VirtualBox

source: vbox/trunk/doc/manual/en_US/user_Networking.xml@ 47991

最後變更 在這個檔案從47991是 47956,由 vboxsync 提交於 11 年 前

doc/manual: minor formatting fix, the excessive newlines spilled over to the final result

檔案大小: 46.0 KB
 
1<?xml version="1.0" encoding="UTF-8"?>
2<!DOCTYPE chapter PUBLIC "-//OASIS//DTD DocBook XML V4.4//EN"
3"http://www.oasis-open.org/docbook/xml/4.4/docbookx.dtd">
4<chapter id="networkingdetails">
5 <title>Virtual networking</title>
6
7 <para>As briefly mentioned in <xref linkend="settings-network" />,
8 VirtualBox provides up to eight virtual PCI Ethernet cards for each virtual
9 machine. For each such card, you can individually select<orderedlist>
10 <listitem>
11 <para>the hardware that will be virtualized as well as</para>
12 </listitem>
13
14 <listitem>
15 <para>the virtualization mode that the virtual card will be operating
16 in with respect to your physical networking hardware on the
17 host.</para>
18 </listitem>
19 </orderedlist></para>
20
21 <para>Four of the network cards can be configured in the "Network" section
22 of the settings dialog in the graphical user interface of VirtualBox. You
23 can configure all eight network cards on the command line via VBoxManage
24 modifyvm; see <xref linkend="vboxmanage-modifyvm" />.</para>
25
26 <para>This chapter explains the various networking settings in more
27 detail.</para>
28
29 <sect1 id="nichardware">
30 <title>Virtual networking hardware</title>
31
32 <para>For each card, you can individually select what kind of
33 <emphasis>hardware</emphasis> will be presented to the virtual machine.
34 VirtualBox can virtualize the following six types of networking
35 hardware:<itemizedlist>
36 <listitem>
37 <para>AMD PCNet PCI II (Am79C970A);</para>
38 </listitem>
39
40 <listitem>
41 <para>AMD PCNet FAST III (Am79C973, the default);</para>
42 </listitem>
43
44 <listitem>
45 <para>Intel PRO/1000 MT Desktop (82540EM);</para>
46 </listitem>
47
48 <listitem>
49 <para>Intel PRO/1000 T Server (82543GC);</para>
50 </listitem>
51
52 <listitem>
53 <para>Intel PRO/1000 MT Server (82545EM);</para>
54 </listitem>
55
56 <listitem>
57 <para>Paravirtualized network adapter (virtio-net).</para>
58 </listitem>
59 </itemizedlist></para>
60
61 <para>The PCNet FAST III is the default because it is supported by nearly
62 all operating systems out of the box, as well as the GNU GRUB boot
63 manager. As an exception, the Intel PRO/1000 family adapters are chosen
64 for some guest operating system types that no longer ship with drivers for
65 the PCNet card, such as Windows Vista.</para>
66
67 <para>The Intel PRO/1000 MT Desktop type works with Windows Vista and
68 later versions. The T Server variant of the Intel PRO/1000 card is
69 recognized by Windows XP guests without additional driver installation.
70 The MT Server variant facilitates OVF imports from other platforms.</para>
71
72 <para>The <emphasis role="bold">"Paravirtualized network adapter
73 (virtio-net)"</emphasis> is special. If you select this, then VirtualBox
74 does <emphasis>not</emphasis> virtualize common networking hardware (that
75 is supported by common guest operating systems out of the box). Instead,
76 VirtualBox then expects a special software interface for virtualized
77 environments to be provided by the guest, thus avoiding the complexity of
78 emulating networking hardware and improving network performance. Starting
79 with version 3.1, VirtualBox provides support for the industry-standard
80 "virtio" networking drivers, which are part of the open-source KVM
81 project.</para>
82
83 <para>The "virtio" networking drivers are available for the following
84 guest operating systems:</para>
85
86 <para><itemizedlist>
87 <listitem>
88 <para>Linux kernels version 2.6.25 or later can be configured to
89 provide virtio support; some distributions also back-ported virtio
90 to older kernels.</para>
91 </listitem>
92
93 <listitem>
94 <para>For Windows 2000, XP and Vista, virtio drivers can be
95 downloaded and installed from the KVM project web page.<footnote>
96 <para><ulink
97 url="http://www.linux-kvm.org/page/WindowsGuestDrivers">http://www.linux-kvm.org/page/WindowsGuestDrivers</ulink>.</para>
98 </footnote></para>
99 </listitem>
100 </itemizedlist></para>
101
102 <para>VirtualBox also has limited support for so-called <emphasis
103 role="bold">jumbo frames</emphasis>, i.e. networking packets with more
104 than 1500 bytes of data, provided that you use the Intel card
105 virtualization and bridged networking. In other words, jumbo frames are
106 not supported with the AMD networking devices; in those cases, jumbo
107 packets will silently be dropped for both the transmit and the receive
108 direction. Guest operating systems trying to use this feature will observe
109 this as a packet loss, which may lead to unexpected application behavior
110 in the guest. This does not cause problems with guest operating systems in
111 their default configuration, as jumbo frames need to be explicitly
112 enabled.</para>
113 </sect1>
114
115 <sect1 id="networkingmodes">
116 <title>Introduction to networking modes</title>
117
118 <para>Each of the eight networking adapters can be separately configured
119 to operate in one of the following modes:<glosslist>
120 <glossentry>
121 <glossterm>Not attached</glossterm>
122
123 <glossdef>
124 <para>In this mode, VirtualBox reports to the guest that a network
125 card is present, but that there is no connection -- as if no
126 Ethernet cable was plugged into the card. This way it is possible
127 to "pull" the virtual Ethernet cable and disrupt the connection,
128 which can be useful to inform a guest operating system that no
129 network connection is available and enforce a
