source: vbox/trunk/doc/manual/en_US/user_Frontends.xml@ 95825

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE chapter PUBLIC "-//OASIS//DTD DocBook XML V4.5//EN"
"http://www.oasis-open.org/docbook/xml/4.5/docbookx.dtd"[
<!ENTITY % all.entities SYSTEM "all-entities.ent">
%all.entities;
]>
<chapter id="remotevm">

  <title>Remote Virtual Machines</title>

  <sect1 id="vrde">

    <title>Remote Display (VRDP Support)</title>

    <para>
      &product-name; can display virtual machines remotely, meaning that
      a virtual machine can execute on one computer even though the
      machine will be displayed on a second computer, and the machine
      will be controlled from there as well, as if the virtual machine
      was running on that second computer.
    </para>

    <para>
      For maximum flexibility, &product-name; implements remote machine
      display through a generic extension interface called the
      VirtualBox Remote Desktop Extension (VRDE). The base open source
      &product-name; package only provides this interface, while
      implementations can be supplied by third parties with
      &product-name; extension packages, which must be installed
      separately from the base package. See
      <xref linkend="intro-installing" />.
    </para>

    <para>
      Oracle provides support for the VirtualBox Remote Display Protocol
      (VRDP) in such an &product-name; extension package.
    </para>

    <para>
      VRDP is a backwards-compatible extension to Microsoft's Remote
      Desktop Protocol (RDP). As a result, you can use any standard RDP
      client to control the remote VM.
    </para>

    <para>
      Even when the extension is installed, the VRDP server is disabled
      by default. It can easily be enabled on a per-VM basis either in
      the VirtualBox Manager in the
      <emphasis role="bold">Display</emphasis> settings, see
      <xref linkend="settings-display" />, or with the
      <command>VBoxManage</command> command, as follows:
    </para>

<screen>$ VBoxManage modifyvm <replaceable>VM-name</replaceable> --vrde on</screen>

    <para>
      By default, the VRDP server uses TCP port <literal>3389</literal>.
      You will need to change the default port if you run more than one
      VRDP server, since the port can only be used by one server at a
      time. You might also need to change it on Windows hosts since the
      default port might already be used by the RDP server that is built
      into Windows itself. Ports 5000 through 5050 are typically not
      used and might be a good choice.
    </para>

    <para>
      The port can be changed either in the
      <emphasis role="bold">Display</emphasis> settings of the graphical
      user interface or with the <option>--vrde-port</option> option of
      the <command>VBoxManage modifyvm</command> command. You can
      specify a comma-separated list of ports or ranges of ports. Use a
      dash between two port numbers to specify a range. The VRDP server
      will bind to <emphasis>one</emphasis> of the available ports from
      the specified list. For example, <command>VBoxManage modifyvm
      <replaceable>VM-name</replaceable> --vrde-port
      5000,5010-5012</command> configures the server to bind to one of
      the ports 5000, 5010, 5011, or 5012. See
      <xref linkend="vboxmanage-modifyvm" />.
    </para>

    <para>
      The actual port used by a running VM can be either queried with
      the <command>VBoxManage showvminfo</command> command or seen in
      the GUI on the <emphasis role="bold">Runtime</emphasis> tab of the
      <emphasis role="bold">Session Information</emphasis> dialog, which
      is accessible from the <emphasis role="bold">Machine</emphasis>
      menu of the VM window.
    </para>

    <para>
      &product-name; supports IPv6. If the host OS supports IPv6 the
      VRDP server will automatically listen for IPv6 connections in
      addition to IPv4.
    </para>

    <sect2 id="rdp-viewers">

      <title>Common Third-Party RDP Viewers</title>

      <para>
        Since VRDP is backwards-compatible to RDP, you can use any
        standard RDP viewer to connect to such a remote virtual machine.
        For this to work, you must specify the IP address of your
        <emphasis>host</emphasis> system, not of the virtual machine, as
        the server address to connect to. You must also specify the port
        number that the VRDP server is using.
      </para>

      <para>
        The following examples are for the most common RDP viewers:
      </para>

