study using edge-based AI workloads shows that today's edge accelerators can provide comparable, and in many cases better, performance, when normalized for power or cost, than traditional edge and cloud servers. They also provide latency and bandwidth benefits for split processing, across and within tiers, when using model compression or model splitting, but require dynamic methods to determine the optimal split across tiers. We find that edge accelerators can support varying degrees of concurrency for multi-tenant inference applications, but lack isolation mechanisms necessary for edge cloud multi-tenant hosting.
more »
« less
- Award ID(s):
- 1908536
- PAR ID:
- 10295597
- Date Published:
- Journal Name:
- IEEE International Symposium on Workload Characterization (IISWC)
- Page Range / eLocation ID:
- 145 to 156
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
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
-
The rapid growth in technology and wide use of internet has increased smart applications such as intelligent transportation control system, and Internet of Things, which heavily rely on an efficient and reliable connectivity network. To overcome high bandwidth work load on the network, as well as minimize latency for real-time applications, the computation can be moved from the central cloud to a distributed edge cloud. The edge computing benefits various smart applications that uses distributed network for data analytics and services. Different from the existing cloud management solutions, edge computing needs to move cloud management services towards distributed heterogeneous edge nodes for multi-tenant user applications. However, existing cloud management services do not offer remote deployment of multi-tenant user applications on the cloud of edge nodes. In this paper, we propose a practical edge cloud software framework for deploying multi-tenant distributed smart applications. Having multiple distributed end nodes, auto discovery of all active end nodes is required for deploying multi-tenant user applications. However, existing cloud solutions require either private network or fixed IP address, which is not achievable for the distributed edge nodes. Most of the edge nodes connected through the public internet without fixed IP, and some of them even connect through IEEE 802.15 based sensor networks. We propose to build a software platform to manage the distributed edge nodes as well as support services to deploy and launch isolated, multi-tenant user applications through a lightweight container. We propose an architectural solution to remotely access edge cloud management services through intermittent internet connections. We open sourced our whole set of software solutions, and analyzed the major performance metrics of the edge cloud platform.more » « less
-
The convergence of edge computing and artificial intelligence requires that inference is performed on-device to provide rapid response with low latency and high accuracy without transferring large amounts of data to the cloud. However, power and size limitations make it challenging for electrical accelerators to support both inference and training for large neural network models. To this end, we propose Trident, a low-power photonic accelerator that combines the benefits of phase change material (PCM) and photonics to implement both inference and training in one unified architecture. Emerging silicon photonics has the potential to exploit the parallelism of neural network models, reduce power consumption and provide high bandwidth density via wavelength division multiplexing, making photonics an ideal candidate for on-device training and inference. As PCM is reconfigurable and non-volatile, we utilize it for two distinct purposes: (i) to maintain resonant wavelength without expensive electrical or thermal heaters, and (ii) to implement non-linear activation function, which eliminates the need to move data between memory and compute units. This multi-purpose use of PCM is shown to lead to significant reduction in energy consumption and execution time. Compared to photonic accelerators DEAP-CNN, CrossLight, and PIXEL, Trident improves energy efficiency by up to 43% and latency by up to 150% on average. Compared to electronic edge AI accelerators Google Coral which utilizes the Google Edge TPU and Bearkey TB96-AI, Trident improves energy efficiency by 11% and 93% respectively. While NVIDIA AGX Xavier is more energy efficient, the reduced data movement and GST activation of Trident reduce latency by 107% on average compared to the NVIDIA accelerator. When compared to the Google Coral and the Bearkey TB96-AI, Trident reduces latency by 1413% and 595% on average.more » « less
-
null (Ed.)Edge cloud data centers (Edge) are deployed to provide responsive services to the end-users. Edge can host more powerful CPUs and DNN accelerators such as GPUs and may be used for offloading tasks from end-user devices that require more significant compute capabilities. But Edge resources may also be limited and must be shared across multiple applications that process requests concurrently from several clients. However, multiplexing GPUs across applications is challenging. With edge cloud servers needing to process a lot of streaming and the advent of multi-GPU systems, getting that data from the network to the GPU can be a bottleneck, limiting the amount of work the GPU cluster can do. The lack of prompt notification of job completion from the GPU can also result in poor GPU utilization. We build on our recent work on controlled spatial sharing of a single GPU to expand to support multi-GPU systems and propose a framework that addresses these challenges. Unlike the state-of-the-art uncontrolled spatial sharing currently available with systems such as CUDA-MPS, our controlled spatial sharing approach uses each of the GPU in the cluster efficiently by removing interference between applications, resulting in much better, predictable, inference latency We also use each of the cluster GPU's DMA engines to offload data transfers to the GPU complex, thereby preventing the CPU from being the bottleneck. Finally, our framework uses the CUDA event library to give timely, low overhead GPU notifications. Our evaluations show we can achieve low DNN inference latency and improve DNN inference throughput by at least a factor of 2.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.1109/IISWC50251.2020.00023
