Proactive hot spot avoidance for Web server dependability

Felber, Pascal; Kaldewey, Tim; Weiss, Stéphane

doi:10.1109/reldis.2004.1353031

Cited by 11 publications

(6 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Redirection of incoming client requests or sessions remains active until the server load is below the threshold. Other studies, such as [4], [3], exploit a double threshold solution, where redirection is activated by a High This latter solution provides a more stable behavior, reducing the number of activation/deactivation of redirection with respect to the single threshold solution, but its higher complexity requires the tuning of the two threshold values to achieve good performance.…”

Section: B Request Managementmentioning

confidence: 99%

“…Decisions concerning request redirection are usually enforced through the evaluation of the load conditions of the local server and (possibly) of its neighbors, and through the comparison of these load values against a fixed, static threshold [1], [2]. Even if this approach has been applied in several fields of ranging from network resource optimization [3] to performance improvement on a local [4], [2] and geographic scale [5], these proposals do not consider pieces of information that could improve redirection decisions, but rely only on a comparison of the server load with static thresholds. This approach has some limits: (a) it does not consider the redirection overhead, that includes the delays introduced by request forwarding and session migration; (b) it does not take into account the computational demand of a user request and, consequently, the impact of its redirection on the load of other servers.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Dynamic Request Management Algorithms for Web-Based Services in Cloud Computing

Lancellotti

Andreolini

Canali

et al. 2011

2011 IEEE 35th Annual Computer Software and Applications Conference

View full text Add to dashboard Cite

Providers of Web-based services can take advantage of many convenient features of cloud computing infrastructures, but they still have to implement request management algorithms that are able to face sudden peaks of requests. We consider distributed algorithms implemented by front-end servers to dispatch and redirect requests among application servers. Current solutions based on load-blind algorithms, or considering just server load and thresholds are inadequate to cope with the demand patterns reaching modern Internet application servers. In this paper, we propose and evaluate a request management algorithm, namely Performance Gain Prediction, that combines several pieces of information (server load, computational cost of a request, user session migration and redirection delay) to predict whether the redirection of a request to another server may result in a shorter response time. To the best of our knowledge, no other study combines information about infrastructure status, user request characteristics and redirection overhead for dynamic request management in cloud computing. Our results show that the proposed algorithm is able to reduce the response time with respect to existing request management algorithms operating on the basis of thresholds.

show abstract

Section: B Request Managementmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Dynamic Request Management Algorithms for Web-Based Services in Cloud Computing

Lancellotti

Andreolini

Canali

et al. 2011

2011 IEEE 35th Annual Computer Software and Applications Conference

View full text Add to dashboard Cite

show abstract

“…For example, existing strategies for redirecting load are founded on one or two thresholds for deciding about activation/deactivation [14]. An approach with fixed threshold(s) is clearly not autonomic and risks to lead the system to unstable oscillations.…”

Section: Related Workmentioning

confidence: 99%

“…It is self-adaptable and guarantees best stability especially when it is enriched by a probabilistic approach to decide about redirection. On the other hand, existing strategies are founded on a threshold basis for activation/deactivation [14], that is not autonomic and risks to lead the system to unstable oscillations.…”

Section: Introductionmentioning

confidence: 99%

Autonomic Request Management Algorithms for Geographically Distributed Internet-Based Systems

Andreolini

Casolari

Colajanni

2008

2008 Second IEEE International Conference on Self-Adaptive and Self-Organizing Systems

View full text Add to dashboard Cite

Supporting Web-based services through geographical distributed clusters of servers is a common solution to the increasing volume and variability of modern traffic. These architectures pose interesting challenges to request management strategies where the most important goal is not to achieve maximum performance, but to guarantee stable and robust results. In this paper, we propose novel request management algorithms that are based on autonomic principles that is, on loose collaboration among the closest nodes and no knowledge about the global system state. Experimental evaluation shows that our autonomic-enhanced algorithms can guarantee robust performance in a variety of settings and reduce standard deviations of the response times with respect to existing request management algorithms.

show abstract

“…Most of them ( [17], [39], [36], [26], [37], [15]) have focused on static content. To improve the application server scalability, application programs or components [28], [2], [44] can be offloaded from the origin server.…”

Section: Related Workmentioning

confidence: 99%

Enabling On-demand Query Result Caching in DotSlash for Handling Web Hotspots Effectively

Zhao

Schulzrinne

2006

2006 1st IEEE Workshop on Hot Topics in Web Systems and Technologies

View full text Add to dashboard Cite

DotSlash is an automated web hotspot rescue system, which allows different web sites to form a mutual-aid community, and use spare capacity in the community to relieve web hotspots experienced by any individual site. DotSlash rescue services enable a web site to build an adaptive distributed web server system on the fly and replicate application programs dynamically, which relieve a spectrum of bottlenecks ranging from access network bandwidth to web servers and application servers. This paper presents DotSlash Qcache services that allow a web site to use on-demand distributed query result caching to greatly reduce the workload at read-mostly databases. The novelty of this work is that our query result caching is on demand and operated based on load conditions, which offers good data consistency for normal load and good scalability with relaxed data consistency under heavy load. DotSlash Qcache services complement DotSlash rescue services; together they provide a comprehensive solution to address different bottlenecks at multitier web sites. Experiments show that using DotSlash a web site can increase its maximum request rate supported by a factor of 10 for the RUBBoS read-only mix.

show abstract

Proactive hot spot avoidance for Web server dependability

Abstract: Abstract

Cited by 11 publications

References 13 publications

Dynamic Request Management Algorithms for Web-Based Services in Cloud Computing

Dynamic Request Management Algorithms for Web-Based Services in Cloud Computing

Autonomic Request Management Algorithms for Geographically Distributed Internet-Based Systems

Enabling On-demand Query Result Caching in DotSlash for Handling Web Hotspots Effectively

Contact Info

Product

Resources

About