Grus: Enabling Latency SLOs for GPU-Accelerated NFV Systems
Grus: Enabling Latency SLOs for GPU-Accelerated NFV Systems
复制标题
DOI：
10.1109/icnp.2018.00025
复制
发表时间：
2018-09
期刊：
影响因子：
--
通讯作者：
Zhilong Zheng;J. Bi;Haiping Wang;Chen Sun;Heng Yu;Hongxin Hu;K. Gao;Jianping Wu
中科院分区：
文献类型：
--
作者：
Zhilong Zheng;J. Bi;Haiping Wang;Chen Sun;Heng Yu;Hongxin Hu;K. Gao;Jianping Wu
图形处理单元（GPU）最近已被开发作为硬件加速器，以提高网络功能虚拟化（NFV）的性能。然而，当多个网络功能（NF）共同位于同一机器中时，GPU加速的NFV系统遭受显著的延迟变化，这阻止了运营商支持延迟服务水平目标（SLO）。现有的研究努力，以解决这个问题，只能保证有限数量的SLO的资源利用效率非常低。在本文中，我们提出了Grus框架，以支持GPU加速的NFV系统中的延迟SLO。Grus深入分析了延迟变化的来源，并提出了三个设计原则：（1）需要动态设置批量大小以限制CPU中的数据包延迟;（2）需要通过PCI-E进行数据传输的重新排序机制以保证停滞时间;（3）需要最大化GPU中的并发性以避免NF执行等待时间。在这些原则的指导下，Grus由两个逻辑层组成，包括基础设施层和调度层。基础架构层配备了一个CPU内可重新排序的工作池，可以调整队列大小和数据包传输顺序，以及GPU内可控并发执行器，以提供最大化的并发性。调度层运行启发式算法来执行准确和快速的调度，以保证基于我们的预测模型的SLO。我们已经实现了Grus的原型。广泛的评估表明，Grus可以显着减少延迟变化，并满足比最先进的解决方案多4.5倍的SLO条款。
Graphics Processing Unit (GPU) has been recently exploited as a hardware accelerator to improve the performance of Network Function Virtualization (NFV). However, GPU-accelerated NFV systems suffer from significant latency variation when multiple network functions (NFs) are co-located in the same machine, which prevents operators from supporting latency Service Level Objectives (SLOs). Existing research efforts to address this problem can only guarantee a limited number of SLOs with very low resource utilization efficiency. In this paper, we present the Grus framework to support latency SLOs in GPU-accelerated NFV systems. Grus thoroughly analyzes the sources of latency variation and proposes three design principles: (1) dynamic batch size setting is needed to bound packet batching latency in CPU; (2) a reordering mechanism for data transfer over PCI-E is required to guarantee the stalling time; and (3) maximizing concurrency in GPU is necessary to avoid NF execution waiting time. Guided by the principles, Grus consists of two logical layers including an infrastructure layer and a scheduling layer. The infrastructure layer is equipped with an in-CPU Reorder-able Worker Pool that could adjust batching size and packet transfer order, and in-GPU Controllable Concurrent Executors to provide maximized concurrency. The scheduling layer runs a heuristic algorithm to perform accurate and fast scheduling to guarantee SLOs based on our prediction models. We have implemented a prototype of Grus. Extensive evaluations demonstrate that Grus can significantly reduce latency variation and satisfy 4.5× more SLO terms than state-of-the-art solutions.