130 reconfiguration.</para>
131 </glossdef>
132 </glossentry>
133
134 <glossentry>
135 <glossterm>Network Address Translation (NAT)</glossterm>
136
137 <glossdef>
138 <para>If all you want is to browse the Web, download files and
139 view e-mail inside the guest, then this default mode should be
140 sufficient for you, and you can safely skip the rest of this
141 section. Please note that there are certain limitations when using
142 Windows file sharing (see <xref linkend="nat-limitations" /> for
143 details).</para>
144 </glossdef>
145 </glossentry>
146
147 <glossentry>
148 <glossterm>Bridged networking</glossterm>
149
150 <glossdef>
151 <para>This is for more advanced networking needs such as network
152 simulations and running servers in a guest. When enabled,
153 VirtualBox connects to one of your installed network cards and
154 exchanges network packets directly, circumventing your host
155 operating system's network stack.</para>
156 </glossdef>
157 </glossentry>
158
159 <glossentry>
160 <glossterm>Internal networking</glossterm>
161
162 <glossdef>
163 <para>This can be used to create a different kind of
164 software-based network which is visible to selected virtual
165 machines, but not to applications running on the host or to the
166 outside world.</para>
167 </glossdef>
168 </glossentry>
169
170 <glossentry>
171 <glossterm>Host-only networking</glossterm>
172
173 <glossdef>
174 <para>This can be used to create a network containing the host and
175 a set of virtual machines, without the need for the host's
176 physical network interface. Instead, a virtual network interface
177 (similar to a loopback interface) is created on the host,
178 providing connectivity among virtual machines and the host.</para>
179 </glossdef>
180 </glossentry>
181
182 <glossentry>
183 <glossterm>Generic networking</glossterm>
184
185 <glossdef>
186 <para>Rarely used modes share the same generic network interface,
187 by allowing the user to select a driver which can be included with
188 VirtualBox or be distributed in an extension pack.</para>
189
190 <para>At the moment there are potentially two available
191 sub-modes:</para>
192
193 <para><glosslist>
194 <glossentry>
195 <glossterm>UDP Tunnel</glossterm>
196
197 <glossdef>
198 <para>This can be used to interconnect virtual machines
199 running on different hosts directly, easily and
200 transparently, over existing network
201 infrastructure.</para>
202 </glossdef>
203 </glossentry>
204
205 <glossentry>
206 <glossterm>VDE (Virtual Distributed Ethernet)
207 networking</glossterm>
208
209 <glossdef>
210 <para>This option can be used to connect to a Virtual
211 Distributed Ethernet switch on a Linux or a FreeBSD host.
212 At the moment this needs compiling VirtualBox from
213 sources, as the Oracle packages do not include it.</para>
214 </glossdef>
215 </glossentry>
216 </glosslist></para>
217 </glossdef>
218 </glossentry>
219 </glosslist></para>
220
221 <para>The following sections describe the available network modes in more
222 detail.</para>
223 </sect1>
224
225 <sect1 id="network_nat">
226 <title>Network Address Translation (NAT)</title>
227
228 <para>Network Address Translation (NAT) is the simplest way of accessing
229 an external network from a virtual machine. Usually, it does not require
230 any configuration on the host network and guest system. For this reason,
231 it is the default networking mode in VirtualBox.</para>
232
233 <para>A virtual machine with NAT enabled acts much like a real computer
234 that connects to the Internet through a router. The "router", in this
235 case, is the VirtualBox networking engine, which maps traffic from and to
236 the virtual machine transparently. In VirtualBox this router is placed
237 between each virtual machine and the host. This separation maximizes
238 security since by default virtual machines cannot talk to each
239 other.</para>
240
241 <para>The disadvantage of NAT mode is that, much like a private network
242 behind a router, the virtual machine is invisible and unreachable from the
243 outside internet; you cannot run a server this way unless you set up port
244 forwarding (described below).</para>
245
246 <para>The network frames sent out by the guest operating system are
247 received by VirtualBox's NAT engine, which extracts the TCP/IP data and
248 resends it using the host operating system. To an application on the host,
249 or to another computer on the same network as the host, it looks like the
250 data was sent by the VirtualBox application on the host, using an IP
251 address belonging to the host. VirtualBox listens for replies to the
252 packages sent, and repacks and resends them to the guest machine on its
253 private network.</para>
254
255 <para>The virtual machine receives its network address and configuration
256 on the private network from a DHCP server integrated into VirtualBox. The
257 IP address thus assigned to the virtual machine is usually on a completely
258 different network than the host. As more than one card of a virtual
259 machine can be set up to use NAT, the first card is connected to the
260 private network 10.0.2.0, the second card to the network 10.0.3.0 and so
261 on. If you need to change the guest-assigned IP range for some reason,
262 please refer to <xref linkend="changenat" />.</para>
263
264 <sect2 id="natforward">
265 <title>Configuring port forwarding with NAT</title>
266
267 <para>As the virtual machine is connected to a private network internal
268 to VirtualBox and invisible to the host, network services on the guest
269 are not accessible to the host machine or to other computers on the same
270 network. However, like a physical router, VirtualBox can make selected
