In a modern chip-multiprocessor system, memory is a shared resource among multiple concurrently executing threads. The memory scheduling algorithm should resolve memory contention by arbitrating memory access in such a way that competing threads progress at a relatively fast and even pace, resulting in high system throughput and fairness. Previously proposed memory scheduling algorithms are predominantly optimized for only one of these objectives: no scheduling algorithm provides the best system throughput and best fairness at the same time.This paper presents a new memory scheduling algorithm that addresses system throughput and fairness separately with the goal of achieving the best of both. The main idea is to divide threads into two separate clusters and employ different memory request scheduling policies in each cluster. Our proposal, Thread Cluster Memory scheduling (TCM), dynamically groups threads with similar memory access behavior into either the latency-sensitive (memory-non-intensive) or the bandwidth-sensitive (memory-intensive) cluster. TCM introduces three major ideas for prioritization: 1) we prioritize the latency-sensitive cluster over the bandwidth-sensitive cluster to improve system throughput; 2) we introduce a "niceness" metric that captures a thread's propensity to interfere with other threads; 3) we use niceness to periodically shuffle the priority order of the threads in the bandwidth-sensitive cluster to provide fair access to each thread in a way that reduces inter-thread interference. On the one hand, prioritizing memory-non-intensive threads significantly improves system throughput without degrading fairness, because such "light" threads only use a small fraction of the total available memory bandwidth. On the other hand, shuffling the priority order of memory-intensive threads improves fairness because it ensures no thread is disproportionately slowed down or starved.We evaluate TCM on a wide variety of multiprogrammed workloads and compare its performance to four previously proposed scheduling algorithms, finding that TCM achieves both the best system throughput and fairness. Averaged over 96 workloads on a 24-core system with 4 memory channels, TCM improves system throughput and reduces maximum slowdown by 4.6%/38.6% compared to ATLAS (previous work providing the best system throughput) and 7.6%/4.6% compared to PAR-BS (previous work providing the best fairness).

A central question in designing server farms today is how to efficiently provision the number of servers to extract the best performance under unpredictable demand patterns while not wasting energy. While one would like to turn servers off when they become idle to save energy, the large setup cost (both, in terms of setup time and energy penalty) needed to switch the server back on can adversely affect performance. The problem is made more complex by the fact that today's servers provide multiple sleep or standby states which trade off the setup cost with the power consumed while the server is 'sleeping'. With so many controls, finding the optimal server farm management policy is an almost intractable problem -How many servers should be on at any given time, how many should be off, and how many should be in some sleep state? In this paper, we employ the popular metric of Energy-Response time Product (ERP) to capture the energy-performance tradeoff, and present the first theoretical results on the optimality of server farm management policies. For a stationary demand pattern, we prove that there exists a very small, natural class of policies that always contains the optimal policy for a single server, and conjecture it to contain a near-optimal policy for multi-server systems. For time-varying demand patterns, we propose a simple, traffic-oblivious policy and provide analytical and empirical evidence for its near-optimality.

Server farms today consume more than 1.5% of the total electricity in the U.S. at a cost of nearly $4.5 billion. Given the rising cost of energy, many industries are now seeking solutions for how to best make use of their available power. An important question which arises in this context is how to distribute available power among servers in a server farm so as to get maximum performance.By giving more power to a server, one can get higher server frequency (speed). Hence it is commonly believed that, for a given power budget, performance can be maximized by operating servers at their highest power levels. However, it is also conceivable that one might prefer to run servers at their lowest power levels, which allows more servers to be turned on for a given power budget. To fully understand the effect of power allocation on performance in a server farm with a fixed power budget, we introduce a queueing theoretic model, which allows us to predict the optimal power allocation in a variety of scenarios. Results are verified via extensive experiments on an IBM BladeCenter.We find that the optimal power allocation varies for different scenarios. In particular, it is not always optimal to run servers at their maximum power levels. There are scenarios where it might be optimal to run servers at their lowest power levels or at some intermediate power levels. Our analysis shows that the optimal power allocation is non-obvious and depends on many factors such as the power-to-frequency relationship in the processors, the arrival rate of jobs, the maximum server frequency, the lowest attainable server frequency and the server farm configuration. Furthermore, our theoretical model allows us to explore more general settings than we can implement, including arbitrarily large server farms and different power-to-frequency curves. Importantly, we show that the optimal power allocation can significantly * Research supported by NSF SMA/PDOS Grant CCR-0615262 and a 2009 IBM Faculty Award.Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. improve server farm performance, by a factor of typically 1.4 and as much as a factor of 5 in some cases.

No abstract

Tackling the questions that systems designers care about, this book brings queueing theory decisively back to computer science. The book is written with computer scientists and engineers in mind and is full of examples from computer systems, as well as manufacturing and operations research. Fun and readable, the book is highly approachable, even for undergraduates, while still being thoroughly rigorous and also covering a much wider span of topics than many queueing books. Readers benefit from a lively mix of motivation and intuition, with illustrations, examples and more than 300 exercises – all while acquiring the skills needed to model, analyze and design large-scale systems with good performance and low cost. The exercises are an important feature, teaching research-level counterintuitive lessons in the design of computer systems. The goal is to train readers not only to customize existing analyses but also to invent their own.