More Related Content
What's hot
What's hot (20)
Viewers also liked
Viewers also liked (17)
Similar to Hypervisors and Virtualization - VMware, Hyper-V, XenServer, and KVM
Similar to Hypervisors and Virtualization - VMware, Hyper-V, XenServer, and KVM (20)
Recently uploaded
Recently uploaded (20)
Hypervisors and Virtualization - VMware, Hyper-V, XenServer, and KVM
- 1. Hypervisors and Virtualization VMware, Hyper-V, XenServer, KVM Maninder Singh Vincent Chu
- 2. Overview • Introduction to Virtualization and Hypervisors • Origins of Hypervisors • Hypervisors and its Classification • Existing Solutions/Products • Advantages and Disadvantages • State of the Art • Future of Hypervisors and Virtualization
- 3. What is Virtualization? • Broadly, the separation of a resource or request for a service from the underlying physical delivery of that service • Concept in which access to a single underlying piece of hardware is coordinated so that multiple guest operating systems can share that single piece of hardware, with no guest operating system being aware that it is actually sharing anything at all • Guest Operating System is the OS that is hosted by the virtualization software on the Host Operating System
- 4. What is Virtualization? • Before using Virtualization, we had: • Single OS per machine • Software and hardware tightly coupled • Underutilized resources (idle time) • Inflexibility • Virtualization gives you: • Hardware independence of operation system and applications • Ability to encapsulate OS and applications in to virtual machines • Ability to provision virtual machines to any system
- 6. How do we get Virtualization? • Hypervisors: The approach to virtualization • Hypervisor also known as the Virtual Machine Monitor • Software that allows multiple operating systems to share a single hardware host • Emulates hardware resources to guest operating systems • Each operating system appears to have host’s processor, memory, and other resources all to itself • In reality, hypervisor controls the resources and allocates them as needed by each guest OS in a synchronized manner
- 7. Overview • Introduction to Virtualization and Hypervisors • Origins of Hypervisors • Hypervisors and its Classification • Existing Solutions/Products • Advantages and Disadvantages • State of the Art • Future of Hypervisors and Virtualization
- 8. Origins of Hypervisors • IBM first developed CP/CMS operating system in 1967, an attempt to build time-sharing systems for mainframe systems • In 1972, IBM’s zSeries line featured Virtualization • Early acceptance and rapid development by developers all over • In 1985, IBM introduced the PR/SM hypervisor to manage logical partitions • Other companies, Sun Microsystems, HP, and SGI joined the race and started selling virtualized software around 2000
- 9. Overview • Introduction to Virtualization and Hypervisors • Origins of Hypervisors • Hypervisors and its Classification • Existing Solutions/Products • Advantages and Disadvantages • State of the Art • Future of Hypervisors and Virtualization
- 10. Classification of Hypervisors There are two types of hypervisor based on architecture: • Type 1 Hypervisor • also known as native or bare metal hypervisor • Type 2 Hypervisor • also known as hosted hypervisor
- 11. Classification of Hypervisors Type 1 Hypervisor • Runs directly on the host’s hardware to manage guest operating systems • Does not require any base server operating system • Direct access to hardware resources • Better performance, scalability, and stability • However, hardware support is limited
- 12. Classification of Hypervisors Type 2 Hypervisor • Hosted on the main operating system • Basically a software installed on an OS • Hypervisor asks OS to make hardware calls • Better compatibility with hardware • Increased overhead affects performance
- 13. Classification of Type 1 VMMs Type 1 Hypervisors can be further classification into two main ways to architect the hypervisor solutions: • Monolithic • Hosts the hypervisor/VMM in a single layer that also includes most of the required components, such as the kernel, device drivers, and the I/O stack • Microkernelized • Uses a very thin, specialized hypervisor that only performs the core tasks of ensuring partition isolation and memory management. This layer does not include the I/O stack or device drivers. • Virtualization stack and hardware-specific device drivers are located in a specialized partition called the parent partition.
