Martin Arlitt

The University of Calgary, Calgary, Alberta, Canada

Are you Martin Arlitt?

Claim your profile

Publications (119)22.58 Total impact

  • [Show abstract] [Hide abstract]
    ABSTRACT: The advent of multi‒core technology motivates new studies to understand how efficiently Web servers utilize such hardware. This paper presents a detailed performance study of a Web server application deployed on a modern eight‒core server. Our study shows that default Web server configurations result in poor scalability with increasing core counts. We study two different types of workloads, namely, a workload with intense TCP/IP related OS activity and the SPECweb2009 Support workload with more application‒level processing. We observe that the scaling behaviour is markedly different for these workloads, mainly because of the difference in the performance of static and dynamic requests. While static requests perform poorly when moving from using one socket to both sockets in the system, the converse is true for dynamic requests. We show that, contrary to what was suggested by previous work, Web server scalability improvement policies need to be adapted based on the type of workload experienced by the server. The results of our experiments reveal that with workload‒specific Web server configuration strategies, a multi‒core server can be utilized up to 80% while still serving requests without significant queuing delays; utilizations beyond 90% are also possible, while still serving requests with ‘acceptable’ response times. Copyright © 2014 John Wiley & Sons, Ltd.
    Concurrency and Computation Practice and Experience 03/2014; · 0.85 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Improving the performance and scalability of Web servers enhances user experiences and reduces the costs of providing Web-based services. The advent of Multi-core technology motivates new studies to understand how efficiently Web servers utilize such hardware. This paper presents a detailed performance study of a Web server application deployed on a modern 2 socket, 4-cores per socket server. Our study show that default, "out-of-the-box" Web server configurations can cause the system to scale poorly with increasing core counts. We study two different types of workloads, namely a workload that imposes intense TCP/IP related OS activity and the SPECweb2009 Support workload, which incurs more application-level processing. We observe that the scaling behaviour is markedly different for these two types of workloads, mainly due to the difference in the performance characteristics of static and dynamic requests. The results of our experiments reveal that with workload-specific Web server configuration strategies a modern Multi-core server can be utilized up to 80% while still serving requests without significant queuing delays; utilizations beyond 90% are also possible, while still serving requests with acceptable response times.
    Proceedings of the 4th ACM/SPEC International Conference on Performance Engineering; 04/2013
  • [Show abstract] [Hide abstract]
    ABSTRACT: Servers in most data centers are often underutilized due to concerns about SLA violations that may result from resource contention as server utilization increases. This low utilization means that neither the capital investment in the servers nor the power consumed is being used as effectively as it could be. In this paper, we present a novel method for managing the collocation of critical (e.g., user interactive) and non-critical (e.g., batch) workloads on virtualized multicore servers. Unlike previous cap-based solutions, our approach improves server utilization while meeting the SLAs of critical workloads by prioritizing resource access using Linux cgroups weights. Extensive experimental results suggest that the proposed work conserving collocation method is able to utilize a server to nearly 100% while keeping the performance loss of critical workloads within the specified limits.
    Integrated Network Management (IM 2013), 2013 IFIP/IEEE International Symposium on; 01/2013
  • [Show abstract] [Hide abstract]
    ABSTRACT: Power-laws are ubiquitous in the Internet and its applications. This tutorial presents a review of power-laws with emphasis on observations from Internet measurements. First, we introduce power-laws and describe two commonly observed power-law distributions, the Pareto and Zipf distributions. Two frequently occurring terms associated with these distributions, specifically heavy tails and long tails, are also discussed. Second, the preferential attachment model, which is a widely used model for generating power-law graph structures, is reviewed. Subsequently, we present several examples of Internet workload properties that exhibit power-law behavior. Finally, we explore several implications of power-laws in computer networks. Using examples from past and present, we review how researchers have studied and exploited power-law properties. We observe that despite the challenges posed, power-laws have been effectively leveraged by researchers to improve the design and performance of Internet-based systems.
