Motivation

Three fundamental abstractions are necessary to describe the operation of

a computing systems:
(1) interpreters/processors, (2) memory, (3) communications links
As the scale of a system and the size of its users grows, it becomes
very challenging to manage its recourses (see three points above)

Resource management issues:

 provision for peak demands  overprovisioning
 heterogeneity of hardware and software
 machine failures
Virtualization is a basic enabler of Cloud Computing, it simplifies
the management of physical resources for the three abstractions

For example, the state of a virtual machine (VM) running under a virtual machine
monitor (VMM) can de saved and migrated to another server to balance the load
For example, virtualization allows users to operate in environments they are familiar
with, rather than forcing them to specific ones

1
 Motivation (cont’d)
 “Virtualization, in computing, refers to the act of creating a
virtual (rather than actual) version of something,
including but not limited to a virtual computer hardware
platform, operating system (OS), storage device, or computer
network resources.” from Wikipedia

 Virtualization abstracts the underlying resources; simplifies

their use; isolates users from one another; and supports
replication which increases the elasticity of a system

2
 Motivation (cont’d)

 Cloud resource virtualization is important for:

 Performance isolation
 as we can dynamically assign and account for resources across
different applications
 System security:
 as it allows isolation of services running on the same hardware
 Performance and reliability:
 as it allows applications to migrate from one platform to another
 The development and management of services offered by a provider

3
Virtualization

 Virtualization simulates the interface to a physical object by:

 Multiplexing: creates multiple virtual objects from one instance of a
physical object. Many virtual objects to one physical. Example - a
processor is multiplexed among a number of processes or threads.
 Aggregation: creates one virtual object from multiple physical
objects. One virtual object to many physical objects. Example - a
number of physical disks are aggregated into a RAID disk.
 Emulation: constructs a virtual object of a certain type from a
different type of a physical object. Example - a physical disk
emulates a Random Access Memory (RAM).
 Multiplexing and emulation. Examples - virtual memory with paging
multiplexes real memory and disk; a virtual address emulates a real
address.

4
Layering and Virtualization

 Layering – a common approach to manage system complexity:

 Simplifies the description of the subsystems; each subsystem is abstracted
through its interfaces with the other subsystems
 Minimises the interactions among the subsystems of a complex system
 With layering we are able to design, implement, and modify the individual
subsystems independently
 Layering in a computer system:
 Hardware
 Software
 Operating system
 Libraries
 Applications

5
Layering and Interfaces

Application Programming Interface (API), Application Binary Interface (ABI), and

Instruction Set Architecture (ISA). An application uses library functions (A1),
makes system calls (A2), and executes machine instructions (A3) (from book) 6
Interfaces

 Instruction Set Architecture (ISA) – at the boundary between

hardware and software.

 Application Binary Interface (ABI) – allows the ensemble consisting of

the application and the library modules to access the hardware; the ABI
does not include privileged system instructions, instead it invokes system
calls.

 Application Program Interface (API) - defines the set of instructions

the hardware was designed to execute and gives the application access to
the ISA; it includes high-level language (HLL) library calls which often
invoke system calls

7
Code portability

 Binaries created by a compiler for a specific ISA and a specific operating

systems are not portable

 It is possible, though, to compile a HLL program for a virtual machine

(VM) environment where portable code is produced and distributed and
then converted by binary translators to the ISA of the host system

 A dynamic binary translation converts blocks of guest instructions

from the portable code to the host instruction and leads to a significant
performance improvement, as such blocks are cached and reused

