Last month, we looked at the first of the seven domains that are on the CompTIA Cloud+ certification entry-level exam (number CV0-001). This month, the focus turns to the second domain — Virtualization — and the five topic areas beneath it:
- Explain the differences between hypervisor types
- Install, configure, and manage virtual machines and devices
- Given a scenario, perform virtual resource migration
- Explain the benefits of virtualization in a cloud environment
- Compare and contrast virtual components used to construct a cloud environment
Virtualization is an important technology in cloud computing because it removes the barrier of needing one-to-one relationships between the physical computer and the operating system. Signifying that importance, these topics (some of which appear again in other domains) make up 19 percent of the exam questions — the second highest percentage of the seven domains. Being an entry-level exam, there is a heavy focus on definitions and knowledge as opposed to actual implementation. That said, each of the five topic areas are examined in order below.
In essence, virtualization means that there is not necessarily a one-to-one relationship between a physical server and a logical (or virtual) server. There could be one physical server that virtually hosts cloud servers for a dozen companies, or there could be several physical servers working together as one logical server. From the end user's side, they have no concept of whether they are interacting with a physical machine or a virtual machine: it is all handled behind the scenes. At the risk of redundancy, virtualization can be thought of as creating virtual (rather than actual) versions of something (a desktop, a server, etc.).
The purpose of going virtual, in almost every instance, is to save money. Providers can achieve economies of scale, because adding additional clients doesn't always require the purchase of additional hardware. Clients can pay only for the services they use and don't have to pay for hardware (or the utilities needed to keep the hardware cool). Developers and other end users can have multiple environments to use without needing to buy additional hardware as well.
To implement virtualization, there needs to be a hypervisor (also known as a virtual machine manager or VMM). The hypervisor allows multiple operating systems to share the same host, and it manages the physical resource allocation to those virtual OSs. There are two types of hypervisors: bare metal (Type I) and operating system dependent (Type II).
With Type I, the hypervisor runs independent of the operating system — booting up before the OS and it is basically the operating system for the physical machine. This setup is most commonly used for server-side virtualization, because the hypervisor itself typically has very low hardware requirements to support its own functions.
Type 1 is generally considered to have better performance than Type 2, simply because there is no host OS involved and the system is dedicated to supporting virtualization. Virtual OSs are run within the hypervisor, and the virtual (guest) OSs are completely independent of each other.
Type II is dependent on the operating system — it cannot boot until the OS is up AND it needs the OS running in order to stay up. Because of this, it is also called the "host OS" and it is commonly used in client-side virtualization environments where multiple OSs are managed on the client machine as opposed to on a server.
An example of this would be a Windows user who wants to run Linux at the same time as Windows. The user could install a hypervisor and then install Linux in the hypervisor and run both OSs concurrently and independently. The downsides of Type 2 are that the host OS consumes resources such as processor time and memory and a host OS failure means that the guest OSs fail as well.
For the exam, remember that when it comes to performance and scalability, Type I is superior to Type II.
When it comes to hypervisor-specific system requirements, bear in mind that running multiple OSs on one physical workstation can require more resources than running a single OS, so the system(s) should be well equipped (CPU, RAM, hard drive space, and network performance). This is especially true for systems running a Type 2 hypervisor, which sits on top of a host OS.
The host OS will need resources too, and it will compete with the VMs for those resources. Additionally, all hosts within a cluster need to be homogenous. Statically allocated IP addresses are recommended and it is important to have sufficient memory, lots of hard drive space, and systems that have current patches installed when issued.
Both proprietary and open source solutions are widely available. Some, like Xen, fork into both proprietary and open source solutions. VMware ESX is offered for free, but you pay for features. Xen is free and open source; ESX is free but not open source (proprietary). KVM is free and open source; Microsoft's Hyper-V is usually free, but not open source (proprietary).
Both consumer and enterprise solutions can be used. Consumer implementations can include embedded deployments, but consumer implementations should not use enterprise applications due to the excessive overhead. As a general rule, "workstation" implementations can be equated to "desktop" use and "cloud" can be equated to the "infrastructure" utilization. Note that Type I hypervisors are more likely to be used in enterprises and Type II by consumers.
Virtual Machines and Devices
The actual option choices for creating, importing, and exporting templates depend on the software being used, but most (Xen, VMware, etc.) have similar options:
- To export, you choose File, Export Template.