      <itemizedlist>

        <listitem>
          <para>
            On Windows, you can use the Microsoft Terminal Services
            Connector, <command>mstsc.exe</command>, that is included
            with Windows. Press the Windows key + R, to display the
            <emphasis role="bold">Run</emphasis> dialog. Enter
            <command>mstsc</command> to start the program. You can also
            find the program in <emphasis role="bold">Start</emphasis>,
            <emphasis role="bold">All Programs</emphasis>,
            <emphasis role="bold">Accessories</emphasis>,
            <emphasis role="bold">Remote Desktop Connection</emphasis>.
            If you use the <emphasis role="bold">Run</emphasis> dialog,
            you can enter options directly. For example:
          </para>

<screen>mstsc 1.2.3.4:3389</screen>

          <para>
            Replace <literal>1.2.3.4</literal> with the host IP address,
            and <literal>3389</literal> with a different port, if
            necessary.
          </para>

          <note>
            <itemizedlist>

              <listitem>
                <para>
                  IPv6 addresses must be enclosed in square brackets to
                  specify a port. For example: <literal>mstsc
                  [fe80::1:2:3:4]:3389</literal>
                </para>
              </listitem>

              <listitem>
                <para>
                  When connecting to localhost in order to test the
                  connection, the addresses <literal>localhost</literal>
                  and <literal>127.0.0.1</literal> might not work using
                  <command>mstsc.exe</command>. Instead, the address
                  <literal>127.0.0.2[:3389]</literal> has to be used.
                </para>
              </listitem>

            </itemizedlist>
          </note>
        </listitem>

        <listitem>
          <para>
            On other systems, you can use the standard open source
            <command>rdesktop</command> program. This ships with most
            Linux distributions.
          </para>

          <para>
            With <command>rdesktop</command>, use a command line such as
            the following:
          </para>

<screen>$ rdesktop -a 16 -N 1.2.3.4:3389</screen>

          <para>
            Replace <literal>1.2.3.4</literal> with the host IP address,
            and <literal>3389</literal> with a different port, if
            necessary. The <option>-a 16</option> option requests a
            color depth of 16 bits per pixel, which we recommend. For
            best performance, after installation of the guest operating
            system, you should set its display color depth to the same
            value. The <option>-N</option> option enables use of the
            NumPad keys.
          </para>
        </listitem>

        <listitem>
          <para>
            You can use the Remmina remote desktop client with VRDP.
            This application is included with some Linux distributions,
            such as Debian and Ubuntu.
          </para>
        </listitem>

        <listitem>
          <para>
            If you run the KDE desktop, you can use
            <command>krdc</command>, the KDE RDP viewer. A typical
            command line is as follows:
          </para>

<screen>$ krdc rdp://1.2.3.4:3389</screen>

          <para>
            Replace <literal>1.2.3.4</literal> with the host IP address,
            and <literal>3389</literal> with a different port, if
            necessary. The <literal>rdp:// </literal> prefix is required
            with <command>krdc</command> to switch it into RDP mode.
          </para>
        </listitem>

        <listitem>
          <para>
            With Sun Ray thin clients you can use
            <command>uttsc</command>, which is part of the Sun Ray
            Windows Connector package. See the Sun Ray documentation for
            details.
          </para>
        </listitem>

      </itemizedlist>

    </sect2>

    <sect2 id="vboxheadless">

      <title>VBoxHeadless, the Remote Desktop Server</title>

      <para>
        While any VM started from the VirtualBox Manager is capable of
        running virtual machines remotely, it is not convenient to have
        to run the full GUI if you never want to have VMs displayed
        locally in the first place. In particular, if you are running
        server hardware whose only purpose is to host VMs, and all your
        VMs are supposed to run remotely over VRDP, then it is pointless
        to have a graphical user interface on the server at all. This is
        especially true for Linux or Oracle Solaris hosts, as the
        VirtualBox Manager comes with dependencies on the Qt and SDL
        libraries. This is inconvenient if you would rather not have the
        X Window system on your server at all.
      </para>

      <para>
        &product-name; therefore comes with a front-end called
        <command>VBoxHeadless</command>, which produces no visible
        output on the host at all, but still can optionally deliver VRDP
        data. This front-end has no dependencies on the X Window system
        on Linux and Oracle Solaris hosts.
      </para>

      <note>
        <para>
          In legacy releases of &product-name;, the headless server was
          called <command>VBoxVRDP</command>. For backwards
          compatibility, the &product-name; installation still includes
          an executable with that name.
        </para>
      </note>