271 services available to the world outside the guest through <emphasis
272 role="bold">port forwarding.</emphasis> This means that VirtualBox
273 listens to certain ports on the host and resends all packets which
274 arrive there to the guest, on the same or a different port.</para>
275
276 <para>To an application on the host or other physical (or virtual)
277 machines on the network, it looks as though the service being proxied is
278 actually running on the host. This also means that you cannot run the
279 same service on the same ports on the host. However, you still gain the
280 advantages of running the service in a virtual machine -- for example,
281 services on the host machine or on other virtual machines cannot be
282 compromised or crashed by a vulnerability or a bug in the service, and
283 the service can run in a different operating system than the host
284 system.</para>
285
286 <para>To configure Port Forwarding you can use the graphical Port
287 Forwarding editor which can be found in the Network Settings dialog
288 for Network Adaptors configured to use NAT. Here you can map host
289 ports to guest ports to allow network traffic to be routed to a
290 specific port in the guest.</para>
291
292 <para>Alternatively command line tool <computeroutput>VBoxManage</computeroutput> could be used;
293 for details, please refer to <xref linkend="vboxmanage-modifyvm" />.</para>
294
295 <para>You will need to know which ports on the guest the service uses
296 and to decide which ports to use on the host (often but not always you
297 will want to use the same ports on the guest and on the host). You can
298 use any ports on the host which are not already in use by a service. For
299 example, to set up incoming NAT connections to an
300 <computeroutput>ssh</computeroutput> server in the guest, use the
301 following command: <screen>VBoxManage modifyvm "VM name" --natpf1 "guestssh,tcp,,2222,,22"</screen>With
302 the above example, all TCP traffic arriving on port 2222 on any host
303 interface will be forwarded to port 22 in the guest. The protocol name
304 <computeroutput>tcp</computeroutput> is a mandatory attribute defining
305 which protocol should be used for forwarding
306 (<computeroutput>udp</computeroutput> could also be used). The name
307 <computeroutput>guestssh</computeroutput> is purely descriptive and will
308 be auto-generated if omitted. The number after
309 <computeroutput>--natpf</computeroutput> denotes the network card, like
310 in other parts of VBoxManage.</para>
311
312 <para>To remove this forwarding rule again, use the following command:
313 <screen>VBoxManage modifyvm "VM name" --natpf1 delete "guestssh"</screen></para>
314
315 <para>If for some reason the guest uses a static assigned IP address not
316 leased from the built-in DHCP server, it is required to specify the
317 guest IP when registering the forwarding rule: <screen>VBoxManage modifyvm "VM name" --natpf1 "guestssh,tcp,,2222,10.0.2.19,22"</screen>This
318 example is identical to the previous one, except that the NAT engine is
319 being told that the guest can be found at the 10.0.2.19 address.</para>
320
321 <para>To forward <emphasis>all</emphasis> incoming traffic from a
322 specific host interface to the guest, specify the IP of that host
323 interface like this:<screen>VBoxManage modifyvm "VM name" --natpf1 "guestssh,tcp,127.0.0.1,2222,,22"</screen>This
324 forwards all TCP traffic arriving on the localhost interface (127.0.0.1)
325 via port 2222 to port 22 in the guest.</para>
326
327 <para>It is possible to configure incoming NAT connections while the
328 VM is running, see <xref linkend="vboxmanage-controlvm"/>.</para>
329 </sect2>
330
331 <sect2 id="nat-tftp">
332 <title>PXE booting with NAT</title>
333
334 <para>PXE booting is now supported in NAT mode. The NAT DHCP server
335 provides a boot file name of the form
336 <computeroutput>vmname.pxe</computeroutput> if the directory
337 <computeroutput>TFTP</computeroutput> exists in the directory where the
338 user's <computeroutput>VirtualBox.xml</computeroutput> file is kept. It
339 is the responsibility of the user to provide
340 <computeroutput>vmname.pxe</computeroutput>.</para>
341 </sect2>
342
343 <sect2 id="nat-limitations">
344 <title>NAT limitations</title>
345
346 <para>There are four <emphasis role="bold">limitations</emphasis> of NAT
347 mode which users should be aware of:</para>
348
349 <glosslist>
350 <glossentry>
351 <glossterm>ICMP protocol limitations:</glossterm>
352
353 <glossdef>
354 <para>Some frequently used network debugging tools (e.g.
355 <computeroutput>ping</computeroutput> or tracerouting) rely on the
356 ICMP protocol for sending/receiving messages. While ICMP support
357 has been improved with VirtualBox 2.1
358 (<computeroutput>ping</computeroutput> should now work), some
359 other tools may not work reliably.</para>
360 </glossdef>
361 </glossentry>
362
363 <glossentry>
364 <glossterm>Receiving of UDP broadcasts is not reliable:</glossterm>
365
366 <glossdef>
367 <para>The guest does not reliably receive broadcasts, since, in
368 order to save resources, it only listens for a certain amount of
369 time after the guest has sent UDP data on a particular port. As a
370 consequence, NetBios name resolution based on broadcasts does not
371 always work (but WINS always works). As a workaround, you can use
372 the numeric IP of the desired server in the
373 <computeroutput>\\server\share</computeroutput> notation.</para>
374 </glossdef>
375 </glossentry>
376
377 <glossentry>
378 <glossterm>Protocols such as GRE are unsupported:</glossterm>
379
380 <glossdef>
381 <para>Protocols other than TCP and UDP are not supported. This
382 means some VPN products (e.g. PPTP from Microsoft) cannot be used.