8
HLL Language Translations

9
History of Virtualization
(from “Modern Operating Systems” 4th Edition, p474 by Tanenbaum and Bos)

 1960’s, IBM: CP/CMS control program: a virtual machine operating

system for the IBM System/360 Model 67
 2000, IBM: z-series with 64-bit virtual address spaces and backward
compatible with the System/360
 1974: Popek and Golberg from UCLA published “Formal Requirements
for Virtualizable Third Generation Architectures” where they listed the
conditions a computer architecture should satisfy to support
virtualization efficiently. The popular x86 architecture that originated in
the 1970s did not support these requirements for decades.
 1990’s, Stanford researchers, VMware: Researchers developed a
new hypervisor and founded VMware, the biggest virtualization company
of today’s. First virtualization solution was is 1999 for x86.
 Today many virtualization solutions: Xen from Cambridge, KVM, Hyper-V,
…
 IBM was the first to produce and sell virtualization for the mainframe.
But, VMware popularised virtualization for the masses.
10
Virtual Machine Monitor (VMM / Hypervisor)
 A virtual machine monitor (VMM/hypervisor) partitions the
resources of computer system into one or more virtual machines
(VMs). Allows several operating systems to run concurrently on a
single hardware platform
 A VM is an execution environment that runs an OS
 VM – an isolated environment that appears to be a whole computer,
but actually only has access to a portion of the computer resources
 A VMM allows:
 Multiple services to share
the same platform
 Live migration - the movement
of a server from one platform to another
 System modification while maintaining
backward compatibility with the original system
 Enforces isolation among the systems, thus security
 A guest operating system is an OS that runs in a VM under the 11
control of the VMM.
VMM Virtualizes the CPU and the Memory
 A VMM (also hypervisor) (howto):
 Traps the privileged instructions executed by a guest OS and
enforces the correctness and safety of the operation
 Traps interrupts and dispatches them to the individual guest
operating systems
 Controls the virtual memory management
 Maintains a shadow page table for each guest OS and replicates any
modification made by the guest OS in its own shadow page table.
This shadow page table points to the actual page frame and it is
used by the Memory Management Unit (MMU) for dynamic address
translation.
 Monitors the system performance and takes corrective actions to
avoid performance degradation. For example, the VMM may swap
out a VM to avoid thrashing.

12
Type 1 and 2 Hypervisors
Type 1 Hypervisor Type 2 Hypervisor

 Taxonomy of VMMs:
1. Type 1 Hypervisor (bare metal, native): supports multiple virtual machines
and runs directly on the hardware (e.g., VMware ESX , Xen, Denali)
2. Type 2 Hypervisor (hosted) VM - runs under a host operating system (e.g.,
user-mode Linux)

13
Examples of Hypervisors

14
Performance and Security Isolation
 The run-time behavior of an application is affected by other applications
running concurrently on the same platform and competing for CPU
cycles, cache, main memory, disk and network access. Thus, it is
difficult to predict the completion time!

 Performance isolation - a critical condition for QoS guarantees in shared

computing environments

 A VMM is a much simpler and better specified system than a traditional

operating system. Example - Xen has approximately 60,000 lines of
code; Denali has only about half: 30,000

 The security vulnerability of VMMs is considerably reduced as the

systems expose a much smaller number of privileged functions. For
example, Xen VMM has 28 hypercalls while Linux has 100s of system
calls
15
Conditions for Efficient Virtualization (from Popek and Goldberg) :

 Conditions for efficient virtualization (from Popek and Goldberg) :

1. A program running under the VMM should exhibit a behavior essentially
identical to that demonstrated when running on an equivalent machine
directly.
2. The VMM should be in complete control of the virtualized resources.
3. A statistically significant fraction of machine instructions must be executed
without the intervention of the VMM. (Why?)

16
Dual-Mode Operation (recap)
 Dual-mode operation allows OS to protect itself and other
system components
 User mode and kernel mode
 Mode bit provided by hardware
 Ability to distinguish when system is running user or kernel code
 Some instructions are privileged, only executable in kernel mode
 System call changes mode to kernel, return resets it to user

17
User-mode vs Kernel-mode (recap)

 Kernel-code (in particular, interrupt handlers) runs in kernel

mode
 the hardware allows all machine instructions to be executed and
allows unrestricted access to memory and I/O ports
 Everything else runs in user mode
 The OS relies very heavily on this hardware-enforced
protection mechanism

18
 Challenges of x86 CPU Virtualization
 Four layers of privilege execution  rings
 User applications run in ring 3
 OS runs in ring 0
 In which ring should the VMM run?
 In ring 0, then, same privileges as an OS  wrong
 In rings 1,2,3, then OS has higher privileges  wrong
 Move the OS to ring 1 and the VMM in ring 0  OK
 Three classes of machine instructions:
1. privileged instructions can be executed
in kernel mode. When attempted to be
executed in user mode, they cause a trap
and so executed in kernel mode.
2. nonprivileged instructions the ones that can be executed in user mode
3. sensitive instructions can be executed in either kernel or user but they
behave differently. Sensitive instructions require special precautions at
execution time.
19
4. sensitive and nonprivileged instructions are hard to virtualize
Techniques for Virtualizing CPU on x86

