Dynamic placement for clustered web applications

Karve, Alexei; Kimbrel, Tracy; Pacifici, G.; Spreitzer, Mike; Steinder, Małgorzata; Sviridenko, Maxim; Tantawi, Asser N.

doi:10.1145/1135777.1135865

Cited by 150 publications

(148 citation statements)

References 19 publications

Supporting

Mentioning

141

Contrasting

Unclassified

Order By: Relevance

“…This can be easily mapped to the different service types that need to be placed on multiple servers, with limited resources. The classic CMKP has been successfully adapted to the service placement problem [22,23]. Karve et al designed a centralized application placement middleware [22].…”

Section: Related Workmentioning

confidence: 99%

Network-aware service placement and selection algorithms on large-scale overlay networks

Famaey¹,

Wauters²,

Turck³

et al. 2011

Computer Communications

View full text Add to dashboard Cite

Currently many service providers offer their services on a private and proprietary hard-and software infrastructure. These infrastructures often share many similarities. Hence we believe a generic service management architecture, that allows service providers to offer a large array of different services on a single infrastructure or multiple providers to offer their services cooperatively, would provide many advantages over current silo-based approaches. Additionally, by allowing the distributed service management components to cooperate in a peerto-peer overlay network, scalability and resilience of the system could be greatly improved.In this paper we propose an optimal algorithm, based on an Integer Linear Programming (ILP) formulation, and several heuristics to support such a generic overlay-based service management architecture. More specifically, we propose algorithms for dynamically allocating server and network resources to a set of services and selecting a suitable service instance for each client. Service instances are placed on a set of servers, taking into account server resource constraints (e.g. CPU and memory). Unlike existing algorithms for this problem, those proposed in this paper also support Service Level Agreements (SLAs), which take the form of Quality of Service demands such as transmission latency constraints and bandwidth requirements. The optimization goal is to maximise the percentage of satisfied demand (answered requests) and minimise the total number of required * Corresponding author; Tel: +32 9 331 49 38; Fax: +32 9 331 48 99Email addresses: jeroen.famaey@intec.ugent.be (J. Famaey), tim.wauters@intec.ugent.be (T. Wauters), filip.deturck@intec.ugent.be (F. De Turck), bart.dhoedt@intec.ugent.be (B. Dhoedt), piet.demeester@intec.ugent.be (P. Demeester) Preprint submitted to ElsevierMarch 1, 2010 overlay servers, while satisfying the SLAs and resource constraints. Additionally, we propose an extension that allows the algorithms to find overlay routing paths to improve the transmission latency for latency-sensitive services.Extensive simulations were performed to evaluate the performance and scalability of the heuristics. They showed that in many cases the heuristics perform close to optimal and they scale well in terms of network size.

show abstract

Section: Related Workmentioning

confidence: 99%

Network-aware service placement and selection algorithms on large-scale overlay networks

Famaey¹,

Wauters²,

Turck³

et al. 2011

Computer Communications

View full text Add to dashboard Cite

show abstract

“…It also models the behavior of clients using Customer Behavior Modeling Graphs to characterize incoming workload. The analytical models developed in CAFe can be used along with the algorithms presented in [4,5] or [6].…”

Section: Related Workmentioning

confidence: 99%

“…Application placement in clustered servers has been studied extensively [1][2][3][4][5][6]. For example, [4][5][6] estimate resource usage of components and place components onto nodes by limiting their sum of resource usage within available node resource.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A Component Assignment Framework for Improved Capacity and Assured Performance in Web Portals

Roy

Gokhale

Dowdy

et al. 2009

On the Move to Meaningful Internet Systems: OTM 2009

View full text Add to dashboard Cite

Abstract. Web portals hosting large-scale internet applications have become popular due to the variety of services they provide to their users. These portals are developed using component technologies. Important design challenges for developers of web portals involve (1) determining the component placement that maximizes the number of users/requests (capacity) without increasing hardware resources and (2) maintaining the performance within certain bounds given by service level agreements (SLAs). The multitude of behavioral patterns presented by users makes it hard to identify the incoming workloads. This paper makes three contributions to the design and evaluation of web portals that address these design challenges. First it introduces an algorithmic framework that combines bin-packing and modeling-based queuing theory to place components onto hardware nodes. This capability is realized by the Component Assignment Framework for multi-tiered internet applications (CAFe). Second, it develops a component-aware queuing model to predict web portal performance. Third, it provides extensive experimental evaluation using the Rice University Bidding System (RUBiS). The results indicate that CAFe can identify opportunities to increase web portal capacity by 25% for a constant amount of hardware resources and typical web application and user workloads.

show abstract

“…The optimization problem that we consider presents a non-linear optimization objective while previous approaches [15,16] to similar problems address only linear optimization objectives. In [17], the authors evaluate a similar problem to that addressed in our work (restricted to transactional applications), and use a simulated annealing optimization algorithm.…”

Section: Related Workmentioning

confidence: 99%

Enabling Resource Sharing between Transactional and Batch Workloads Using Dynamic Application Placement

Carrera

Steinder²,

Whalley³

et al. 2008

Middleware 2008

View full text Add to dashboard Cite

Abstract. We present a technique that enables existing middleware to fairly manage mixed workloads: batch jobs and transactional applications. The technique leverages a generic application placement controller, which dynamically allocates compute resources to application instances. The controller works towards a fairness goal while also trying to maximize individual workload performance. We use relative performance functions to drive the application placement controller. Such functions are derived from workload-specific performance models-in the case of transactional workloads, we use queuing theory to build the performance model. For batch workloads, we evaluate a candidate placement by calculating long-term estimates of the completion times that are achievable with that placement according to a scheduling policy. In this paper, we propose a lowest relative performance first scheduling policy as a way to also achieve fair resource allocation among batch jobs. Our technique permits collocation of the workload types on the same physical hardware, and leverages control mechanisms such as suspension and migration to perform online system reconfiguration. In our experiments we demonstrate that our technique maximizes mixed workload performance while providing service differentiation based on highlevel performance goals.

show abstract

Dynamic placement for clustered web applications

Cited by 150 publications

References 19 publications

Network-aware service placement and selection algorithms on large-scale overlay networks

Network-aware service placement and selection algorithms on large-scale overlay networks

A Component Assignment Framework for Improved Capacity and Assured Performance in Web Portals

Enabling Resource Sharing between Transactional and Batch Workloads Using Dynamic Application Placement

Contact Info

Product

Resources

About