383 There are other VPN products which use simply TCP and UDP.</para>
384 </glossdef>
385 </glossentry>
386
387 <glossentry>
388 <glossterm>Forwarding host ports &lt; 1024 impossible:</glossterm>
389
390 <glossdef>
391 <para>On Unix-based hosts (e.g. Linux, Solaris, Mac OS X) it is
392 not possible to bind to ports below 1024 from applications that
393 are not run by <computeroutput>root</computeroutput>. As a result,
394 if you try to configure such a port forwarding, the VM will refuse
395 to start.</para>
396 </glossdef>
397 </glossentry>
398 </glosslist>
399
400 <para>These limitations normally don't affect standard network use. But
401 the presence of NAT has also subtle effects that may interfere with
402 protocols that are normally working. One example is NFS, where the
403 server is often configured to refuse connections from non-privileged
404 ports (i.e. ports not below 1024).</para>
405 </sect2>
406 </sect1>
407
408 <sect1 id="network_nat_service">
409 <title>Network Address Translation Service (experimental)</title>
410
411 <para>The Network Address Translation (NAT) service works in a similar way
412 to a home router, grouping the systems using it into a network and
413 preventing systems outside of this network from directly accessing systems
414 inside it, but letting systems inside communicate with each other and with
415 systems outside using TCP and UDP over IPv4 and IPv6.</para>
416
417 <para>A NAT service is attached to an internal network. Virtual machines
418 which are to make use of it should be attached to that internal network.
419 The name of internal network is chosen when the NAT service is created and
420 the internal network will be created if it does not already exist. An
421 example command to create a NAT network is:
422 </para>
423 <para><screen>VBoxManage natnetwork add -t nat-int-network -n "192.168.15.0/24" -e</screen></para>
424 <para>
425 Here, "nat-int-network" is the name of the internal network to be used and
426 "192.168.15.0/24" is the network address and mask of the NAT service
427 interface. By default in this static configuration the gateway will be
428 assigned the address 192.168.15.1 (the address following the interface
429 address), though this is subject to change. To attach a DHCP server to the
430 internal network, we modify the example as follows:</para>
431 <para><screen>VBoxManage natnetwork add -t nat-int-network -n "192.168.15.0/24" -e -h on</screen></para>
432 <para> or to add a DHCP server to the network after creation:</para>
433 <para><screen>VBoxManage natnetwork modify -t nat-int-network -h on</screen></para>
434 <para>To disable it again, use:</para>
435 <para><screen>VBoxManage natnetwork modify -t nat-int-network -h off</screen></para>
436 <para>A current limitation of the DHCP server is that it does not provide
437 nameserver information, so systems using it should be explicitly configured
438 to use the right DNS servers.</para>
439 <para>To start the NAT service, use the following command:</para>
440 <para><screen>VBoxManage natnetwork start -t nat-int-network</screen></para>
441 <para>If the network has a DHCP server attached then it will start together
442 with the NAT network service.</para>
443 <para><screen>VBoxManage natnetwork stop -t nat-int-network</screen> stops
444 the NAT network service, together with DHCP server if any.</para>
445 <para>To delete the NAT network service use:</para>
446 <para><screen>VBoxManage natnetwork remove -t nat-int-network</screen></para>
447 <para>This command does not remove the DHCP server if one is enabled on the
448 internal network.</para>
449 <para>Port-forwarding is supported (using the "-p" switch for IPv4 and "-P"
450 for IPv6):</para>
451 <para><screen>VBoxManage natnetwork modify -t nat-int-network -p "ssh:tcp:[]:10022:[192.168.15.15]:22"</screen></para>
452 <para>This adds a port-forwarding rule from the host's TCP 10022 port to
453 the port 22 on the guest with IP address 192.168.15.15. To delete the rule,
454 use:</para>
455 <para><screen>VBoxManage natnetwork modify -t nat-int-network -p delete ssh</screen></para>
456 <para>To see the list of registered NAT networks, use:</para>
457 <para><screen>VBoxManage list natnetworks</screen></para>
458 </sect1>
459
460 <sect1 id="network_bridged">
461 <title>Bridged networking</title>
462
463 <para>With bridged networking, VirtualBox uses a device driver on your
464 <emphasis>host</emphasis> system that filters data from your physical
465 network adapter. This driver is therefore called a "net filter" driver.
466 This allows VirtualBox to intercept data from the physical network and
467 inject data into it, effectively creating a new network interface in
468 software. When a guest is using such a new software interface, it looks to
469 the host system as though the guest were physically connected to the
470 interface using a network cable: the host can send data to the guest
471 through that interface and receive data from it. This means that you can
472 set up routing or bridging between the guest and the rest of your
473 network.</para>
474
475 <para>For this to work, VirtualBox needs a device driver on your host
476 system. The way bridged networking works has been completely rewritten
477 with VirtualBox 2.0 and 2.1, depending on the host operating system. From
478 the user perspective, the main difference is that complex configuration is
479 no longer necessary on any of the supported host operating
480 systems.<footnote>
481 <para>For Mac OS X and Solaris hosts, net filter drivers were already
482 added in VirtualBox 2.0 (as initial support for Host Interface
483 Networking on these platforms). With VirtualBox 2.1, net filter
484 drivers were also added for the Windows and Linux hosts, replacing the
485 mechanisms previously present in VirtualBox for those platforms;
486 especially on Linux, the earlier method required creating TAP
487 interfaces and bridges, which was complex and varied from one
488 distribution to the next. None of this is necessary anymore. Bridged
489 network was formerly called "Host Interface Networking" and has been
490 renamed with version 2.2 without any change in functionality.</para>
491 </footnote></para>
492
493 <para><note>
494 <para>Even though TAP is no longer necessary on Linux with bridged
495 networking, you <emphasis>can</emphasis> still use TAP interfaces for
496 certain advanced setups, since you can connect a VM to any host
497 interface -- which could also be a TAP interface.</para>
498 </note>To enable bridged networking, all you need to do is to open the
499 Settings dialog of a virtual machine, go to the "Network" page and select
500 "Bridged network" in the drop down list for the "Attached to" field.
