Mark S. Squillante

On maximizing service-level-agreement profits

We present a methodology for maximizing profits in a general class of e-commerce environments. The cost model is based on revenues that are generated when Quality-of-Service (QoS) guarantees are satisfied and on penalties that are incurred otherwise. The corresponding QoS criteria are derived from multiclass Service-Level-Agreements (SLAs) between service providers and their clients, which include the tail distributions of the per-class delays in addition to more standard QoS metrics such as throughput and mean delays. Our approach consists of formulating the optimization problem as a network flow model with a separable set of concave objective functions based on queueing-theoretic formulas, where the SLA classes are taken into account in both the constraints and the objective function. This problem is then solved via a fixed-point iteration. Numerous experiments illustrate the benefits of our approach.We present a methodology for maximizing profits in a general class of e-commerce environments. The cost model is based on revenues that are generated when Quality-of-Service (QoS) guarantees are satisfied and on penalties that are incurred otherwise. The corresponding QoS criteria are derived from multiclass Service-Level-Agreements (SLAs) between service providers and their clients, which include the tail distributions of the per-class delays in addition to more standard QoS metrics such as throughput and mean delays. Our approach consists of formulating the optimization problem as a network flow model with a separable set of concave objective functions based on queueing-theoretic formulas, where the SLA classes are taken into account in both the constraints and the objective function. This problem is then solved via a fixed-point iteration. Numerous experiments illustrate the benefits of our approach.

Failure data analysis of a large-scale heterogeneous server environment

The growing complexity of hardware and software mandates the recognition of fault occurrence in system deployment and management. While there are several techniques to prevent and/or handle faults, there continues to be a growing need for an in-depth understanding of system errors and failures and their empirical and statistical properties. This understanding can help evaluate the effectiveness of different techniques for improving system availability, in addition to developing new solutions. In this paper, we analyze the empirical and statistical properties of system errors and failures from a network of nearly 400 heterogeneous servers running a diverse workload over a year. While improvements in system robustness continue to limit the number of actual failures to a very small fraction of the recorded errors, the failure rates are significant and highly variable. Our results also show that the system error and failure patterns are comprised of time-varying behavior containing long stationary intervals. These stationary intervals exhibit various strong correlation structures and periodic patterns, which impact performance but also can be exploited to address such performance issues.

Using processor-cache affinity information in shared-memory multiprocessor scheduling

In a shared-memory multiprocessor system, it may be more efficient to schedule a task on one processor than on another if relevant data already reside in a particular processors cache. The effects of this type of processor affinity are examined. It is observed that tasks continuously alternate between executing at a processor and releasing this processor due to I/O, synchronization, quantum expiration, or preemption. Queuing network models of different abstract scheduling policies are formulated, spanning the range from ignoring affinity to fixing tasks on processors. These models are solved via mean value analysis, where possible, and by simulation otherwise. An analytic cache model is developed and used in these scheduling models to include the effects of an initial burst of cache misses experienced by tasks when they return to a processor for execution. A mean-value technique is also developed and used in the scheduling models to include the effects of increased bus traffic due to these bursts of cache misses. Only a small amount of affinity information needs to be maintained for each task. The importance of having a policy that adapts its behavior to changes in system load is demonstrated. >

Optimal crawling strategies for web search engines

Web Search Engines employ multiple so-called crawlers to maintain local copies of web pages. But these web pages are frequently updated by their owners, and therefore the crawlers must regularly revisit the web pages to maintain the freshness of their local copies. In this paper, we propose a two-part scheme to optimize this crawling process. One goal might be the minimization of the average level of staleness over all web pages, and the scheme we propose can solve this problem. Alternatively, the same basic scheme could be used to minimize a possibly more important search engine embarrassment level metric: The frequency with which a client makes a search engine query and then clicks on a returned url only to find that the result is incorrect. The first part our scheme determines the (nearly) optimal crawling frequencies, as well as the theoretically optimal times to crawl each web page. It does so within an extremely general stochastic framework, one which supports a wide range of complex update patterns found in practice. It uses techniques from probability theory and the theory of resource allocation problems which are highly computationally efficient -- crucial for practicality because the size of the problem in the web environment is immense. The second part employs these crawling frequencies and ideal crawl times as input, and creates an optimal achievable schedule for the crawlers. Our solution, based on network flow theory, is exact as well as highly efficient. An analysis of the update patterns from a highly accessed and highly dynamic web site is used to gain some insights into the properties of page updates in practice. Then, based on this analysis, we perform a set of detailed simulation experiments to demonstrate the quality and speed of our approach.

