A Survey of Resource Management in Containerized Cloud

Introduction Cloud computing ensures network based on-demand resource provisioning through the configurable shared pool of resources. It is being considered as a successful data centre architecture by many big as well as small and medium enterprises. Infrastructure as a Service (IaaS) model of Cloud provides hardware resources like CPU, memory, network bandwidth and storage. Ondemand resource provisioning in IaaS is materialised through server virtualization by means of hypervisor that divides physical resources into logical units known as Virtual Machines (VM) and assigns VMs to the clients on demand. Resources configured in a virtual machine can be increased or decreased as per the client’s requirement seamlessly. Virtual machines can also be migrated to a different physical machine in case source physical machine fails or it is short of resources. Thus virtualization-based cloud computing promises reliability, scalability and business continuity. In recent years, due to the emergence of big data, mobile computing and IoT, cloud-based workload has increased as well as services are experiencing random change in workload. Workload fluctuation eventually changes resource requirements of cloud-based services. Responding to a sudden change in resource requirements is a big challenge in front of cloud service providers. Virtual machines boot up with their own independent operating system. This makes them bulky and thus inefficient to respond quickly to a sudden change in resource demand (Zhang et al., 2018). Linux container is emerging as a lightweight alternative to a virtual machine. Linux container is an encapsulation of an application along with all libraries and framework components on which the application has runtime dependencies. A container is a logical unit similar to VM, but unlike VM, instead of having its own operating system, containers share the kernel of host operating system. This is also known as operating system virtualization

Read Full Article