      <para>
        To start a virtual machine with <command>VBoxHeadless</command>,
        you have the following options:
      </para>

      <itemizedlist>

        <listitem>
          <para>
            Use the <command>VBoxManage</command> command, as follows:
          </para>

<screen>$ VBoxManage startvm <replaceable>VM-name</replaceable> --type headless</screen>

          <para>
            The <option>--type</option> option causes &product-name; to
            use <command>VBoxHeadless</command> as the front-end to the
            internal virtualization engine, instead of the Qt front-end.
          </para>
        </listitem>

        <listitem>
          <para>
            Use the <command>VBoxHeadless</command> command, as follows:
          </para>

<screen>VBoxHeadless --startvm <replaceable>uuid</replaceable>|<replaceable>vmname</replaceable></screen>

          <para>
            This way of starting the VM helps troubleshooting problems
            reported by <command>VBoxManage startvm</command>, because
            you can sometimes see more detailed error messages,
            especially for early failures before the VM execution is
            started. In normal situations <command>VBoxManage
            startvm</command> is preferred, since it runs the VM
            directly as a background process which has to be done
            explicitly when directly starting with
            <command>VBoxHeadless</command>. The full documentation of
            the command is in <xref linkend="man_vboxheadless"/>.
          </para>
        </listitem>

        <listitem>
          <para>
            Start <command>VBoxHeadless</command> from the VirtualBox
            Manager GUI, by pressing the Shift key when starting a
            virtual machine or by selecting
            <emphasis role="bold">Headless Start</emphasis> from the
            <emphasis role="bold">Machine</emphasis> menu.
          </para>
        </listitem>

      </itemizedlist>

      <para>
        When you use the <command>VBoxHeadless</command> command to
        start a VM, the VRDP server will be enabled according to the VM
        configuration. You can override the VM's setting using
        <option>--vrde</option> command line parameter. To enable the
        VRDP server, start the VM as follows:
      </para>

<screen>VBoxHeadless --startvm <replaceable>uuid</replaceable>|<replaceable>vmname</replaceable> --vrde on</screen>

      <para>
        To disable the VRDP server:
      </para>

<screen>VBoxHeadless --startvm <replaceable>uuid</replaceable>|<replaceable>vmname</replaceable> --vrde off</screen>

      <para>
        To have the VRDP server enabled depending on the VM
        configuration, as for other front-ends:
      </para>

<screen>VBoxHeadless --startvm <replaceable>uuid</replaceable>|<replaceable>vmname</replaceable> --vrde config</screen>

      <para>
        This command is the same as the following:
      </para>

<screen>VBoxHeadless --startvm <replaceable>uuid</replaceable>|<replaceable>vmname</replaceable></screen>

      <para>
        If you start the VM with <command>VBoxManage startvm</command>
        then the configuration settings of the VM are always used.
      </para>

    </sect2>

    <sect2 id="headless-vm-steps">

      <title>Step by Step: Creating a Virtual Machine on a Headless Server</title>

      <para>
        The following instructions describe how to create a virtual
        machine on a headless server over a network connection. This
        example creates a virtual machine, establishes an RDP connection
        and installs a guest operating system. All of these tasks are
        done without having to touch the headless server. You need the
        following prerequisites:
      </para>

      <itemizedlist>

        <listitem>
          <para>
            &product-name; on a server machine with a supported host
            operating system. The &product-name; Extension Pack for the
            VRDP server must be installed, see <xref linkend="vrde"/>.
            The procedures assume a Linux server is used.
          </para>
        </listitem>

        <listitem>
          <para>
            An ISO file accessible from the server, containing the
            installation data for the guest operating system to install.
            Windows XP is used in the example.
          </para>
        </listitem>

        <listitem>
          <para>
            A terminal connection to that host through which you can
            access a command line, such as <command>ssh</command>.
          </para>
        </listitem>

        <listitem>
          <para>
            An RDP viewer on the remote client. See
            <xref linkend="rdp-viewers" /> for examples.
          </para>
        </listitem>

      </itemizedlist>

      <para>
        Note that on the server machine, since we will only use the
        headless server, Qt and the X Window system are not required.
      </para>

