VMware Software-Defined Storage. Martin Hosken
Читать онлайн книгу.following provides a brief summary of the content in each of the 10 chapters:
Chapter 1 : Software-Defined Storage Design This chapter provides an overview of where vSphere storage technology is today, and how we've reached this point. This chapter also introduces software-defined storage, the economics of storage resources, and enabling storage as a service.
Chapter 2 : Classic Storage Models and Constructs This chapter covers the legacy and classic storage technologies that have been used in the VMware infrastructure for the last decade. This chapter provides the background required for you to understand the focus of this book, VMware vSphere's next-generation storage technology design.
Chapter 3 : Fabric Connectivity and Storage I/O Architecture This chapter presents storage connectivity and fabric architecture, which is relevant for legacy storage technologies as well as next-generation solutions including Virtual Volumes.
Chapter 4 : Policy-Driven Storage Design with Virtual SAN This chapter addresses all of the design considerations associated with VMware's Virtual SAN storage technology. The chapter provides detailed coverage of Virtual SAN functionality, design factors, and architectural considerations.
Chapter 5 : Virtual SAN Stretched Cluster Design This chapter focuses on one type of Virtual SAN solution, stretched cluster design. This type of solution has specific design and implementation considerations that are addressed in depth. This chapter also provides an example Virtual SAN stretched architecture design as a reference.
Chapter 6 : Designing for Web-Scale Virtual SAN Platforms This chapter addresses specific considerations associated with large-scale deployments of Virtual SAN hyper-converged infrastructure, commonly referred to as web-scale.
Chapter 7 Virtual SAN Use Case Library This chapter provides an overview of Virtual SAN use cases. It also provides a detailed solution architecture for a cloud management platform that you can use as a reference.
Chapter 8 : Policy-Driven Storage Design with Virtual Volumes This chapter provides detailed coverage of VMware's Virtual Volumes technology and its associated policydriven storage concepts This chapter also provides a lowlevel knowledge transfer as well as addressing in detail the design factors and architectural concepts associated with implementing Virtual Volumes
Chapter 9 : Delivering a Storage-as-a-Service Design This chapter explains how IT organizations and service providers can design and deliver storage as a service in a cloud-enabled data center by using VMware's cloud management platform technologies.
Chapter 10 : Monitoring and Storage Operations Design To ensure that a storage design can deliver an operationally efficient storage platform end to end, this final chapter covers storage monitoring and alerting design in the software-defined storage data center.
Chapter 1
Software-Defined Storage Design
VMware is the global leader in providing virtualization solutions. The VMware ESXi software provides a hypervisor platform that abstracts CPU, memory, and storage resources to run multiple virtual machines concurrently on the same physical server.
To successfully design a virtual infrastructure, other products are required in addition to the hypervisor, in order to manage, monitor, automate, and secure the environment. Fortunately, VMware also provides many of the products required to design an end-to-end solution, and to develop an infrastructure that is software driven, as opposed to hardware driven. This is commonly described as the software-defined data center (SDDC), illustrated in Figure 1.1.
Figure 1.1 Software-defined data center conceptual model
The SDDC is not a single product sold by VMware or anyone else. It is an approach whereby management and orchestration tools are configured to manage, monitor, and operationalize the entire infrastructure. This might include products such as vSphere, NSX, vRealize Automation, vRealize Operations Manager, and Virtual SAN from VMware, but it could also include solutions such as VMware Integrated OpenStack, CloudStack, or any custom cloud-management solution that can deliver the required platform management and orchestration capabilities.
The primary aim of the SDDC is to decouple the infrastructure from its underlying hardware, in order to allow software to take advantage of the physical network, server, and storage. This makes the SDDC location-independent, and as such, it may be housed in a single physical data center, span multiple private data centers, or even extend into hybrid and public cloud facilities.
From the end user’s perspective, applications that are delivered from an SDDC are consumed in exactly the same way as they otherwise would be – through mobile, desktop, and virtual desktop interfaces – from anywhere, any time, with any device.
However, with the SDDC infrastructure decoupled from the physical hardware, the operational model of a virtual machine – with on-demand provisioning, isolation, mobility, speed, and agility – can be replicated for the entire data-center environment (including networking and storage), with complete visibility, security, and scale.
The overall aim is that an SDDC can be achieved with the customer’s existing physical infrastructure, and also provide the flexibility for added capacity and new deployments.
Software-Defined Compute
In this book, software-defined compute refers to the compute virtualization of the x86 architecture. What is virtualization? If you don’t know the answer to this question, you’re probably reading the wrong book, but in any case, let’s make sure we’re on the same page.
In the IT industry, the term virtualization can refer to various technologies. However, from a VMware perspective, virtualization is the technique used for abstracting the physical hardware away from the operating system. This technique allows multiple guest operating systems (logical servers or desktops) to run concurrently on a single physical server. This allows these logical servers to become a portable virtual compute resource, called virtual machines. Each virtual machine runs its own guest operating system and applications in an isolated manner.
Compute virtualization is achieved by a hypervisor layer, which exists between the hardware of the physical server and the virtual machines. The hypervisor is used to provide hardware resources, such as CPU, memory, and network to all the virtual machines running on that physical host. A physical server can run numerous virtual machines, depending on the hardware resources available.
Although a virtual machine is a logical entity, to its operating system and end users, it seems like a physical host with its own CPU, memory, network controller, and disks. However, all virtual machines running on a host share the same underlying physical hardware, but each taking its own share in an isolated manner. From the hypervisor’s perspective, each virtual machine is simply a discrete set of files, which include a configuration file, virtual disk files, log files, and so on.
It is VMware’s ESXi software that provides the hypervisor platform, which is designed from the ground up to run multiple virtual machines concurrently, on the same physical server hardware.
Software-Defined Networking
Traditional