view README @ 2674:611881eac889

bitkeeper revision 1.1159.1.230 (416d965arlQEDDF33KaiMFJy_o-s4g)

Better random MAC addresses.
date Wed Oct 13 20:55:54 2004 +0000 (2004-10-13)
parents 11be1dfb262b
children 7266d3bd3b1f 4b524192e62b b6703e4c6e16
line source
1 ###############################
2 __ __ ____ ___
3 \ \/ /___ _ __ |___ \ / _ \
4 \ // _ \ '_ \ __) || | | |
5 / \ __/ | | | / __/ | |_| |
6 /_/\_\___|_| |_| |_____(_)___/
8 ###############################
10 University of Cambridge Computer Laboratory
11 28 Aug 2004
15 About the Xen Virtual Machine Monitor
16 =====================================
18 "Xen" is a Virtual Machine Monitor (VMM) originally developed by the
19 Systems Research Group of the University of Cambridge Computer
20 Laboratory, as part of the UK-EPSRC funded XenoServers project.
22 The XenoServers project aims to provide a "public infrastructure for
23 global distributed computing", and Xen plays a key part in that,
24 allowing us to efficiently partition a single machine to enable
25 multiple independent clients to run their operating systems and
26 applications in an environment providing protection, resource
27 isolation and accounting. The project web page contains further
28 information along with pointers to papers and technical reports:
31 Xen has since grown into a project in its own right, enabling us to
32 investigate interesting research issues regarding the best techniques
33 for virtualizing resources such as the CPU, memory, disk and network.
34 The project has been bolstered by support from Intel Research
35 Cambridge, and HP Labs, who are now working closely with us. We're
36 also in receipt of support from Microsoft Research Cambridge to port
37 Windows XP to run on Xen.
39 Xen enables multiple operating system images to execute concurrently
40 on the same hardware with very low performance overhead --- much lower
41 than commercial offerings for the same x86 platform.
43 This is achieved by requiring OSs to be specifically ported to run on
44 Xen, rather than allowing unmodified OS images to be used. Crucially,
45 only the OS needs to be changed -- all of the user-level application
46 binaries, libraries etc can run unmodified. Hence the modified OS
47 kernel can typically just be dropped into any existing OS distribution
48 or installation.
50 Xen currently runs on the x86 architecture, but could in principle be
51 ported to others. In fact, it would have been rather easier to write
52 Xen for pretty much any other architecture as x86 is particularly
53 tricky to handle. A good description of Xen's design, implementation
54 and performance is contained in our October 2003 SOSP paper, available
55 at
56 [update: work to port Xen to x86_64 and IA64 is underway]
58 Five different operating systems have been ported to run on Xen:
59 Linux 2.4/2.6, Windows XP, NetBSD, FreeBSD and Plan 9.
61 The Linux 2.4 port (currently Linux 2.4.26) works very well -- we
62 regularly use it to host complex applications such as PostgreSQL,
63 Apache, BK servers etc. It runs every user-space applications we've
64 tried. We refer to our version of Linux ported to run on Xen as
65 "XenLinux", although really it's just standard Linux ported to a new
66 virtual CPU architecture that we call xen-x86.
68 NetBSD has been ported to Xen by Christian Limpach, and will hopefully
69 soon become part of the standard release. Work on a FreeBSD port has
70 been started by Kip Macy, and we hope to see this complete for the 2.0
71 release. Ron Minnich has been working on Plan 9.
73 The Windows XP port is nearly finished. It's running user space
74 applications and is generally in pretty good shape thanks to some hard
75 work by a team over the summer. Of course, there are issues with
76 releasing this code to others. We should be able to release the
77 source and binaries to anyone that has signed the Microsoft academic
78 source license, which these days has very reasonable terms. We are in
79 discussions with Microsoft about the possibility of being able to make
80 binary releases to a larger user community. Obviously, there are
81 issues with product activation in this environment which need to be
82 thought through.
84 So, for the moment, you only get to run Linux 2.4/2.6 and NetBSD on
85 Xen, but we hope this will change before too long. Even running
86 multiple copies of the same OS can be very useful, as it provides a
87 means of containing faults to one OS image, and also for providing
88 performance isolation between the various OS, enabling you to either
89 restrict, or reserve resources for, particular VM instances.
91 It's also useful for development -- each version of Linux can have
92 different patches applied, enabling different kernels to be tried
93 out. For example, the "vservers" patch used by PlanetLab applies
94 cleanly to our ported version of Linux.
96 We've successfully booted over 128 copies of Linux on the same machine
97 (a dual CPU hyperthreaded Xeon box) but we imagine that it would be
98 more normal to use some smaller number, perhaps 10-20.
100 A common question is "how many virtual machines can I run on hardware
101 xyz?". The answer is very application dependent, but the rule of thumb
102 is that you should expect to be able to run the same workload under
103 multiple guest OSes that you could run under a single Linux instance,
104 with an additional overhead of a few MB per OS instance.
