In earlier times, many companies deployed hardware servers to provide services to network users and customers. In an effort to be efficient, each service or workload such as email, web, print, directory and file was delivered from a separate server. This had the advantage that work could be carried out on an individual server without directly affecting the other machines. Additionally, failure of a single server did not interrupt services supplied by the remaining functioning systems.
Some downsides to this arrangement include the potential of multiple points of hardware failure, multiple complex hardware systems to manage, introduction of a new service could potentially require costly hardware acquisition and high energy costs. Moreover, very rarely was the full power of the hardware harnessed by the running applications. On top of all that, server redundancy added complexity and costs.
Server virtualization has been adopted as a model to address the shortcomings of the multiple physical server architecture while preserving the advantages. In server virtualization, a software application known as a hypervisor runs on top of the physical hardware and creates an environment in which multiple virtual servers, also known as virtual machines (VMs) can be created and operationalized.
Each VM is fully contained withits own processor, RAM, input/output resources and operating system (OS). Note that the hardware resources within the VM environment are really emulations of the corresponding physical hardware resources.The hypervisor abstracts the hardware from the context of the VMs making each VM independent of the hardware below it.
To address the issues associated with shared storage in the clustered environment, the hyperconverged infrastructure (HCI) was introduced. In this case, storage is virtualized using a technology known as software defined storage (SDS). Simply put, each physical server is outfitted with redundant locally attached storage devices. Each server is connected through a high-speed network fabric that allows storage device sharing across the entire cluster. Should failure of any storage device occur, the workloads on that failed device are instantly available on the remaining devices installed throughout the cluster.
The minimum network configuration required for redundancy with two nodes consists of redundant connections to two network switches. If desired, VLANs may be incorporated to separate storage and management traffic. A recommended configuration is illustrated in the following diagram.
Other possible configurations exist. Please contact TRUSTWORTHY Systems Inc. for further information.
Let TRUSTWORTHY Systems Inc. be the catalyst for risky behavior, change and development of a better digital experience within your organization.