Performance implications of multi-tier application deployments on Infrastructure-as-a-Service clouds: Towards performance modeling

Abstract

a b s t r a c tHosting a multi-tier application using an Infrastructure-as-a-Service (IaaS) cloud requires deploying components of the application stack across virtual machines (VMs) to provide the application's infrastructure while considering factors such as scalability, fault tolerance, performance and deployment costs (# of VMs). This paper presents results from an empirical study which investigates implications for application performance and resource requirements (CPU, disk and network) resulting from how multi-tier applications are deployed to IaaS clouds. We investigate the implications of: (1) component placement across VMs, (2) VM memory size, (3) VM hypervisor type (KVM vs. Xen), and (4) VM placement across physical hosts (provisioning variation). All possible deployment configurations for two multi-tier application variants are tested. One application variant was computationally bound by the application middleware, the other bound by geospatial queries. The best performing deployments required as few as 2 VMs, half the number required for VM-level service isolation, demonstrating potential cost savings when components can be consolidated. Resource utilization (CPU time, disk I/O, and network I/O) varied with component deployment location, VM memory allocation, and the hypervisor used (Xen or KVM) demonstrating how application deployment decisions impact required resources. Isolating application components using separate VMs produced performance overhead of ∼1%-2%. Provisioning variation of VMs across physical hosts produced overhead up to 3%. Relationships between resource utilization and performance were assessed using multiple linear regression to develop a model to predict application deployment performance. Our model explained over 84% of the variance and predicted application performance with mean absolute error of only ∼0.3 s with CPU time, disk sector reads, and disk sector writes serving as the most powerful predictors of application performance.

Risks and rewards of cloud computing in the UK public sector: A reflection on three Organisational case studies

“…& Easier Application Migration: According to Youseff et al (2008) and Lloyd et al (2013), in an IaaS cloud environment application migration is focused on efficient moving of applications back and forth between a virtualised data centre and private cloud environment and public clouds. However, there were intermittent problems with the virtualisation and storage area network software deployed in the case study cloud solution, which made this a difficult transition for UKLA1.…”

FCMS: A fuzzy controller for CPU and memory consolidation under SLA constraints

“…Second, we must be able to coordinate the allocation of CPU and memory, in order to avoid the arising of the effects caused by the interactions of the mechanisms used to carry out them. In particular, as reported in the literature, when the amount of memory allocated to a given VM is changed, the utilization of the corresponding vCPU suddenly increases for a magnitude and a duration that vary according to the magnitude of memory variation and the type of operation (i.e., memory increment or decrement), and shows a bursty behavior.…”

Heuristic virtual machine allocation for multi-tier Ambient Assisted Living applications in a cloud data center

“…There are an increasing number of papers that focus on performance modeling of typical multi-tier applications [33][34][35]. Authors in [33] explore several results about resource allocation and application performance.…”

“…Authors in [33] explore several results about resource allocation and application performance. Their results can provide some directions for the deployment of multi-tier applications in the cloud infrastructure.…”