QoS-driven server migration for Internet data centers

Ranjan, Supranamaya; Rolia, Jerry; Fu, Huirong; Knightly, Edward W.

doi:10.1109/iwqos.2002.1006569

Cited by 125 publications

(103 citation statements)

References 8 publications

Supporting

Mentioning

101

Contrasting

Unclassified

Order By: Relevance

“…It should be pointed out that a number of similar models for simple, replicable applications have been proposed in recent works (Doyle et al, 2003;Ranjan et al, 2002;Villela et al, 2004) and any of these could potentially be used by our provisioning algorithm. The comparison of our model with these other models is not relevant to our current discussion of overload management and therefore beyond the scope of this work.…”

Section: Article In Pressmentioning

confidence: 99%

“…Some papers have addressed the problem of provisioning resources at the granularity of individual servers as in our work. Ranjan et al (2002) consider the problem of dynamically varying the number of servers assigned to a single service hosted on a data center. Their objective is to minimize the number of servers needed to meet the service's QoS targets.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Cataclysm: Scalable overload policing for internet applications

Urgaonkar

Shenoy

2008

Journal of Network and Computer Applications

View full text Add to dashboard Cite

In this paper we present the Cataclysm server platform for handling extreme overloads in hosted Internet applications. The primary contribution of our work is to develop a low overhead, highly scalable admission control technique for Internet applications. Cataclysm provides several desirable features, such as guarantees on response time by conducting accurate size-based admission control, revenue maximization at multiple time-scales via preferential admission of important requests and dynamic capacity provisioning, and the ability to be operational even under extreme overloads. Cataclysm can transparently trade-off the accuracy of its decision making with the intensity of the workload allowing it to handle incoming rates of several tens of thousands of requests/second. We implement a prototype Cataclysm hosting platform on a Linux cluster and demonstrate the benefits of our integrated approach using a variety of workloads. r

show abstract

Section: Article In Pressmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Cataclysm: Scalable overload policing for internet applications

Urgaonkar

Shenoy

2008

Journal of Network and Computer Applications

View full text Add to dashboard Cite

show abstract

“…Modeling of a singletier system, such as a simple HTTP server, has been studied extensively. Even for a multi-tier structure which is employed ubiquitously for most servers, the system is usually abstracted as the most bottle-necked tier only: in [16], only the application tier for the e-commerce systems are modeled by a M/GI/1/PS queue; similarly in [19] the application tier with N node cluster is modeled by a G/G/N queue. Recently B. Urgaonkar et al proposed analytic models for both open and closed multi-tier systems [17,18].…”

Section: Related Workmentioning

confidence: 99%

R-Capriccio: A Capacity Planning and Anomaly Detection Tool for Enterprise Services with Live Workloads

Zhang¹,

Cherkasova²,

Mathews³

et al. 2007

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. As the complexity of IT systems increases, performance management and capacity planning become the largest and most difficult expenses to control. New methodologies and modeling techniques that explain large-system behavior and help predict their future performance are now needed to effectively tackle the emerging performance issues. With the multi-tier architecture paradigm becoming an industry standard for developing scalable client-server applications, it is important to design effective and accurate performance prediction models of multi-tier applications under an enterprise production environment and a real workload mix. To accurately answer performance questions for an existing production system with a real workload mix, we design and implement a new capacity planning and anomaly detection tool, called R-Capriccio, that is based on the following three components: i) a Workload Profiler that exploits locality in existing enterprise web workloads and extracts a small set of most popular, core client transactions responsible for the majority of client requests in the system; ii) a Regression-based Solver that is used for deriving the CPU demand of each core transaction on a given hardware; and iii) an Analytical Model that is based on a network of queues that models a multi-tier system. To validate R-Capriccio, we conduct a detailed case study using the access logs from two heterogeneous production servers that represent customized client accesses to a popular and actively used HP Open View Service Desk application.

show abstract

“…In (Liu et al, 2005) authors proposes the design of online feedback control algorithms to dynamically adjust entitlement values for a resource container on a server shared by multiple applications. The problem of provisioning resources in cluster architectures has been addressed in (Appleby et al, 2001;Ranjan et al, 2002) by allocating entire machines (dedicated model) and in (Chandra et al, 2003a;Pradhan et al, 2002;Uragonkar and Shenoy, 2004) by sharing node resources among multiple applications (shared model). Cataclysm (Sotomayor and Childers, 2005) performs overload control by bringing together admission control, adaptive service degradation and dynamic provisioning of platform resources, demonstrating that the most effective way to handle overloading must consider a combination of techniques.…”

Section: Related Workmentioning

confidence: 99%

Dynamic Resource Provisioning for Self-Adaptive Heterogeneous Workloads in SMP Hosting Platforms

2007

Proceedings of the Second International Conference on E-Business

View full text Add to dashboard Cite

Abstract:We introduce a novel approach that allows heterogeneous applications run together on a shared hosting platform, dynamically sharing the platform's resources. The proposed approach has been validated by a proofof-concept prototype which uses a global processor manager to distribute the platform's processors among two (or more) heterogeneous applications, i.e. a Tomcat application server and a Globus grid middleware. Our evaluation demonstrates the benefit of including bidirectional communication between applications and the OS for efficiently managing the resources and preventing the degradation of an applications performance, especially when the hosting platform is fully overloaded. For the sake of simplicity, we have modified the applications so that they communicate with the resource manager, although other techniques can be applied to avoid these modifications. Running different applications in a shared platform and being able to assign priorities between them provides important benefits.

show abstract

QoS-driven server migration for Internet data centers

Cited by 125 publications

References 8 publications

Cataclysm: Scalable overload policing for internet applications

Cataclysm: Scalable overload policing for internet applications

R-Capriccio: A Capacity Planning and Anomaly Detection Tool for Enterprise Services with Live Workloads

Dynamic Resource Provisioning for Self-Adaptive Heterogeneous Workloads in SMP Hosting Platforms

Contact Info

Product

Resources

About