Flexible Framework for Elasticity in Cloud Computing

TL;DR: This thesis investigated all the aspects of elasticity to manage efficiently the resources provisioning and de-provisioning in cloud computing and solved such challenges as how to overcome the problem of over- Provisioning and under-Provisioning.

Abstract: Recently, cloud computing has been gaining more popularity and has received a great deal of attention from both industrial and academic worlds. Industries and application providers have moved or plan to move to clouds in order to focus on their core business. This frees them from the burden and cost of managing their physical servers in local data center infrastructures. However, the main factor motivating the use of cloud computing is its ability to provide resources according to the customer’s needs or what is referred to as elastic provisioning and de-provisioning. Therefore, elasticity is one of the key features in cloud computing that dynamically adjusts the amount of allocated resources to meet changes in workload demands. The workload of cloud applications usually varies drastically over time and hence maintaining sufficient resources to meet peak requirements can be costly, and will increase the application provider’s functional cost. Conversely, if providers cut the costs by maintaining only minimum computing resources, there will not be sufficient resources to meet peak requirements and cause bad performance, violating Service Level Agreement (SLA). Therefore, adapting cloud applications during their execution according to demand variation is a challenging task. In addition, cloud elasticity is diverse and heterogeneous because it encompasses different approaches, policies, purposes, and applications. Furthermore, elasticity can be applied at the infrastructure level or application level. The infrastructure is powered by a certain virtualization technology such as VMware, Xen, containers or a provider-specific virtualization platform. We are interested in investigating: How to overcome the problem of over-provisioning and under-provisioning? How to guaranty resource availability? How to overcome the problems of heterogeneity and resource granularity? How to standardize, unify elasticity solutions and model its diversity at a high level of abstraction to manage its different aspects? In this thesis, we solved such challenges and we investigated all the aspects of elasticity to manage efficiently the resources provisioning and de-provisioning in cloud computing. It extended the state-of-the-art by making the following three contributions. Firstly, an up-to-date state-of-the-art of the cloud elasticity which reviews different works related to elasticity for both Virtual Machines (VMs) and containers. Secondly, ElasticDocker, an approach to manage container elasticity including vertical elasticity, live migration, and elasticity combination between different virtualization techniques. Thirdly, Model-Driven Elasticity Management with OCCI (MoDEMO), a new unified, standard-based, model-driven, highly extensible, highly reconfigurable elasticity management framework that supports multiple elasticity policies, both vertical and horizontal elasticities, different virtualization techniques and multiple cloud providers.