      <orderedlist>

        <listitem>
          <para>
            On the headless server, create a new virtual machine. For
            example:
          </para>

<screen>VBoxManage createvm --name "Windows XP" --ostype WindowsXP --register</screen>

          <para>
            If you do not specify <option>--register</option>, you will
            have to manually use the <command>registervm</command>
            command later.
          </para>

          <para>
            You do not need to specify <option>--ostype</option>, but
            doing so selects some sensible default values for certain VM
            parameters. For example, the RAM size and the type of the
            virtual network device. To get a complete list of supported
            operating systems you can use the following command:
          </para>

<screen>VBoxManage list ostypes</screen>
        </listitem>

        <listitem>
          <para>
            Make sure the settings for the VM are appropriate for the
            guest operating system that we will install. For example:
          </para>

<screen>VBoxManage modifyvm "Windows XP" --memory 256 --acpi on --boot1 dvd --nic1 nat</screen>
        </listitem>

        <listitem>
          <para>
            Create a virtual hard disk for the VM. For example, to
            create a 10 GB virtual hard disk:
          </para>

<screen>VBoxManage createhd --filename "WinXP.vdi" --size 10000</screen>
        </listitem>

        <listitem>
          <para>
            Add an IDE Controller to the new VM. For example:
          </para>

<screen>VBoxManage storagectl "Windows XP" --name "IDE Controller"
  --add ide --controller PIIX4</screen>
        </listitem>

        <listitem>
          <para>
            Set the VDI file you created as the first virtual hard disk
            of the new VM. For example:
          </para>

<screen>VBoxManage storageattach "Windows XP" --storagectl "IDE Controller"
 --port 0 --device 0 --type hdd --medium "WinXP.vdi"</screen>
        </listitem>

        <listitem>
          <para>
            Attach the ISO file that contains the operating system
            installation that you want to install later to the virtual
            machine. This is done so that the VM can boot from it.
          </para>

<screen>VBoxManage storageattach "Windows XP" --storagectl "IDE Controller"
 --port 0 --device 1 --type dvddrive --medium /full/path/to/iso.iso</screen>
        </listitem>

        <listitem>
          <para>
            Enable the VirtualBox Remote Desktop Extension, the VRDP
            server, as follows:
          </para>

<screen>VBoxManage modifyvm "Windows XP" --vrde on</screen>
        </listitem>

        <listitem>
          <para>
            Start the virtual machine using the
            <command>VBoxHeadless</command> command:
          </para>

<screen>VBoxHeadless --startvm "Windows XP"</screen>

          <para>
            If the configuration steps worked, you should see a
            copyright notice. If you are returned to the command line,
            then something did not work correctly.
          </para>
        </listitem>

        <listitem>
          <para>
            On the client machine, start the RDP viewer and connect to
            the server. See <xref linkend="rdp-viewers" /> for details
            of how to use various common RDP viewers.
          </para>

          <para>
            The installation routine of your guest operating system
            should be displayed in the RDP viewer.
          </para>
        </listitem>

      </orderedlist>

    </sect2>

    <sect2 id="usb-over-rdp">

      <title>Remote USB</title>

      <para>
        As a special feature additional to the VRDP support,
        &product-name; also supports remote USB devices over the wire.
        That is, an &product-name; guest that runs on one computer can
        access the USB devices of the remote computer on which the VRDP
        data is being displayed the same way as USB devices that are
        connected to the actual host. This enables running of virtual
        machines on an &product-name; host that acts as a server, where
        a client can connect from elsewhere that needs only a network
        adapter and a display capable of running an RDP viewer. When USB
        devices are plugged into the client, the remote &product-name;
        server can access them.
      </para>

      <para>
        For these remote USB devices, the same filter rules apply as for
        other USB devices. See <xref linkend="settings-usb" />. All you
        have to do is specify Remote, or Any, when setting up these
        rules.
      </para>

      <para>
        Accessing remote USB devices is only possible if the RDP client
        supports this extension. Some versions of
        <command>uttsc</command>, a client tailored for the use with Sun
        Ray thin clients, support accessing remote USB devices. RDP
        clients for other platforms will be provided in future
        &product-name; versions.
      </para>