501 Finally, select desired host interface from the list at the bottom of the
502 page, which contains the physical network interfaces of your systems. On a
503 typical MacBook, for example, this will allow you to select between "en1:
504 AirPort" (which is the wireless interface) and "en0: Ethernet", which
505 represents the interface with a network cable.</para>
506
507 <note><para>Bridging to a wireless interface is done differently from
508 bridging to a wired interface, because most wireless adapters do not
509 support promiscuous mode. All traffic has to use the MAC address of the
510 host's wireless adapter, and therefore VirtualBox needs to replace the
511 source MAC address in the Ethernet header of an outgoing packet to make
512 sure the reply will be sent to the host interface. When VirtualBox sees
513 an incoming packet with a destination IP address that belongs to one of
514 the virtual machine adapters it replaces the destination MAC address in
515 the Ethernet header with the VM adapter's MAC address and passes it on.
516 VirtualBox examines ARP and DHCP packets in order to learn the IP
517 addresses of virtual machines.</para></note>
518
519 <para>Depending on your host operating system, the following limitations
520 should be kept in mind:<itemizedlist>
521 <listitem>
522 <para>On <emphasis role="bold">Macintosh</emphasis> hosts,
523 functionality is limited when using AirPort (the Mac's wireless
524 networking) for bridged networking. Currently, VirtualBox supports
525 only IPv4 over AirPort. For other protocols such as IPv6 and IPX,
526 you must choose a wired interface.</para>
527 </listitem>
528
529 <listitem>
530 <para>On <emphasis role="bold">Linux</emphasis> hosts, functionality
531 is limited when using wireless interfaces for bridged networking.
532 Currently, VirtualBox supports only IPv4 over wireless. For other
533 protocols such as IPv6 and IPX, you must choose a wired
534 interface.</para>
535
536 <para>Also, setting the MTU to less than 1500 bytes on wired
537 interfaces provided by the sky2 driver on the Marvell Yukon II EC
538 Ultra Ethernet NIC is known to cause packet losses under certain
539 conditions.</para>
540
541 <para>Some adapters strip VLAN tags in hardware. This does not allow
542 to use VLAN trunking between VM and the external network with
543 pre-2.6.27 Linux kernels nor with host operating systems other than
544 Linux.</para>
545 </listitem>
546
547 <listitem>
548 <para>On <emphasis role="bold">Solaris</emphasis> hosts, there is no
549 support for using wireless interfaces. Filtering guest traffic using
550 IPFilter is also not completely supported due to technical
551 restrictions of the Solaris networking subsystem. These issues would
552 be addressed in a future release of Solaris 11.</para>
553
554 <para>Starting with VirtualBox 4.1, on Solaris 11 hosts (build 159
555 and above), it is possible to use Solaris' Crossbow Virtual Network
556 Interfaces (VNICs) directly with VirtualBox without any additional
557 configuration other than each VNIC must be exclusive for every guest
558 network interface. With VirtualBox 2.0.4 and above, VNICs can be
559 used but with the following caveats:</para>
560
561 <itemizedlist>
562 <listitem>
563 <para>A VNIC cannot be shared between multiple guest network
564 interfaces, i.e. each guest network interface must have its own,
565 exclusive VNIC.</para>
566 </listitem>
567
568 <listitem>
569 <para>The VNIC and the guest network interface that uses the
570 VNIC must be assigned identical MAC addresses.</para>
571 </listitem>
572 </itemizedlist>
573
574 <para>When using VLAN interfaces with VirtualBox, they must be named
575 according to the PPA-hack naming scheme (e.g. "e1000g513001"), as
576 otherwise the guest may receive packets in an unexpected
577 format.</para>
578 </listitem>
579 </itemizedlist></para>
580 </sect1>
581
582 <sect1 id="network_internal">
583 <title>Internal networking</title>
584
585 <para>Internal Networking is similar to bridged networking in that the VM
586 can directly communicate with the outside world. However, the "outside
587 world" is limited to other VMs on the same host which connect to the same
588 internal network.</para>
589
590 <para>Even though technically, everything that can be done using internal
591 networking can also be done using bridged networking, there are security
592 advantages with internal networking. In bridged networking mode, all
593 traffic goes through a physical interface of the host system. It is
594 therefore possible to attach a packet sniffer (such as Wireshark) to the
595 host interface and log all traffic that goes over it. If, for any reason,
596 you prefer two or more VMs on the same machine to communicate privately,
597 hiding their data from both the host system and the user, bridged
598 networking therefore is not an option.</para>
599
600 <para>Internal networks are created automatically as needed, i.e. there is
601 no central configuration. Every internal network is identified simply by
602 its name. Once there is more than one active virtual network card with the
603 same internal network ID, the VirtualBox support driver will automatically
604 "wire" the cards and act as a network switch. The VirtualBox support
605 driver implements a complete Ethernet switch and supports both
606 broadcast/multicast frames and promiscuous mode.</para>