1. Full virtualization with binary translation

2. OS-assisted Virtualization or Paravirtualization
3. Hardware assisted virtualization

20
 Techniques for Virtualizing CPU on x86
Full virtualization – a guest OS can run unchanged under the VMM as if
it was running directly on the hardware platform. Each VM runs an exact
copy of the actual hardware.
 Binary translation rewrites parts of the code on the fly to replace sensitive
but not privileged instructions with safe code to emulate the original instruction
 “The hypervisor translates all operating system instructions on the fly and
caches the results for future use, while user level instructions run unmodified at
native speed.” (from VMware paper)
 Examples: VMware, Microsoft Virtual Server
 Advantages:
 No hardware assistance,
 No modifications of the guest OS
 Isolation, Security
 Disadvantages:
 Speed of execution

21
 Techniques for Virtualizing CPU on x86
Paravirtualization – “involves modifying the OS kernel to replace non-
virtualizable instructions with hypercalls that communicate directly with the
virtualization layer hypervisor. The hypervisor also provides hypercall
interfaces for other critical kernel operations such as memory
management, interrupt handling and time keeping. “ (from VMware paper)

 Advantage: faster execution, lower virtualization overhead

 Disadvantage: poor portability
 Examples: Xen, Denali

22
Full Virtualization and Paravirtualization

23
 Techniques for Virtualizing CPU on x86
 Hardware Assisted Virtualization – “a new CPU execution mode
feature that allows the VMM to run in a new root mode below ring 0. As
depicted in Figure 7, privileged and sensitive calls are set to
automatically trap to the hypervisor, removing the need for either
binary translation or paravirtualization“ (from VMware paper)

 Advantage: even faster execution

 Examples: Intel VT-x, Xen 3.x

24
VT-x, a Major Architectural Enhancement
 In 2005 Intel released two Pentium 4 models supporting VT-x.
 VT-x supports two modes of operations (Figure (a)):
1. VMX root - for VMM operations.
2. VMX non-root - support a VM.
 And a new data structure called the Virtual Machine Control Structure
including host-state and guest-state areas (Figure (b)).
 VM entry - the processor state is loaded from the guest-state of the VM
scheduled to run; then the control is transferred from VMM to the VM.
 VM exit - saves the processor state in the guest-state area of the running
VM; then it loads the processor state from the host-state area, finally
transfers control to the VMM.

25
Xen - a VMM based on Paravirtualization
 The goal of the Cambridge group - design a VMM capable of scaling to
about 100 VMs running standard applications and services without any
modifications to the Application Binary Interface (ABI). (2003, Computing
Laboratory, Cambridge University)
 Linux, Minix, NetBSD, FreeBSD and others can operate as
paravirtualized Xen guest OS running on x86, x86-64, Itanium, and
ARM architectures.
 Xen domain - ensemble of address spaces hosting a guest OS and
applications running under the guest OS. Runs on a virtual CPU.
 Dom0 - dedicated to execution of Xen control functions and
privileged instructions.
 DomU - a user domain.
 Applications make system calls using hypercalls processed by Xen;
privileged instructions issued by a guest OS are paravirtualized and
must be validated by Xen.

26
Xen

27
Dom0 Components
 XenStore – a Dom0 process.
 Supports a system-wide registry and naming service.
 Implemented as a hierarchical key-value storage.
 A watch function informs listeners of changes of the key in storage
they have subscribed to.
 Communicates with guest VMs via shared memory using Dom0
privileges.
 Toolstack - responsible for creating, destroying, and managing the
resources and privileges of VMs.
 To create a new VM, a user provides a configuration file describing
memory and CPU allocations and device configurations.
 Toolstack parses this file and writes this information in XenStore.
 Takes advantage of Dom0 privileges to map guest memory, to load
a kernel and virtual BIOS and to set up initial communication
channels with XenStore and with the virtual console when a new VM
is created. 28
Strategies for virtual memory management, CPU multiplexing,
and I/O devices

29
Linux Containers
 A Linux Container is a Linux process (or processes) that is a virtual
environment with its own process network space. (lightweight process
virtualization)
 Containers share portions of the host kernel
 Containers use:
 Namespaces: per-process isolation of OS resources (filesystem, network and user ids)
 Cgroups: resource management and accounting per process