    IEEE Network 01/2013; 27(1):59-64. · 2.85 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Cyber physical systems such as buildings contain entities (devices, appliances, etc.) that consume a multitude of resources (power, water, etc.). Efficient operation of these entities is important for reducing operating costs and environmental footprint of buildings. In this paper, we propose an entity characterization framework based on a finite state machine abstraction. Each state in the state machine is characterized in terms of distributions of sustainability or performance metrics of interest. This framework provides a basis for anomaly detection, assessment, prediction and usage pattern discovery. We demonstrate the usefulness of the framework using data from actual building entities. In particular, we apply our methodology to chillers and cooling towers, components of a building HVAC system.
    Proceedings of the Fourth ACM Workshop on Embedded Sensing Systems for Energy-Efficiency in Buildings; 11/2012
  • 10/2012;
  • [Show abstract] [Hide abstract]
    ABSTRACT: A world consisting of billions of service-oriented client devices and thousands of data centers can deliver a diverse range of services, from social networking to management of our natural resources. However, these services must scale in order to meet the fundamental needs of society. To enable such scaling, the total cost of ownership of the data centers that host the services and comprise the vast majority of service delivery costs will need to be reduced. As energy drives the total cost of ownership of data centers, there is a need for a new paradigm in design and management of data centers that minimizes energy used across their lifetimes, from “cradle to cradle”. This tutorial article presents a blueprint for a “net-zero data center”: one that offsets any electricity used from the grid via adequate on-site power generation that gets fed back to the grid at a later time. We discuss how such a data center addresses the total cost of ownership, illustrating that contrary to the oft-held view of sustainability as “paying more to be green”, sustainable data centers—built on a framework that focuses on integrating supply and demand management from end-to-end—can concurrently lead to lowest cost and lowest environmental impact.
    ACM Journal on Emerging Technologies in Computing Systems (JETC). 10/2012; 8(4).
  • Conference Paper: Adaptive green hosting
    [Show abstract] [Hide abstract]
    ABSTRACT: The growing carbon footprint of Web hosting centers contributes to climate change and could harm the public's perception of Web hosts and Internet services. A pioneering cadre of Web hosts, called green hosts, lower their footprints by cutting into their profit margins to buy carbon offsets. This paper argues that an adaptive approach to buying carbon offsets can increase a green host's total profit by exploiting daily, bursty patterns in Internet service workloads. We make the case in three steps. First, we present a realistic, geographically distributed service that meets strict SLAs while using green hosts to lower its carbon footprint. We show that the service routes requests between competing hosts differently depending on its request arrival rate and on how many carbon offsets each host provides. Second, we use empirical traces of request arrivals to compute how many carbon offsets a host should provide to maximize its profit. We find that diurnal fluctuations and bursty surges interrupted long contiguous periods where the best carbon offset policy held steady, leading us to propose a reactive approach. For certain hosts, our approach can triple the profit compared to a fixed approach used in practice. Third, we simulate 9 services with diverse carbon footprint goals that distribute their workloads across 11 Web hosts worldwide. We use real data on the location of Web hosts and their provided carbon offset policies to show that adaptive green hosting can increase profit by 152% for one of today's larger green hosts.
    Proceedings of the 9th international conference on Autonomic computing; 09/2012
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Exploring new options for publishing and content delivery offers an enormous opportunity to improve the state of the art and further modernize academic and professional publications.
    Computer 09/2012; 45(9):101-105. · 1.68 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Commercial buildings are significant consumers of electricity. The first step towards better energy management in commercial buildings is monitoring consumption. However, instrumenting every electrical panel in a large commercial building is expensive and wasteful. In this paper, we propose a greedy meter (sensor) placement algorithm based on maximization of information gained, subject to a cost constraint. The algorithm provides a near-optimal solution guarantee. Furthermore, to identify power saving opportunities, we use an unsupervised anomaly detection technique based on a low-dimensional embedding. Further, to better manage resources such as lighting and HVAC, we propose a semi-supervised approach combining hidden Markov models (HMM) and a standard classifier to model occupancy based on readily available port-level network statistics.
    08/2012;