    </sect2>

    <sect2 id="vbox-auth">

      <title>RDP Authentication</title>

      <para>
        For each virtual machine that is remotely accessible using RDP,
        you can individually determine if and how client connections are
        authenticated. For this, use the <command>VBoxManage
        modifyvm</command> command with the
        <option>--vrde-auth-type</option> option. See
        <xref linkend="vboxmanage-modifyvm" />. The following methods of
        authentication are available:
      </para>

      <itemizedlist>

        <listitem>
          <para>
            The <emphasis role="bold">null</emphasis> method means that
            there is no authentication at all. Any client can connect to
            the VRDP server and thus the virtual machine. This is very
            insecure and only to be recommended for private networks.
          </para>
        </listitem>

        <listitem>
          <para>
            The <emphasis role="bold">external</emphasis> method
            provides external authentication through a special
            authentication library. &product-name; ships with two
            special authentication libraries:
          </para>

          <orderedlist>

            <listitem>
              <para>
                The default authentication library,
                <command>VBoxAuth</command>, authenticates against user
                credentials of the hosts. Depending on the host
                platform, this means the following:
              </para>

              <itemizedlist>

                <listitem>
                  <para>
                    On Linux hosts, <command>VBoxAuth.so</command>
                    authenticates users against the host's PAM system.
                  </para>
                </listitem>

                <listitem>
                  <para>
                    On Windows hosts, <command>VBoxAuth.dll</command>
                    authenticates users against the host's WinLogon
                    system.
                  </para>
                </listitem>

                <listitem>
                  <para>
                    On Mac OS X hosts, <command>VBoxAuth.dylib</command>
                    authenticates users against the host's directory
                    service.
                  </para>
                </listitem>

              </itemizedlist>

              <para>
                In other words, the external method by default performs
                authentication with the user accounts that exist on the
                host system. Any user with valid authentication
                credentials is accepted. For example, the username does
                not have to correspond to the user running the VM.
              </para>
            </listitem>

            <listitem>
              <para>
                An additional library called
                <command>VBoxAuthSimple</command> performs
                authentication against credentials configured in the
                <literal>extradata</literal> section of a virtual
                machine's XML settings file. This is probably the
                simplest way to get authentication that does not depend
                on a running and supported guest. The following steps
                are required:
              </para>

              <orderedlist>

                <listitem>
                  <para>
                    Enable <command>VBoxAuthSimple</command> with the
                    following command:
                  </para>

<screen>VBoxManage setproperty vrdeauthlibrary "VBoxAuthSimple"</screen>
                </listitem>

                <listitem>
                  <para>
                    To enable the library for a particular VM, you must
                    switch authentication to external, as follows:
                  </para>

<screen>VBoxManage modifyvm <replaceable>VM-name</replaceable> --vrde-auth-type external</screen>

                  <para>
                    Replace <replaceable>VM-name</replaceable> with the
                    VM name or UUID.
                  </para>
                </listitem>

                <listitem>
                  <para>
                    You then need to configure users and passwords by
                    writing items into the machine's extradata. Since
                    the XML machine settings file, into whose
                    <literal>extradata</literal> section the password
                    needs to be written, is a plain text file,
                    &product-name; uses hashes to encrypt passwords. The
                    following command must be used:
                  </para>

<screen>VBoxManage setextradata <replaceable>VM-name</replaceable> "VBoxAuthSimple/users/<replaceable>user</replaceable>" <replaceable>hash</replaceable></screen>

                  <para>
                    Replace <replaceable>VM-name</replaceable> with the
                    VM name or UUID, <replaceable>user</replaceable>
                    with the user name who should be allowed to log in
                    and <replaceable>hash</replaceable> with the
                    encrypted password. The following command example
                    obtains the hash value for the password
                    <literal>secret</literal>:
                  </para>

<screen>$ VBoxManage internalcommands passwordhash "secret"
2bb80d537b1da3e38bd30361aa855686bde0eacd7162fef6a25fe97bf527a25b</screen>

                  <para>
                    You then use <command>VBoxManage
                    setextradata</command> to store this value in the
                    machine's <literal>extradata</literal> section.
                  </para>

                  <para>
                    As a combined example, to set the password for the
                    user <literal>john</literal> and the machine
                    <literal>My VM</literal> to
                    <literal>secret</literal>, use this command:
                  </para>

