Home > Articles

Introduction to Oracle Databases on Virtual Infrastructure

  • Print
  • + Share This
99.9% of all database or data management systems should be considered candidates for virtualization on vSphere. In this chapter from Virtualizing Oracle Databases on vSphere, the authors argue that Oracle databases and software are prime candidates to consider migrating to virtualized infrastructure.
This chapter is from the book

Oracle databases and software run successfully on vSphere and provide significant scalability, availability, and performance benefits. In fact, virtualization quite simply makes Oracle better. What makes the transition smooth is that the Oracle database administrator’s (DBA) skill set, deployment technique, and responsibilities do not change when transitioning from a physical to virtual environment. However, it is important that the DBA’s scope of responsibility does increase in breadth.

In years past, the DBA’s concerns moved into the nedtwork realm as Oracle introduced horizontal scalability with Oracle Cluster (later to become Oracle Parallel Server and finally Oracle Real Application Cluster [RAC]). Over the past decade, the focus of the DBA grew to include storage as Oracle introduced Automatic Storage Management (ASM), and thus, the DBA was confronted with a lack of willingness on the part of the storage administrator to manage the ASM instance. Similarly today, we see the realm of the DBA extend into the virtualization arena. Importantly, it is axiomatic that basic database administration skills do not change when virtualization is included in the stack. This is because ESXi does not alter the kernel of any guest operating system (OS). Likewise, it is equally important for the DBA to embrace the fact that the some components of the stack that affect the database have been extended into the virtualized infrastructure, such as networking, storage access, processing capability, and memory. Consequently, the areas of concern for the DBA have been extended, respectively.

Throughout this book, we maintain a conversational tone along with a thematic approach to the organization centered on the idea of the four V’s. Often, it can be both entertaining and memorable to point out certain technology industry trends. The trend of the monopolization of specific letters by certain well-known companies is an example. We all know who dominates the use of the letter f or t or i or even O. Ironically, VMware prominently uses the uppercase letter V to begin the name VMware, and VMW is the acronym most often associated with VMware. However the main product vSphere begins with a lowercase v, and it is commonplace for the individual features to begin with the lowercase letter v. The thematic approach referred to earlier will be centered on that letter V, although no adherence to the case will be necessary.

The four V’s are viability, value, versatility, and vision and are shown in Figure 1-1. The first step in any early implementation process is to convince the critical stakeholders such as the DBAs and information technology (IT) managers that virtualized infrastructure is a viable alternative to nonvirtualized infrastructure. Notice the subtlety in the reference to “nonvirtualized infrastructure” as opposed to the more common reference of “physical infrastructure.” Subsequent to the viability concerns being satisfied, we move on to the value discussion. The value of vSphere as a platform of virtualized hardware for business-critical applications (BCAs) is discussed in great detail as the chapters of this book unfold. The discussion eventually leads to the versatility of vSphere and VMware broadly. VMW field personnel will proudly state to all prospective customers that “We are agnostic to both logical and physical architectures.” And finally, no discussion on this subject is complete without acknowledgment of the future vision of vSphere’s capabilities and VMware as a corporation. The choice of transitioning a company’s entire IT architecture to this wonderful “Platform of Virtualized Hardware” is succinct in description, profound in consequence, but colossal in potential. Therefore, every potential stakeholder in the company should have an understanding of this vision.

Figure 1-1

Figure 1-1 Four V’s approach

Virtualization with ESXi and vSphere and the Software-Defined Datacenter

ESX or ESXi is the world’s leading x86 hypervisor. Hypervisors were first introduced at IBM in the 1960s by abstracting the machine’s supervisor state and allowing multiple virtual machines (VMs) to run simultaneously in separate VM contexts. A hypervisor, sometimes scientifically referred to as a virtual machine monitor (VMM), is software that allows for the creation, management, and runtime execution of independent VMs running their own guest operating systems. The physical machine that the hypervisor runs on is referred to as the host machine.

A number of hypervisor types exist. A Type 1 hypervisor, such as ESXi, runs on the bare metal of the computer. The VMs are created on the layer above ESXi and the guest OS runs within that second layer. ESXi has a minimal memory footprint (144M for vSphere version 5.x). A Type 2 hypervisor runs within a base OS, and therefore the guest operating systems run on the third layer above the hardware. See Figure 1-2.