106 One key feature in this new release of Xen is `live migration'. This
107 enables virtual machines instances to be dynamically moved between
108 physical Xen machines, with typical downtimes of just a few tens of
109 milliseconds. This is really useful for admins that want to take a
110 node down for maintenance, or to load balance a large number of
111 virtual machines across a cluster.
115 Hardware support
116 ================
118 Xen is intended to be run on server-class machines, and the current
119 list of supported hardware very much reflects this, avoiding the need
120 for us to write drivers for "legacy" hardware. It is likely that some
121 desktop chipsets will fail to work properly with the default Xen
122 configuration: specifying 'noacpi' or 'ignorebiostables' when booting
123 Xen may help in these cases.
125 Xen requires a "P6" or newer processor (e.g. Pentium Pro, Celeron,
126 Pentium II, Pentium III, Pentium IV, Xeon, AMD Athlon, AMD Duron).
127 Multiprocessor machines are supported, and we also have basic support
128 for HyperThreading (SMT), although this remains a topic for ongoing
129 research. We're also working on an AMD x86_64 port (though Xen should
130 run on Opterons in 32-bit mode just fine).
132 Xen can currently use up to 4GB of memory. It's possible for x86
133 machines to address more than that (64GB), but it requires using a
134 different page table format (3-level rather than 2-level) that we
135 currently don't support. Adding 3-level PAE support wouldn't be
136 difficult, but we'd also need to add support to all the guest
137 OSs. Volunteers welcome!
139 In contrast to previous Xen versions, in Xen 2.0 device drivers run
140 within a privileged guest OS rather than within Xen itself. This means
141 that we should be compatible with the full set of device hardware
142 supported by Linux. The default XenLinux build contains support for
143 relatively modern server-class network and disk hardware, but you can
144 add suppport for other hardware by configuring your XenLinux kernel in
145 the normal way (e.g. "make xconfig").
148 Building Xen and XenLinux
149 =========================
151 The public master BK repository for the 2.0 release lives at:
152 bk://
154 To fetch a local copy, install the BitKeeper tools, then run:
155 'bk clone bk://'
157 You can do a complete build of Xen, the control tools, and the
158 XenLinux kernel images with "make world". This can take 10 minutes
159 even on a fast machine. If you're on an SMP machine you may wish to
160 give the '-j4' argument to make to get a parallel build. All of the
161 files that are built are placed under the ./install directory. You
162 can then install everything to the standard system directories
163 (e.g. /boot, /usr/bin, /usr/lib/python/ etc) by typing "make install".
165 Take a look in install/boot/:
166 install/boot/xen.gz The Xen 'kernel' (formerly image.gz)
167 install/boot/vmlinuz-2.4.27-xen0 Domain 0 XenLinux kernel (xenolinux.gz)
168 install/boot/vmlinuz-2.4.27-xenU Unprivileged XenLinux kernel
170 The difference between the two Linux kernels that are built is
171 due to the configuration file used for each. The "U" suffixed
172 unprivileged version doesn't contain any of the physical hardware
173 device drivers, so is 30% smaller and hence may be preferred for
174 your non-privileged domains.
176 The install/boot directory will also contain the config files
177 used for building the XenLinux kernels, and also versions of Xen
178 and XenLinux kernels that contain debug symbols (xen-syms and
179 vmlinux-syms-2.4.27-xen0) which are essential for interpreting crash
180 dumps.
182 Inspect the Makefile if you want to see what goes on during a
183 build. Building Xen and the tools is straightforward, but XenLinux is
184 more complicated. The makefile needs a 'pristine' linux kernel tree
185 which it will then add the Xen architecture files to. You can tell the
186 makefile the location of the appropriate linux compressed tar file by
187 setting the LINUX_SRC environment variable
188 (e.g. "LINUX_SRC=/tmp/linux-2.4.27.tar.gz make world") or by placing
189 the tar file somewhere in the search path of LINUX_SRC_PATH which
190 defaults to ".:..". If the makefile can't find a suitable kernel tar
191 file it attempts to download it from, but this won't work
192 if you're behind a firewall.
194 After untaring the pristine kernel tree, the makefile uses the
195 'mkbuildtree' script to add the Xen patches the kernel. "make world"
196 then build two different XenLinux images, one with a "-xen0" extension
197 which contains hardware device drivers and is intended to be used in
198 the first virtual machine ("domain 0"), and one with a "-xenU"
199 extension that just contains virtual-device drivers. The latter can be
200 used for all non hardware privileged domains, and is substantially
201 smaller than the other kernel with its selection of hardware drivers.
203 If you don't want to use bitkeeper to download the source, you can
204 download prebuilt binaries and src tar balls from the project
205 downloads page:
207 Using the domain control tools
208 ==============================
210 Before starting domains you'll need to start the node management
211 daemon: "xend start".
212 The primary tool for starting and controlling domains is "xm".
213 "xm help <cmd>" will tell you how to use it.
215 README.CD contains some example invocations.
217 Further documentation is in docs/ (e.g., docs/Xen-HOWTO), and also in