<screen>VBoxManage setextradata "My VM" "VBoxAuthSimple/users/john"
    2bb80d537b1da3e38bd30361aa855686bde0eacd7162fef6a25fe97bf527a25b</screen>
                </listitem>

              </orderedlist>
            </listitem>

          </orderedlist>
        </listitem>

        <listitem>
          <para>
            The <emphasis role="bold">guest</emphasis> authentication
            method performs authentication with a special component that
            comes with the Guest Additions. As a result, authentication
            is not performed on the host, but with the guest user
            accounts.
          </para>

          <para>
            This method is currently still in testing and not yet
            supported.
          </para>
        </listitem>

      </itemizedlist>

      <para>
        In addition to the methods described above, you can replace the
        default external authentication module with any other module.
        For this, &product-name; provides a well-defined interface that
        enables you to write your own authentication module. This is
        described in detail in the &product-name; Software Development
        Kit (SDK) reference. See <xref linkend="VirtualBoxAPI" />.
      </para>

    </sect2>

    <sect2 id="vrde-crypt">

      <title>RDP Encryption</title>

      <para>
        RDP features data stream encryption, which is based on the RC4
        symmetric cipher, with keys up to 128-bit. The RC4 keys are
        replaced at regular intervals, every 4096 packets.
      </para>

      <para>
        RDP provides the following different authentication methods:
      </para>

      <itemizedlist>

        <listitem>
          <para>
            <emphasis role="bold">RDP 4</emphasis> authentication was
            used historically. With RDP 4, the RDP client does not
            perform any checks in order to verify the identity of the
            server it connects to. Since user credentials can be
            obtained using a man in the middle (MITM) attack, RDP4
            authentication is insecure and should generally not be used.
          </para>
        </listitem>

        <listitem>
          <para>
            <emphasis role="bold">RDP 5.1</emphasis> authentication
            employs a server certificate for which the client possesses
            the public key. This way it is guaranteed that the server
            possess the corresponding private key. However, as this
            hard-coded private key became public some years ago, RDP 5.1
            authentication is also insecure.
          </para>
        </listitem>

        <listitem>
          <para>
            <emphasis role="bold">RDP 5.2 or later</emphasis>
            authentication uses Enhanced RDP Security, which means that
            an external security protocol is used to secure the
            connection. RDP 4 and RDP 5.1 use Standard RDP Security. The
            VRDP server supports Enhanced RDP Security with TLS protocol
            and, as a part of the TLS handshake, sends the server
            certificate to the client.
          </para>

          <para>
            The <literal>Security/Method</literal> VRDE property sets
            the desired security method, which is used for a connection.
            Valid values are as follows:
          </para>

          <itemizedlist>

            <listitem>
              <para>
                <emphasis role="bold">Negotiate.</emphasis> Both
                Enhanced (TLS) and Standard RDP Security connections are
                allowed. The security method is negotiated with the
                client. This is the default setting.
              </para>
            </listitem>

            <listitem>
              <para>
                <emphasis role="bold">RDP.</emphasis> Only Standard RDP
                Security is accepted.
              </para>
            </listitem>

            <listitem>
              <para>
                <emphasis role="bold">TLS.</emphasis> Only Enhanced RDP
                Security is accepted. The client must support TLS.
              </para>

              <para>
                The version of OpenSSL used by &product-name; supports
                TLS versions 1.0, 1.1, 1.2, and 1.3.
              </para>
            </listitem>

          </itemizedlist>

          <para>
            For example, the following command enables a client to use
            either Standard or Enhanced RDP Security connection:
          </para>

<screen>vboxmanage modifyvm <replaceable>VM-name</replaceable> --vrde-property "Security/Method=negotiate"</screen>

          <para>
            If the <literal>Security/Method</literal> property is set to
            either Negotiate or TLS, the TLS protocol will be
            automatically used by the server, if the client supports
            TLS. However, in order to use TLS the server must possess
            the Server Certificate, the Server Private Key and the
            Certificate Authority (CA) Certificate. The following
            example shows how to generate a server certificate.
          </para>

          <orderedlist>

            <listitem>
              <para>
                Create a CA self signed certificate.
              </para>

<screen>openssl req -new -x509 -days 365 -extensions v3_ca \
  -keyout ca_key_private.pem -out ca_cert.pem</screen>
            </listitem>

