User-facing, latency-sensitive services, such as websearch, underutilize their computing resources during daily periods of low traffic. Reusing those resources for other tasks is rarely done in production services since the contention for shared resources can cause latency spikes that violate the service-level objectives of latency-sensitive tasks. The resulting under-utilization hurts both the affordability and energy efficiency of large-scale datacenters. With the slowdown in technology scaling caused by the sunsetting of Moore's law, it becomes important to address this opportunity. We present Heracles, a feedback-based controller that enables the safe colocation of best-effort tasks alongside a latency-critical service. Heracles dynamically manages multiple hardware and software isolation mechanisms, such as CPU, memory, and network isolation, to ensure that the latency-sensitive job meets latency targets while maximizing the resources given to best-effort tasks. We evaluate Heracles using production latency-critical and batch workloads from Google and demonstrate average server utilizations of 90% without latency violations across all the load and colocation scenarios that we evaluated. CCS Concepts: r Computer systems organization → Cloud computing; r Software and its engineering → Scheduling
Abstract. Energy efficiency is becoming increasingly important in the operation of networking infrastructure, especially in enterprise and data center networks. Researchers have proposed several strategies for energy management of networking devices. However, we need a comprehensive characterization of power consumption by a variety of switches and routers to accurately quantify the savings from the various power savings schemes. In this paper, we first describe the hurdles in network power instrumentation and present a power measurement study of a variety of networking gear such as hubs, edge switches, core switches, routers and wireless access points in both stand-alone mode and a production data center. We build and describe a benchmarking suite that will allow users to measure and compare the power consumed for a large set of common configurations at any switch or router of their choice. We also propose a network energy proportionality index, which is an easily measurable metric, to compare power consumption behaviors of multiple devices.
Analysis of technology and application trends reveals a growing imbalance in the peak compute-to-memory-capacity ratio for future servers. At the same time, the fraction contributed by memory systems to total datacenter costs and power consumption during typical usage is increasing. In response to these trends, this paper reexamines traditional compute-memory co-location on a single system and details the design of a new general-purpose architectural building block-a memory blade-that allows memory to be "disaggregated" across a system ensemble. This remote memory blade can be used for memory capacity expansion to improve performance and for sharing memory across servers to reduce provisioning and power costs. We use this memory blade building block to propose two new system architecture solutions-(1) page-swapped remote memory at the virtualization layer, and (2) block-access remote memory with support in the coherence hardware-that enable transparent memory expansion and sharing on commodity-based systems. Using simulations of a mix of enterprise benchmarks supplemented with traces from live datacenters, we demonstrate that memory disaggregation can provide substantial performance benefits (on average 10X) in memory constrained environments, while the sharing enabled by our solutions can improve performance-per-dollar by up to 87% when optimizing memory provisioning across multiple servers.
Abstract-Networking devices today consume a non-trivial amount of energy and it has been shown that this energy consumption is largely independent of the load through the devices. With a strong need to curtail the rising operational costs of IT infrastructure, there is a tremendous opportunity for introducing energy awareness in the design and operation of enterprise and data center networks. We focus on these networks as they are under the control of a single administrative domain in which network-wide control can be consistently applied. In this paper, we describe and analyze three approaches to saving energy in single administrative domain networks, without significantly impacting the networks' ability to provide the expected levels of performance and availability. We also explore the tradeoffs between conserving energy and meeting performance and availability requirements. We conduct an extensive case study of our algorithms by simulating a real Web 2.0 workload in a real data center network topology using power characterizations that we obtain from real network hardware. Our results indicate that for our workload and data center scenario, 16% power savings (with no performance penalty and small decrease in availability) can be obtained merely by appropriately adjusting the active network elements (links). Significant additional savings (up to 75%) can be obtained by incorporating network traffic management and server workload consolidation.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.