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Workload generators are widely used for testing the performance of Web-based systems. Typically, these tools are also used to collect measurements such as throughput and end-user response times that are often used to characterize the QoS provided by a system to its users. However, our study finds that Web workload generation is more difficult than it seems. In examining the popular RUBiS client generator, we found that reported response times could be grossly inaccurate, and that the generated workloads were less realistic than expected, causing server scalability to be incorrectly estimated. Using experimentation, we demonstrate how the Java virtual machine and the Java network library are the root causes of these issues. Our work serves as an example of how to verify the behavior of a Web workload generator. Copyright © 2011 John Wiley & Sons, Ltd.
    Software Practice and Experience 04/2012; 42(5):629 - 647. · 1.01 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Workload characterization is important for understanding how systems and services are used in practice and to help identify design improvements. To better understand the longitudinal workload dynamics of chunk-based content delivery systems, this paper analyzes the BitTorrent usage as observed from two different vantage points. Using two simultaneously collected 48-week long traces, we analyze the differences in download characteristics and popularity dynamics observed locally at a university campus versus at a global scale. We find that campus users typically download larger files and are early adopters of new content, in the sense that they typically download files well before the time at which the global popularity of the files peak. The noticeable exception is music files, which the campus users are late to download. We also find that there typically is high churn in the set of files that are popular each week, both locally and globally, and that the most popular files peak significantly later than their release date. These findings provide insights that may improve the efficiency of content sharing locally, and thus increase the scalability of the global system.
    Proceedings of the 13th international conference on Passive and Active Measurement; 03/2012
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: This position paper explores research challenges facing carbon-aware cloud applications. These applications run inside of a renewable-energy datacenter, provision resources on demand, and seek to minimize their use of carbon-heavy, grid energy. First, we argue that carbon-aware applications need new provisioning policies that address the uncertainty of renewable energy. Second, we argue that renewable-energy datacenters need mechanisms to determine the contribution of grid energy to specific application workloads. We propose first-cut solutions to these problems and present preliminary results for carbon-aware Web server running on a small renewable-energy cluster.
    01/2012;
  • 01/2012;
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Cyberlockers have recently become a very popular means of distributing content. Today, cyberlocker traffic accounts for a non-negligible fraction of the total Internet traffic volume, and is forecasted to grow significantly in the future. The underlying protocol used in cyberlockers is HTTP, and increased usage of these services could drastically alter the characteristics of Web traffic. In light of the evolving nature of Web traffic, updated traffic models are required to capture this change. Despite their popularity, there has been limited work on understanding the characteristics of traffic flows originating from cyberlockers. Using a year-long trace collected from a large campus network, we present a comprehensive characterization study of cyberlocker traffic at the transport layer. We use a combination of flow-level and host-level characteristics to provide insights into the behavior of cyberlockers and their impact on networks. We also develop statistical models that capture the salient features of cyberlocker traffic. Studying the transport-layer interaction is important for analyzing reliability, congestion, flow control, and impact on other layers as well as Internet hosts. Our results can be used in developing improved traffic simulation models that can aid in capacity planning and network traffic management.
    Local Computer Networks (LCN), 2012 IEEE 37th Conference on; 01/2012
  • [Show abstract] [Hide abstract]
    ABSTRACT: The World Wide Web and the services it provides are continually evolving. Even for a single time instant, it is a complex task to methodologically determine the infrastructure over which these services are provided and the corresponding effect on user perceived performance. For such tasks, researchers typically rely on active measurements or large numbers of volunteer users. In this paper, we consider an alternative approach, which we refer to as passive crowd-based monitoring. More specifically, we use passively collected proxy logs from a global enterprise to observe differences in the quality of service (QoS) experienced by users on different continents. We also show how this technique can measure properties of the underlying infrastructures of different Web content providers. While some of these properties have been observed using active measurements, we are the first to show that many of these properties (such as location of servers) can be obtained using passive measurements of actual user activity. Passive crowd-based monitoring has the advantages that it does not add any overhead on Web infrastructure, it does not require any specific software on the clients, but still captures the performance and infrastructure observed by actual Web usage.
    Communications (ICC), 2012 IEEE International Conference on; 01/2012