Figure 1-2

Figure 1-2 Type 1 and Type 2 hypervisor

It is also important to point out that ESXi is nonparavirtualized. Paravirtualization is discussed later in the book, but it is important to understand that this means that no guest OS kernel is altered and that there therefore exists a perfect state of abstraction between the guest OS and the hardware.

ESXi is a hypervisor, but vSphere is a “platform of virtualized hardware.” It is logical hardware and should always be described and understood as such. And from this point on, we refer to virtualized hardware and nonvirtualized hardware.

The software-defined datacenter (SDDC) is a philosophy of architecture, not an actual product, although it does imply comprehensive full-stack virtualization. In a complete SDDC, all elements of the datacenter are virtualized. VMware has been very successful virtualizing the server components to include processing with virtual CPUs (vCPUs) and memory allocated to the VM (we avoid using the phrase virtual memory because it has other well-accepted connotations), but virtualization of the network and storage have been elusive. It is true that common terms such as virtual disk (VMDK), virtual network interface card (vNIC), and virtual distributed switch (vDS) all imply virtualization, but that is not the reality. An important attribute of true virtualization is abstraction. See Figure 1-3.

Figure 1-3

Figure 1-3 VMware SDDC (high level)

As VMware moves into the next phase of technology, the SDDC will include not only the virtualization of the server but also the virtualization of the network and storage. In 2012, VMware acquired Nicira Corporation, whose technology has led to the Network Virtualization and Security platform (NSX), which does meet anyone’s strict definition of true network virtualization. Figure 1-4 shows vSphere with NSX incorporated. Subsequent chapters cover specific networking recommendations.

Figure 1-4

Figure 1-4 VMware NSX

For more information on VMware NSX, refer to the link below.

www.vmware.com/products/nsx.

Storage virtualization comes in many flavors. Often, the phrase is used to describe the storage paradigms used only with “stretch clusters,” in which a single logical unit (LUN) of storage will exist in two different physical locations but synchronicity will be maintained through disk mirroring. Systems such as IBM SAN Volume Controller (SVC), shown in Figure 1-5, HP 3PAR Peer Persistence, and EMC VPLEX, shown in Figure 1-6 (formally Yotta-Yotta technology), correctly come to mind when the phrase storage virtualization is used. In the VMware context, we understand storage virtualization as an intrinsic part of the SDDC, and we include advanced VMware storage capabilities such as virtual storage-area network (vSAN), the soon-to-be released virtual volumes (vVols), and ­vFlashReadCache (vFRC) among others as the essential components of the storage virtualization paradigm.

Figure 1-5

Figure 1-5 IBM SVC stretch cluster

Figure 1-6

Figure 1-6 EMC VPLEX stretch cluster

For more information on HP 3PAR, refer to hp.com or to the link below.

www8.hp.com/us/en/products/storage-software/product-detail.html?oid=5335710#!tab=features.

Together, the tried-and-true virtualization of the server through vSphere when combined with the more recent ideas of network and storage virtualization constitute the SDDC VMware style. And throughout this book, we refer to each component of this evolutionary leap in datacenter design. Paul Maritz, the former CEO of VMware and present CEO of Pivotal Solutions, once referred to VMware as “the magic pixie dust that was changing the world.” We agree, although we constantly have to point out to customers that despite the magical illusion presented by VMware products, the laws of physics still apply, and so each implementation must consider the limits of those pesky rules of Newton and Einstein other­wise risk disappointment. However, it is not overheated rhetoric to state that VMware is changing the world. For example, it would be difficult to find a single corporate entity that has facilitated more reduction in power consumption throughout its customer base. More importantly, the name VMware itself implies a tectonic-like shift in the world of technology. Computing started with hardware, and then intrepid minds developed software to effectively use that hardware; but to comprehensively tie software to hardware, the world needed VMware.

For more information on EMC VPLEX stretch clusters, refer to the link below.

www.emc.com/storage/vplex/vplex.htm.

  • + Share This
  • 🔖 Save To Your Account