 Examples for using containers:

 https://www.dotcloud.com/
 https://www.heroku.com/

30
opensource.com
Xen (old) Implementation on x86 Architecture

 Xen runs at privilege Level 0, the guest OS at Level 1, and applications

at Level 3.
 The x86 architecture does not support either the tagging of TLB entries
or the software management of the TLB. Thus, address space
switching, when the VMM activates a different OS, requires a complete
TLB flush; this has a negative impact on the performance.
 Solution - load Xen in a 64 MB segment at the top of each address
space and delegate the management of hardware page tables to the
guest OS with minimal intervention from Xen. This region is not
accessible or re-mappable by the guest OS.
 A guest OS must register with Xen a description table with the
addresses of exception handlers for validation.

31
Xen Abstractions for Networking and I/O
 Each domain has one or more Virtual Network Interfaces (VIFs) which
support the functionality of a network interface card. A VIF is attached
to a Virtual Firewall-Router (VFR).
 Split drivers have a front-end in the DomU and the back-end in Dom0;
the two communicate via a ring in shared memory.
 Ring - a circular queue of descriptors allocated by a domain and
accessible within Xen. Descriptors do not contain data, the data buffers
are allocated off-band by the guest OS.
 Two rings of buffer descriptors, one for packet sending and one for
packet receiving, are supported.
 To transmit a packet:
 a guest OS enqueues a buffer descriptor to the send ring,
 then Xen copies the descriptor and checks safety,
 copies only the packet header, not the payload, and
 executes the matching rules.
32
Xen I/O
Xen zero-copy semantics
for data transfer using I/O
rings.
(a) The communication
between a guest
domain and the driver
domain over an I/O
and an event channel;
NIC is the Network
Interface Controller.
(b) The circular ring of
buffers.

33
Xen 2.0

 Optimization of:
 Virtual interface - takes advantage of the capabilities of some
physical NICs, such as checksum offload.
 I/O channel - rather than copying a data buffer holding a packet,
each packet is allocated in a new page and then the physical page
containing the packet is re-mapped into the target domain.
 Virtual memory - takes advantage of the superpage and global page
mapping hardware on Pentium and Pentium Pro processors. A
superpage entry covers 1,024 pages of physical memory and the
address translation mechanism maps a set of contiguous pages to a
set of contiguous physical pages. This helps reduce the number of
TLB misses.

34
Xen Network Architecture
The original architecture The optimised architecture

35
Performance Measurements

A comparison of send and receive data rates for a native Linux

system, the Xen driver domain, an original Xen guest domain, and an
optimised Xen guest domain.

36
Performance Comparison of Virtual Machines

 Compare the performance of Xen and OpenVZwith, a standard

operating system, a plain vanilla Linux.
 The questions examined are:
 How the performance scales up with the load?
 What is the impact of a mix of applications?
 What are the implications of the load assignment on individual
servers?
 The main conclusions:
 The virtualization overhead of Xen is considerably higher than that
of OpenVZ and that this is due primarily to L2-cache misses.
 The performance degradation when the workload increases is also
noticeable for Xen.
 Hosting multiple tiers of the same application on the same server is
not an optimal solution.

37
The Darker Side of Virtualization

 In a layered structure, a defense mechanism at some layer can be

disabled by malware running at a layer below it.
 It is feasible to insert a rogue VMM, a Virtual-Machine Based Rootkit
(VMBR) between the physical hardware and an operating system.
 Rootkit - malware with a privileged access to a system.
 The VMBR can enable a separate malicious OS to run surreptitiously
and make this malicious OS invisible to the guest OS and to the
application running under it.
 Under the protection of the VMBR, the malicious OS could:
 observe the data, the events, or the state of the target system.
 run services, such as spam relays or distributed denial-of-service
attacks.
 interfere with the application.

38
The Darker Side of Virtualization (con’t)

The insertion of a Virtual-Machine Based Rootkit (VMBR) as the lowest

layer of the software stack running on the physical hardware; (a) below
an operating system; (b) below a legitimate virtual machine monitor. The
VMBR enables a malicious OS to run surreptitiously and makes it invisible
to the genuine or the guest OS and to the application.
39
Summary

 Virtualization (Chapter 5, Sections 5.1-5.8)

 Layering and virtualization.
 Virtual machine monitor.
 Virtual machine.
 x86 support for virtualization.
 Xen.

40