607
608 <para>In order to attach a VM's network card to an internal network, set
609 its networking mode to "internal networking". There are two ways to
610 accomplish this:</para>
611
612 <para><itemizedlist>
613 <listitem>
614 <para>You can use a VM's "Settings" dialog in the VirtualBox
615 graphical user interface. In the "Networking" category of the
616 settings dialog, select "Internal Networking" from the drop-down
617 list of networking modes. Now select the name of an existing
618 internal network from the drop-down below or enter a new name into
619 the entry field.</para>
620 </listitem>
621
622 <listitem>
623 <para>You can use <screen>VBoxManage modifyvm "VM name" --nic&lt;x&gt; intnet</screen>
624 Optionally, you can specify a network name with the command <screen>VBoxManage modifyvm "VM name" --intnet&lt;x&gt; "network name"</screen>
625 If you do not specify a network name, the network card will be
626 attached to the network <computeroutput>intnet</computeroutput> by
627 default.</para>
628 </listitem>
629 </itemizedlist></para>
630
631 <para>Unless you configure the (virtual) network cards in the guest
632 operating systems that are participating in the internal network to use
633 static IP addresses, you may want to use the DHCP server that is built
634 into VirtualBox to manage IP addresses for the internal network. Please
635 see <xref linkend="vboxmanage-dhcpserver" /> for details.</para>
636
637 <para>As a security measure, the Linux implementation of internal
638 networking only allows VMs running under the same user ID to establish an
639 internal network.</para>
640 </sect1>
641
642 <sect1 id="network_hostonly">
643 <title>Host-only networking</title>
644
645 <para>Host-only networking is another networking mode that was added with
646 version 2.2 of VirtualBox. It can be thought of as a hybrid between the
647 bridged and internal networking modes: as with bridged networking, the
648 virtual machines can talk to each other and the host as if they were
649 connected through a physical Ethernet switch. Similarly, as with internal
650 networking however, a physical networking interface need not be present,
651 and the virtual machines cannot talk to the world outside the host since
652 they are not connected to a physical networking interface.</para>
653
654 <para>Instead, when host-only networking is used, VirtualBox creates a new
655 software interface on the host which then appears next to your existing
656 network interfaces. In other words, whereas with bridged networking an
657 existing physical interface is used to attach virtual machines to, with
658 host-only networking a new "loopback" interface is created on the host.
659 And whereas with internal networking, the traffic between the virtual
660 machines cannot be seen, the traffic on the "loopback" interface on the
661 host can be intercepted.</para>
662
663 <para>Host-only networking is particularly useful for preconfigured
664 virtual appliances, where multiple virtual machines are shipped together
665 and designed to cooperate. For example, one virtual machine may contain a
666 web server and a second one a database, and since they are intended to
667 talk to each other, the appliance can instruct VirtualBox to set up a
668 host-only network for the two. A second (bridged) network would then
669 connect the web server to the outside world to serve data to, but the
670 outside world cannot connect to the database.</para>
671
672 <para>To change a virtual machine's virtual network interface to "host
673 only" mode:<itemizedlist>
674 <listitem>
675 <para>either go to the "Network" page in the virtual machine's
676 settings notebook in the graphical user interface and select
677 "Host-only networking", or</para>
678 </listitem>
679
680 <listitem>
681 <para>on the command line, type <computeroutput>VBoxManage modifyvm
682 "VM name" --nic&lt;x&gt; hostonly</computeroutput>; see <xref
683 linkend="vboxmanage-modifyvm" /> for details.</para>
684 </listitem>
685 </itemizedlist></para>
686
687 <para>For host-only networking, like with internal networking, you may
688 find the DHCP server useful that is built into VirtualBox. This can be
689 enabled to then manage the IP addresses in the host-only network since
690 otherwise you would need to configure all IP addresses
691 statically.<itemizedlist>
692 <listitem>
693 <para>In the VirtualBox graphical user interface, you can configure
694 all these items in the global settings via "File" -&gt; "Settings"
695 -&gt; "Network", which lists all host-only networks which are
696 presently in use. Click on the network name and then on the "Edit"
697 button to the right, and you can modify the adapter and DHCP
698 settings.</para>
699 </listitem>
700
701 <listitem>
702 <para>Alternatively, you can use <computeroutput>VBoxManage
703 dhcpserver</computeroutput> on the command line; please see <xref
704 linkend="vboxmanage-dhcpserver" /> for details.</para>
705 </listitem>
706 </itemizedlist></para>
707 <para><note>On Linux and Mac OS X hosts the number of host-only interfaces is
708 limited to 128. There is no such limit for Solaris and Windows hosts.</note></para>
709 </sect1>
710
711 <sect1 id="network_udp_tunnel">
712 <title>UDP Tunnel networking</title>
713
714 <para>This networking mode allows to interconnect virtual machines running