- 14. Classification of Type 1 VMMs Hardware Hypervisor VM 1 (Admin) VM 2 VM 3 Drivers Monolithic Hypervisor Hardware Hypervisor VM 1 (“Parent”) VM 2 (“Child”) VM 3 (“Child”) Drivers Drivers Drivers Virtualization Stack Microkernelized Hypervisor
- 15. Virtualization Techniques There are multiple approaches to running the guest operating system. These include: • Full Virtualization • Paravirtualization • Also known as OS assisted virtualization • Hardware-assisted virtualization • Also known as accelerated virtualization, hardware virtual machine (HVM)
- 16. Virtualization Techniques Full Virtualization • Completely abstracted from the underlying hardware by virtualization layer • Guest OS unaware that it is a guest • Hypervisor translates all OS calls on-the-fly • No hardware assistance or modification; flexibility
- 17. Virtualization Techniques Paravirtualization • An efficient and lightweight virtualization technique • The hypervisor provides an API and the Guest OS calls that API, requiring OS modifications. • Does not require virtualization extensions from the host CPU • Guests and control domains require kernel support and drivers • Enable near-native performance • paravirt-ops code supported by Linux kernel as of version 2.6.23
- 18. Virtualization Techniques Hardware-assisted virtualization • Enables efficient full virtualization using help from hardware capabilities, primarily from the host processors • Unmodified guest OS; no API calls made • Hypervisor traps sensitive calls • Added to x86 processors (Intel VT-x or AMD-V) in 2006
- 19. Virtualization Techniques FULL VIRTUALIZATION PARAVIRTUALIZATION HARDWARE ASSISTED Technique Direct execution Hypercalls Exit to Root Mode on privileged instructions Guest OS modification Unmodified guest OS; excellent compatibility Guest OS codified to issue hypercalls; poor compatibility Unmodified guest OS; excellent compatibility Performance Good Better in certain cases Fair; certain lags in binary translation but will improve over time Used By VMware, Microsoft, KVM VMware, Xen VMware, Xen, Microsoft, Parallels Guest OS hypervisor independent? Yes XenLinux runs only on Xen Hypervisor Yes
- 20. Overview • Introduction to Virtualization and Hypervisors • Origins of Hypervisors • Hypervisors and its Classification • Existing Solutions/Products • Advantages and Disadvantages • State of the Art • Future of Hypervisors and Virtualization
- 21. Leading Hypervisors • Type 1 • VMware ESX and ESXi • Microsoft Hyper-V • Citrix Systems XenServer • Type 2 • VMware Workstation • Oracle VM VirtualBox • Microsoft Virtual PC • Parallels Desktop
- 22. VMware ESX and ESXi • Enterprise virtualization platform offered by VMware • ESX runs on bare metal, without an OS • Monolithic architecture • Includes own Linux kernel, which is started first and then loads VMware’s vmkernel component • ESXi, an upgrade from ESX, does not contain the Linux kernel. Directly loads from vmkernel • Vmkernel has three interfaces: • Hardware • Guest systems • Service console (Linux based console operating system)
- 24. VMware ESXi • ESXi is a smaller footprint version of ESX • Does not include the ESX Service Console • All VMware related agents and 3rd party agents run directly on vmkernel • Ultra-thin architecture, highly reliable, small code-base • More secure as there is less patching • Uses Direct Console User Interface (DCUI) for management
- 26. ESX and ESXi comparison CAPABILITY ESX ESXI Service Console Present Removed Troubleshooting performed via Service Console ESXi Shell Secure Syslog Not Supported Supported Management Network Service Console Interface VMKernel Interface Hardware Monitoring 3 rd Party agents installed in Service console Via CIM Providers Software patches and updates Needed as similar to Linux operation system Few pacthes because of small footprint and more secure vSphere web Access Only experimental Full managenet capability via vSPhere web client Locked Down Mode Not present Present . Lockdown mode prevents remote users to login to the host Rapid deployment via Auto Deploy Not supported Supported Custom Image creation Not supported Supported VMkernel Network Used for vMotion, Fault Tolarance, Stoarge Connectivity Management Network , vMotion, Fault Tolerance, Storage Connectivity, ISCSI port binding
- 27. Microsoft Hyper-V • Introduced in 2008, it is Microsoft’s Type 1 Hypervisor for x86-64 systems • Has a microkernel design • Hyper-V comes in two variants: • Hyper-V Server: A stand-alone version • Variant of core installation of Windows Server 2008 • Includes full Hyper-V functionality • Mostly CLI for configuration but MMC can be installed for administration • An installable role version for Windows Server • Installed as a role on Windows Server 2008/2012 • Configuration and administration done by Remote Desktop or management consoles for much easier graphical control
- 28. Microsoft Hyper-V Architecture • Implements virtual machines in terms of partitions • Partition is a logical unit of isolation in which an OS is running • Virtualization layer runs in parent partition giving direct access to hardware resources • Parent partition creates child partitions for guest OSs • Compared to VMware ESX/ESXi, all hardware calls go through root or parent partition • Access to hardware is controlled via root partition though VMBus. This is indirect driver model. Paravirtualized. • Hyper-V installs on hardware and places original OS in the root partition
- 30. Citrix Systems XenServer • An open-source Type-1 or bare-metal hypervisor • Originated as a research project at the University of Cambridge in 2003 • Commercialize by XenSource Inc. • Developed and made available as open-source software under GNU General Public Licence (GPL), version 2 • Acquired by Citrix in 2007 • Commercial versions: Citrix XenServer, Oracle VM • Small footprint and interface; uses a microkernel design • Supports Paravirtualization • Supports IA-32, x86-64 and ARM instruction sets.