Analysis and characterization of large-scale Web server access patterns and performance

In this paper we develop a general methodology for characterizing the access patterns of Web server requests based on a time‐series analysis of finite collections of observed data from real systems. Our approach is used together with the access logs from the IBM Web site for the Olympic Games to demonstrate some of its advantages over previous methods and to construct a particular class of benchmarks for large‐scale heavily‐accessed Web server environments. We then apply an instance of this class of benchmarks to analyze aspects of large‐scale Web server performance, demonstrating some additional problems with methods commonly used to evaluate Web server performance at different request traffic intensities.

Performance implications of failures in large-scale cluster scheduling

As we continue to evolve into large-scale parallel systems, many of them employing hundreds of computing engines to take on mission-critical roles, it is crucial to design those systems anticipating and accommodating the occurrence of failures. Failures become a commonplace feature of such large-scale systems, and one cannot continue to treat them as an exception. Despite the current and increasing importance of failures in these systems, our understanding of the performance impact of these critical issues on parallel computing environments is extremely limited. In this paper we develop a general failure modeling framework based on recent results from large-scale clusters and then we exploit this framework to conduct a detailed performance analysis of the impact of failures on system performance for a wide range of scheduling policies. Our results demonstrate that such failures can have a significant impact on the mean job response time and mean job slowdown under existing scheduling policies that ignore failures. We therefore investigate different scheduling mechanisms and policies to address these performance issues. Our results show that periodic checkpointing of jobs seems to do little to ease this problem. On the other hand, we demonstrate that information about the spatial and temporal correlation of failure occurrences can be very useful in designing a scheduling (job allocation) strategy to enhance system performance, with the former providing the greatest benefits.

Evaluation of Multithreaded Uniprocessors for Commercial Application Environments

As memory speeds grow at a considerably slower rate than processor speeds, memory accesses are starting to dominate the execution time of processors, and this will likely continue into the future. This trend will be exacerbated by growing miss rates due to commercial applications, object-oriented programming and micro-kernel based operating systems. We examine the use of coarse-grained multithreading to address this important problem in uniprocessor on-line transaction processing environments where there is a natural, coarse-grained parallelism between the tasks resulting from transactions being executed concurrently, with no application software modifications required. Our results suggest that multithreading can provide significant performance improvements for uniprocessor commercial computing environments.

Performance analysis of job scheduling policies in parallel supercomputing environments

The authors analyze three general classes of scheduling policies under a workload typical of large-scale scientific computing. These policies differ in the manner in which processors are partitioned among the jobs as well as the way in which jobs are prioritized for execution on the partitions. The results indicate that existing static schemes to not perform well under varying workloads. Adaptive policies tend to make better scheduling decisions, but their ability to adjust to workload changes is limited. Dynamic partitioning policies, on the other hand, yield the best performance and can be tuned to provide desired performance differences among jobs with varying resource demands.

Efficiently serving dynamic data at highly accessed web sites

We present architectures and algorithms for efficiently serving dynamic data at highly accessed Web sites together with the results of an analysis motivating our design and quantifying its performance benefits. This includes algorithms for keeping cached data consistent so that dynamic pages can be cached at the Web server and dynamic content can be served at the performance level of static content. We show that our system design is able to achieve cache hit ratios close to 100% for cached data which is almost never obsolete by more than a few seconds, if at all. Our architectures and algorithms provide more than an order of magnitude improvement in performance using an order of magnitude fewer servers over that obtained under conventional methods.

A Hierarchical Approach for the Resource Management of Very Large Cloud Platforms

Worldwide interest in the delivery of computing and storage capacity as a service continues to grow at a rapid pace. The complexities of such cloud computing centers require advanced resource management solutions that are capable of dynamically adapting the cloud platform while providing continuous service and performance guarantees. The goal of this paper is to devise resource allocation policies for virtualized cloud environments that satisfy performance and availability guarantees and minimize energy costs in very large cloud service centers. We present a scalable distributed hierarchical framework based on a mixed-integer nonlinear optimization of resource management acting at multiple timescales. Extensive experiments across a wide variety of configurations demonstrate the efficiency and effectiveness of our approach.