715 on different hosts.</para>
716
717 <para>Technically this is done by encapsulating Ethernet frames sent or
718 received by the guest network card into UDP/IP datagrams, and sending them
719 over any network available to the host.</para>
720
721 <para>UDP Tunnel mode has three parameters:<glosslist>
722 <glossentry>
723 <glossterm>Source UDP port</glossterm>
724
725 <glossdef>
726 <para>The port on which the host listens. Datagrams arriving on
727 this port from any source address will be forwarded to the
728 receiving part of the guest network card.</para>
729 </glossdef>
730 </glossentry>
731
732 <glossentry>
733 <glossterm>Destination address</glossterm>
734
735 <glossdef>
736 <para>IP address of the target host of the transmitted
737 data.</para>
738 </glossdef>
739 </glossentry>
740
741 <glossentry>
742 <glossterm>Destination UDP port</glossterm>
743
744 <glossdef>
745 <para>Port number to which the transmitted data is sent.</para>
746 </glossdef>
747 </glossentry>
748 </glosslist></para>
749
750 <para>When interconnecting two virtual machines on two different hosts,
751 their IP addresses must be swapped. On single host, source and destination
752 UDP ports must be swapped.</para>
753
754 <para>In the following example host 1 uses the IP address 10.0.0.1 and
755 host 2 uses IP address 10.0.0.2. Configuration via command-line:<screen> VBoxManage modifyvm "VM 01 on host 1" --nic&lt;x&gt; generic
756 VBoxManage modifyvm "VM 01 on host 1" --nicgenericdrv&lt;x&gt; UDPTunnel
757 VBoxManage modifyvm "VM 01 on host 1" --nicproperty&lt;x&gt; dest=10.0.0.2
758 VBoxManage modifyvm "VM 01 on host 1" --nicproperty&lt;x&gt; sport=10001
759 VBoxManage modifyvm "VM 01 on host 1" --nicproperty&lt;x&gt; dport=10002</screen>
760 and <screen> VBoxManage modifyvm "VM 02 on host 2" --nic&lt;y&gt; generic
761 VBoxManage modifyvm "VM 02 on host 2" --nicgenericdrv&lt;y&gt; UDPTunnel
762 VBoxManage modifyvm "VM 02 on host 2" --nicproperty&lt;y&gt; dest=10.0.0.1
763 VBoxManage modifyvm "VM 02 on host 2" --nicproperty&lt;y&gt; sport=10002
764 VBoxManage modifyvm "VM 02 on host 2" --nicproperty&lt;y&gt; dport=10001</screen></para>
765
766 <para>Of course, you can always interconnect two virtual machines on the
767 same host, by setting the destination address parameter to 127.0.0.1 on
768 both. It will act similarly to "Internal network" in this case, however
769 the host can see the network traffic which it could not in the normal
770 Internal network case.</para>
771
772 <para><note>
773 On Unix-based hosts (e.g. Linux, Solaris, Mac OS X) it is not possible
774 to bind to ports below 1024 from applications that are not run by
775 <computeroutput>root</computeroutput>. As a result, if you try to
776 configure such a source UDP port, the VM will refuse to start.
777 </note></para>
778 </sect1>
779
780 <sect1 id="network_vde">
781 <title>VDE networking</title>
782
783 <para>Virtual Distributed Ethernet (VDE<footnote>
784 <para>VDE is a project developed by Renzo Davoli, Associate Professor
785 at the University of Bologna, Italy.</para>
786 </footnote>) is a flexible, virtual network infrastructure system,
787 spanning across multiple hosts in a secure way. It allows for L2/L3
788 switching, including spanning-tree protocol, VLANs, and WAN emulation. It
789 is an optional part of VirtualBox which is only included in the source
790 code.</para>
791
792 <para>The basic building blocks of the infrastructure are VDE switches,
793 VDE plugs and VDE wires which inter-connect the switches.</para>
794
795 <para>The VirtualBox VDE driver has one parameter:<glosslist>
796 <glossentry>
797 <glossterm>VDE network</glossterm>
798
799 <glossdef>
800 <para>The name of the VDE network switch socket to which the VM
801 will be connected.</para>
802 </glossdef>
803 </glossentry>
804 </glosslist></para>
805
806 <para>The following basic example shows how to connect a virtual machine
807 to a VDE switch:</para>
808
809 <para><orderedlist>
810 <listitem>
811 <para>Create a VDE switch: <screen>vde_switch -s /tmp/switch1</screen></para>
812 </listitem>
813
814 <listitem>
815 <para>Configuration via command-line: <screen>VBoxManage modifyvm "VM name" --nic&lt;x&gt; generic</screen>
816 <screen>VBoxManage modifyvm "VM name" --nicgenericdrv&lt;x&gt; VDE</screen>
817 To connect to automatically allocated switch port, use: <screen>VBoxManage modifyvm "VM name" --nicproperty&lt;x&gt; network=/tmp/switch1</screen>
818 To connect to specific switch port &lt;n&gt;, use: <screen>VBoxManage modifyvm "VM name" --nicproperty&lt;x&gt; network=/tmp/switch1[&lt;n&gt;]</screen>
819 The latter option can be useful for VLANs.</para>
820 </listitem>
821
822 <listitem>
823 <para>Optionally map between VDE switch port and VLAN: (from switch
824 CLI) <screen>vde$ vlan/create &lt;VLAN&gt;</screen> <screen>vde$ port/setvlan &lt;port&gt; &lt;VLAN&gt;</screen></para>
825 </listitem>
826 </orderedlist></para>
827
828 <para>VDE is available on Linux and FreeBSD hosts only. It is only
829 available if the VDE software and the VDE plugin library from the
830 VirtualSquare project are installed on the host system<footnote>
831 <para>For Linux hosts, the shared library libvdeplug.so must be
832 available in the search path for shared libraries</para>
833 </footnote>. For more information on setting up VDE networks, please see