  • Source
    Martin Arlitt
    [Show abstract] [Hide abstract]
    ABSTRACT: Time tends to pass more quickly than we would like. Sometimes it is helpful to reflect on what you have accomplished, and to derive what you have learned from the experiences. These "lessons learned" may then be leveraged by yourself or others in the future. Occasionally, an external event will motivate this self reflection. For me, it was the 50th anniversary reunion of the St. Walburg Eagles, held in July 2011. The Eagles are a full-contact (ice) hockey team I played with between 1988 and 1996 (the Eagles ceased operations twice during this period, which limited me to four seasons playing with them), while attending university. What would I tell my friends and former teammates that I had been doing for the past 15 years? After some thought, I realized that my time as an Eagle had prepared me for a research career, in ways I would never have imagined. This article (an extended version with color photos is available in [1]) shares some of these similarities, to motivate others to reflect on their own careers and achievements, and perhaps make proactive changes as a result.
    Computer Communication Review - CCR. 01/2012;
  • [Show abstract] [Hide abstract]
    ABSTRACT: Web-based services are used by many organizations to support their customers and employees. An important consideration in developing such services is ensuring the Quality of Service (QoS) that users experience is acceptable. Recent years have seen a shift toward deploying Web service son multi-core hardware. Leveraging the performance benefits of multi-core hardware is a non-trivial task. In particular, systematic Web server benchmarking techniques are needed so organizations can verify their ability to meet customer QoS objectives while effectively utilizing such hardware. However, our recent experiences suggest that the multi-core era imposes significant challenges to Web server benchmarking. In particular, due to limitations of current hardware monitoring tools, we found that a large number of experiments are needed to detect complex bottlenecks that can arise in a multi-core system due to contention for shared resources such as cache hierarchy, memory controllers and processor inter-connects. Furthermore, multiple load generator instances are needed to adequately stress multi-core hardware. This leads to practical challenges in validating and managing the test results. This paper describes the automation strategies we employed to overcome these challenges. We make our test harness available for other researchers and practitioners working on similar studies.
    01/2012;
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Reduction of resource consumption in data centers is becoming a growing concern for data center designers, operators and users. Accordingly, interest in the use of renewable energy to provide some portion of a data center's overall energy usage is also growing. One key concern is that the amount of renewable energy necessary to satisfy a typical data center's power consumption can lead to prohibitively high capital costs for the power generation and delivery infrastructure, particularly if on-site renewables are used. In this paper, we introduce a method to operate a data center with renewable energy that minimizes dependence on grid power while minimizing capital cost. We achieve this by integrating data center demand with the availability of resource supplies during operation. We discuss results from the deployment of our method in a production data center.
    Thermal and Thermomechanical Phenomena in Electronic Systems (ITherm), 2012 13th IEEE Intersociety Conference on; 01/2012
  • [Show abstract] [Hide abstract]
    ABSTRACT: Commercial buildings are significant consumers of electricity. In this paper, we collect and analyze six weeks of data from 39 power meters in three buildings of a campus of a large company. We use an unsupervised anomaly detection technique based on a low-dimensional embedding to identify power saving opportunities. Further, to better manage resources such as lighting and HVAC, we develop occupancy models based on readily available port-level network logs. We propose a semi-supervised approach that combines hidden Markov models (HMM) with standard classifiers such as naive Bayes and support vector machines (SVM). This two step approach simplifies the occupancy model while achieving good accuracy. The experimental results over ten office cubicles show that the maximum error is less than 15% with an average error of 9.3%. We demonstrate that using our occupancy models, we can potentially reduce the lighting load on one floor (about 45 kW) by about 9.5%.
    11/2011;

Publication Stats

4k Citations
22.58 Total Impact Points

Institutions

  • 2002–2012
    • The University of Calgary
      • • Department of Electrical and Computer Engineering
      • • Department of Computer Science
      Calgary, Alberta, Canada
  • 2000–2012
    • Hewlett-Packard
      Palo Alto, California, United States
  • 2005
    • Palo Alto University
      Palo Alto, California, United States
    • University of Illinois, Urbana-Champaign
      Urbana, Illinois, United States
  • 2000–2004
    • FX Palo Alto Laboratory
      Palo Alto, California, United States
  • 1970–1999
    • University of Saskatchewan
      • Department of Computer Science
      Saskatoon, Saskatchewan, Canada