- 31. Citrix Systems XenServer Architecture • Guest types: The hypervisor can run fully virtualized guests, or paravirtualized guests. • A running instance of a virtual machine is called a domain or guest. • Employs a special domain called domain 0 which contains: • Drivers for the hardware • Toolstack to control VMs. • A control stack to: • manage virtual machine creation, • destruction, and • configuration.
- 32. Citrix Systems XenServer Architecture
- 33. Kernel-based Virtual Machine (KVM) • Initially developed by Qumranet later acquired by Redhat in 2008 • Full virtualized solution with a small code base for Linux on x86 hardware with virtualization extensions (Intel VT or AMD-V) • Designed as small, light kernel module to leverage the facilities provided by hardware support of virtualization • Originally supported x86 processors but now has been ported to IBM S/390, Intel IA-64 • Free, open source virtualization architecture. Kernel component comes standard in vanilla Linux (2.6.20)
- 34. Kernel-based Virtual Machine (KVM) • KVM hypervisor Type 1 or Type 2: Still up for discussion • Implemented as a kernel module, allowing Linux to become a hypervisor simply by loading it • Device appears in /dev/kvm. Allows control by ioctl() system calls to create new VMs, assign memory, etc • Hardware emulation or platform virtualization controlled by QEMU-kvm
- 35. Overview • Introduction to Virtualization and Hypervisors • Origins of Hypervisors • Hypervisors and its Classification • Existing Solutions/Products • Advantages and Disadvantages • State of the Art • Future of Hypervisors and Virtualization
- 36. Advantages • Consolidation of hardware resources • Ease of administration • Centralization of resources • Significant cost savings • Less hardware • Energy and power savings • Fault tolerance through clustering • Ease of deployment and management • Better use from hardware • Virtualization enables higher utilization rates of hardware because each server supports enough virtual machines to increase its utilization from the typical 15% to as much as 80%.
- 37. Advantages (continued) • Scalability • Additional processing power, network bandwidth, and storage capacity can be accomplished quickly and easily by apportioning additional available resources from the host to the guest VM. • High availability • Mitigating downtime of VMs by dynamically allocating more resources to a guest as needed, migrating VMs between different hosts with zero downtime, fault tolerance, and backing up snapshots of the running systems state. • Software installation made easy • Software vendors can deliver their products preinstalled in VMs (also known as virtual appliances), much of the traditional installation and configuration work associated with software will disappear.
- 38. Disadvantages • Due to the demands of server consolidation, VMs tend to consume more CPU and memory, and require greater disk I/O bandwidth than physical servers with comparable computing loads.
- 39. Overview • Introduction to Virtualization and Hypervisors • Origins of Hypervisors • Hypervisors and its Classification • Existing Solutions/Products • Advantages and Disadvantages • State of the Art • Future of Hypervisors and Virtualization
- 40. State of the Art • Virtual Private Server (VPS): • A virtual machine sold as a service by an Internet hosting service. • Amazon Elastic Compute Cloud (EC2) • Amazon.com’s cloud computing platform, Amazon Web Services (AWS) • Uses Xen as its hypervisor • Microsoft Azure • Microsoft's cloud application platform. • Powered by Microsoft Hyper-V • Google Compute Engine (GCE) • Infrastructure as a Service (IaaS) component of Google Cloud Platform • Uses KVM as the hypervisor • Rackspace Cloud uses VMware • Used in enterprise IT
- 41. Overview • Introduction to Virtualization and Hypervisors • Origins of Hypervisors • Hypervisors and its Classification • Existing Solutions/Products • Advantages and Disadvantages • State of the Art • Future of Hypervisors and Virtualization
- 42. The Future • Single-purpose Appliances • Trending towards each VM running and being specialized for a single application, instead of a full operating system. • Each instance of each application runs in an isolated VM as though it has the entire machine to itself • Unikernel or exokernel, virtual library operating systems, are hypervisor-operating system hybrids that are currently being developed that make it possible to perform as much compile-time work as possible to eliminate unnecessary features from the final VM. • Thus, creating a specialize VM for a specific application • Reducing the amount of active code running in the VM, and • Reducing the attack surface for remote code execution exploits and serious data leaks; better security.
- 43. The Future • Single-purpose Appliances: • Mirage OS
- 44. Any Questions?
- 45. Thanks