            <listitem>
              <para>
                Generate a server private key and a request for signing.
              </para>

<screen>openssl genrsa -out server_key_private.pem
openssl req -new -key server_key_private.pem -out server_req.pem</screen>
            </listitem>

            <listitem>
              <para>
                Generate the server certificate.
              </para>

<screen>openssl x509 -req -days 365 -in server_req.pem \
  -CA ca_cert.pem -CAkey ca_key_private.pem -set_serial 01 -out server_cert.pem</screen>
            </listitem>

          </orderedlist>

          <para>
            The server must be configured to access the required files.
            For example:
          </para>

<screen>vboxmanage modifyvm <replaceable>VM-name</replaceable> \
  --vrde-property "Security/CACertificate=path/ca_cert.pem"</screen>

<screen>vboxmanage modifyvm <replaceable>VM-name</replaceable> \
  --vrde-property "Security/ServerCertificate=path/server_cert.pem"</screen>

<screen>vboxmanage modifyvm <replaceable>VM-name</replaceable> \
  --vrde-property "Security/ServerPrivateKey=path/server_key_private.pem"</screen>
        </listitem>

      </itemizedlist>

      <para>
        As the client that connects to the server determines what type
        of encryption will be used, with <command>rdesktop</command>,
        the Linux RDP viewer, use the <option>-4</option> or
        <option>-5</option> options.
      </para>

    </sect2>

    <sect2 id="vrde-multiconnection">

      <title>Multiple Connections to the VRDP Server</title>

      <para>
        The VRDP server of &product-name; supports multiple simultaneous
        connections to the same running VM from different clients. All
        connected clients see the same screen output and share a mouse
        pointer and keyboard focus. This is similar to several people
        using the same computer at the same time, taking turns at the
        keyboard.
      </para>

      <para>
        The following command enables multiple connection mode:
      </para>

<screen>VBoxManage modifyvm <replaceable>VM-name</replaceable> --vrde-multi-con on</screen>

    </sect2>

    <sect2 id="vrde-multimonitor">

      <title>Multiple Remote Monitors</title>

      <para>
        To access two or more remote VM displays you have to enable the
        VRDP multiconnection mode. See
        <xref linkend="vrde-multiconnection"/>.
      </para>

      <para>
        The RDP client can select the virtual monitor number to connect
        to using the <literal>domain</literal> login parameter
        (<option>-d</option>). If the parameter ends with
        <literal>@</literal> followed by a number, &product-name;
        interprets this number as the screen index. The primary guest
        screen is selected with <literal>@1</literal>, the first
        secondary screen is <literal>@2</literal>, and so on.
      </para>

      <para>
        The Microsoft RDP 6 client does not let you specify a separate
        domain name. Instead, enter
        <literal><replaceable>domain</replaceable>\<replaceable>username</replaceable></literal>
        in the <emphasis role="bold">Username</emphasis> field. For
        example, <literal>@2\<replaceable>name</replaceable></literal>.
        <replaceable>name</replaceable> must be supplied, and must be
        the name used to log in if the VRDP server is set up to require
        credentials. If it is not, you may use any text as the username.
      </para>

    </sect2>

    <sect2 id="vrde-videochannel">

      <title>VRDP Video Redirection</title>

      <para>
        The VRDP server can redirect video streams from the guest to the
        RDP client. Video frames are compressed using the JPEG algorithm
        allowing a higher compression ratio than standard RDP bitmap
        compression methods. It is possible to increase the compression
        ratio by lowering the video quality.
      </para>

      <para>
        The VRDP server automatically detects video streams in a guest
        as frequently updated rectangular areas. As a result, this
        method works with any guest operating system without having to
        install additional software in the guest. In particular, the
        Guest Additions are not required.
      </para>

      <para>
        On the client side, however, currently only the Windows 7 Remote
        Desktop Connection client supports this feature. If a client
        does not support video redirection, the VRDP server falls back
        to regular bitmap updates.
      </para>

      <para>
        The following command enables video redirection:
      </para>

<screen>VBoxManage modifyvm <replaceable>VM-name</replaceable> --vrde-video-channel on</screen>

      <para>
        The quality of the video is defined as a value from 10 to 100
        percent, representing a JPEG compression level, where lower
        numbers mean lower quality but higher compression. The quality
        can be changed using the following command:
      </para>