834 the documentation accompanying the software.<footnote>
835 <para><ulink
836 url="http://wiki.virtualsquare.org/wiki/index.php/VDE_Basic_Networking">http://wiki.virtualsquare.org/wiki/index.php/VDE_Basic_Networking</ulink>.</para>
837 </footnote></para>
838 </sect1>
839
840 <sect1 id="network_bandwidth_limit">
841 <title>Limiting bandwidth for network I/O</title>
842
843 <para>Starting with version 4.2, VirtualBox allows for limiting the
844 maximum bandwidth used for network transmission. Several network adapters
845 of one VM may share limits through bandwidth groups. It is possible
846 to have more than one such limit.</para>
847 <note><para>VirtualBox shapes VM traffic only in the transmit direction,
848 delaying the packets being sent by virtual machines. It does not limit
849 the traffic being received by virtual machines.</para>
850 </note>
851
852 <para>Limits are configured through
853 <computeroutput>VBoxManage</computeroutput>. The example below creates a
854 bandwidth group named "Limit", sets the limit to 20 Mbit/s and assigns the
855 group to the first and second adapters of the VM:<screen>VBoxManage bandwidthctl "VM name" add Limit --type network --limit 20m
856VBoxManage modifyvm "VM name" --nicbandwidthgroup1 Limit
857VBoxManage modifyvm "VM name" --nicbandwidthgroup2 Limit</screen></para>
858
859 <para>All adapters in a group share the bandwidth limit, meaning that in the
860 example above the bandwidth of both adapters combined can never exceed 20
861 Mbit/s. However, if one adapter doesn't require bandwidth the other can use the
862 remaining bandwidth of its group.</para>
863
864 <para>The limits for each group can be changed while the VM is running,
865 with changes being picked up immediately. The example below changes the
866 limit for the group created in the example above to 100 Kbit/s:<screen>VBoxManage bandwidthctl "VM name" set Limit --limit 100k</screen></para>
867
868 <para>To completely disable shaping for the first adapter of VM use the
869 following command:
870 <screen>VBoxManage modifyvm "VM name" --nicbandwidthgroup1 none</screen></para>
871
872 <para>It is also possible to disable shaping for all adapters assigned to a
873 bandwidth group while VM is running, by specifying the zero limit for the
874 group. For example, for the bandwidth group named "Limit" use:
875 <screen>VBoxManage bandwidthctl "VM name" set Limit --limit 0</screen></para>
876 </sect1>
877 <sect1 id="network_performance">
878 <title>Improving network performance</title>
879
880 <para>VirtualBox provides a variety of virtual network adapters that can be
881 "attached" to the host's network in a number of ways. Depending on which
882 types of adapters and attachments are used the network performance will
883 be different. Performance-wise the <emphasis>virtio</emphasis> network
884 adapter is preferable over <emphasis>Intel PRO/1000</emphasis> emulated
885 adapters, which are preferred over <emphasis>PCNet</emphasis> family of
886 adapters. Both <emphasis>virtio</emphasis> and <emphasis>Intel PRO/1000
887 </emphasis> adapters enjoy the benefit of segmentation and checksum
888 offloading. Segmentation offloading is essential for high performance as
889 it allows for less context switches, dramatically increasing the sizes
890 of packets that cross VM/host boundary.</para>
891 <note><para>Neither <emphasis>virtio</emphasis> nor <emphasis>Intel PRO/1000
892 </emphasis> drivers for Windows XP support segmentation
893 offloading. Therefore Windows XP guests never reach the same
894 transmission rates as other guest types. Refer to MS Knowledge base
895 article 842264 for additional information.</para>
896 </note>
897 <para>Three attachment types: <emphasis>internal</emphasis>,
898 <emphasis>bridged</emphasis> and <emphasis>host-only</emphasis>, have
899 nearly identical performance, the <emphasis>internal</emphasis> type
900 being a little bit faster and using less CPU cycles as the packets never
901 reach the host's network stack. The <emphasis>NAT</emphasis> attachment
902 is the slowest (and safest) of all attachment types as it provides
903 network address translation. The generic driver attachment is special and
904 cannot be considered as an alternative to other attachment types.</para>
905 <para>The number of CPUs assigned to VM does not improve network
906 performance and in some cases may hurt it due to increased concurrency in
907 the guest.</para>
908 <para>Here is the short summary of things to check in order to improve
909 network performance:</para>
910 <para><orderedlist>
911 <listitem>
912 <para>Whenever possible use <emphasis>virtio</emphasis> network
913 adapter, otherwise use one of <emphasis>Intel PRO/1000</emphasis>
914 adapters;</para>
915 </listitem>
916 <listitem>
917 <para>Use <emphasis>bridged</emphasis> attachment instead of
918 <emphasis>NAT</emphasis></para>;
919 </listitem>
920 <listitem>
921 <para>Make sure segmentation offloading is enabled in the guest OS.
922 Usually it will be enabled by default. You can check and modify
923 offloading settings using <computeroutput>ethtool</computeroutput>
924 command in Linux guests.</para>
925 </listitem>
926 </orderedlist></para>
927 </sect1>
928</chapter>
注意: 瀏覽 TracBrowser 來幫助您使用儲存庫瀏覽器

© 2024 Oracle Support Privacy / Do Not Sell My Info Terms of Use Trademark Policy Automated Access Etiquette