- To import, you choose File, Import.
Depending on the importing/exporting that are doing, you often have to agree to the terms of a EULA.
Know that each virtual desktop (often called a virtual desktop interface or VDI) will usually need full network access, and configuring the permissions for each can be time consuming to configure without templates. The virtual machine create a virtual NIC and allows you to manage the resources of that NIC appropriately.
Theoretically, the virtual NIC would not have to be connected to the physical NIC — an administrator could create an entire virtual network within the virtual environment where the virtual machines just talk to each other — but that is not normally practical in the real world.
In most situations, the virtual NIC will be connected to the physical NIC and configuring a virtual switch within the hypervisor normally does this. That virtual switch manages the traffic to and from the virtual NICs and logically attaches to the physical NIC. Because of this, network bandwidth is often the biggest bottleneck when running multiple virtual OSs
Guest Tools are helpers added after the VM/OS has been installed. With VMware, for example, install VMware Tools on workstation menu (this is available as an ISO file). A clone is a copy of an existing virtual machine. Changes made to a clone do not affect the parent virtual machine. Changes made to the parent virtual machine do not appear in a clone.
A snapshot is a point-in-time copy of the virtual machine. File-Level backups are incremental backups of virtual machines. An image backup is an online backup of the virtual machine(s). Virtual disk limits are based on the virtual machines used. The following table shows the maximums for VMware vSphere 5.1:
A VLAN makes it possible for VNIC to communicate with other network devices. The Virtual NIC needs an IP address, subnet mask, and default gateway values the same any physical NIC would.
Virtual Resource Migration
Before migrating to a virtual platform, it is important to plan carefully. Not every physical server is a perfect candidate for migration. You need to create a baseline. Try not to overprovision or underprovision. Know that migration will entail some downtime and plan for the least disruption possible. With online migration, the source computer stays up during migration. With offline migration, the source computer taken offline during migration.
Reasons for migrating can include performance issues, testing needs, upgrading existing systems, and better resource utilization. The three possible migration types are:
- Physical to Virtual (P2V) — from a physical to virtual
- Virtual to Virtual (V2V) — from one virtual to another
- Virtual to Physical (V2P) — from a virtual to a physical
Virtualization in a Cloud Environment
Virtualization simplifies the sharing of resources and it is possible to share almost any resource: the processor, the disk, network, memory, etc.
Elasticity offers the ability to scale up resources as needed and it has a number of other benefits that go along with it: the time to service — the mean time to implement — is quicker inside rather than outside the virtual model, resource pooling is possible as are multitenant models, it is scalable not only up but also down, and applications are both available and portable.
Using network and application isolation it is possible to increase security and control resources. When planning isolation, think of security, chargeback, etc. Infrastructure consolidation can range from SaaS to IaaS and allow multiple machines to run on the same host.
A virtual datacenter appears the same as a physical datacenter from an administration standpoint and features elasticity, scalability, etc. A big benefit of the virtual center is that it can employ a pay-as-you-go model.
There are a number of terms to know related to virtualization for this topic:
Virtual NIC: While software only, it allows interaction with other devices on network and has MAC/IP address, network configuration settings, etc.
Virtual HBA: Enables a single physical Fibre Channel HBA port to function as multiple logical ports, each with its own identity.
Virtual Router: Software only, but acts like hardware router.
Shared Memory: Virtual memory settings can be changed as needed. This can usually be configured as a static value or dynamically.
Virtual CPU: It is installed on the guest virtual machine and appears the same as a physical CPU. A vCPU is also known as a virtual processor.
Shared storage: Can be done on SAN, NAS, etc. Virtual machine sees only "physical disk."
Clustered storage: Using multiple devices can increase performance. Microsoft Clustering Services would be an alternative to this.
NPIV(N_port ID Virtualization): Multiple hosts share the same physical fibre channel port ID. This is used for High Availability with a SAN.
While not appearing in the CompTIA list, know that the word emulator is tossed around often. The word is often used synonymously with hypervisor but they aren't exactly the same. While the hypervisor can support multiple OSs, technically an emulator appears to work the same as one specific OS. It is helpful to keep this difference in mind.
Summing It Up
There are seven domains on the CompTIA Cloud+ certification exam (CV0-001) and this month we walked through the topics covered by the second one. Next month, the focus will move to the third domain, Infrastructure, and what you should know about it as you study for the exam.