<screen>VBoxManage modifyvm <replaceable>VM-name</replaceable> --vrde-video-channel-quality 75</screen>

    </sect2>

    <sect2 id="vrde-customization">

      <title>VRDP Customization</title>

      <para>
        You can disable display output, mouse and keyboard input, audio,
        remote USB, or clipboard individually in the VRDP server.
      </para>

      <para>
        The following commands change the corresponding server settings:
      </para>

<screen>$ VBoxManage modifyvm <replaceable>VM-name</replaceable> --vrde-property Client/DisableDisplay=1
$ VBoxManage modifyvm <replaceable>VM-name</replaceable> --vrde-property Client/DisableInput=1
$ VBoxManage modifyvm <replaceable>VM-name</replaceable> --vrde-property Client/DisableUSB=1
$ VBoxManage modifyvm <replaceable>VM-name</replaceable> --vrde-property Client/DisableAudio=1
$ VBoxManage modifyvm <replaceable>VM-name</replaceable> --vrde-property Client/DisableClipboard=1
$ VBoxManage modifyvm <replaceable>VM-name</replaceable> --vrde-property Client/DisableUpstreamAudio=1</screen>

      <para>
        To reenable a feature, use a similar command without the
        trailing 1. For example:
      </para>

<screen>$ VBoxManage modifyvm <replaceable>VM-name</replaceable> --vrde-property Client/DisableDisplay=</screen>

    </sect2>

  </sect1>

  <sect1 id="teleporting">

    <title>Teleporting</title>

    <para>
      &product-name; supports <emphasis>teleporting</emphasis>.
      Teleporting is moving a virtual machine over a network from one
      &product-name; host to another, while the virtual machine is
      running. This works regardless of the host operating system that
      is running on the hosts. You can teleport virtual machines between
      Oracle Solaris and Mac OS X hosts, for example.
    </para>

    <para>
      Teleporting requires that a machine be currently running on one
      host, which is called the <emphasis>source</emphasis>. The host to
      which the virtual machine will be teleported is called the
      <emphasis>target</emphasis>. The machine on the target is then
      configured to wait for the source to contact the target. The
      machine's running state will then be transferred from the source
      to the target with minimal downtime.
    </para>

    <para>
      Teleporting happens over any TCP/IP network. The source and the
      target only need to agree on a TCP/IP port which is specified in
      the teleporting settings.
    </para>

    <para>
      At this time, there are a few prerequisites for this to work, as
      follows:
    </para>

    <itemizedlist>

      <listitem>
        <para>
          On the target host, you must configure a virtual machine in
          &product-name; with exactly the same hardware settings as the
          machine on the source that you want to teleport. This does not
          apply to settings which are merely descriptive, such as the VM
          name, but obviously for teleporting to work, the target
          machine must have the same amount of memory and other hardware
          settings. Otherwise teleporting will fail with an error
          message.
        </para>
      </listitem>

      <listitem>
        <para>
          The two virtual machines on the source and the target must
          share the same storage, hard disks as well as floppy disks and
          CD/DVD images. This means that they either use the same iSCSI
          targets or that the storage resides somewhere on the network
          and both hosts have access to it using NFS or SMB/CIFS.
        </para>

        <para>
          This also means that neither the source nor the target machine
          can have any snapshots.
        </para>
      </listitem>

    </itemizedlist>

    <para>
      To configure teleporting, perform the following steps:
    </para>

    <orderedlist>

      <listitem>
        <para>
          On the <emphasis>target</emphasis> host, configure the virtual
          machine to wait for a teleport request to arrive when it is
          started, instead of actually attempting to start the machine.
          This is done with the following <command>VBoxManage</command>
          command:
        </para>

<screen>VBoxManage modifyvm <replaceable>targetvmname</replaceable> --teleporter on --teleporter-port <replaceable>port</replaceable></screen>

        <para>
          <replaceable>targetvmname</replaceable> is the name of the
          virtual machine on the target host and
          <replaceable>port</replaceable> is a TCP/IP port number to be
          used on both the source and the target hosts. For example, use
          6000. See <xref linkend="vboxmanage-modifyvm" />.
        </para>
      </listitem>

      <listitem>
        <para>
          Start the VM on the target host. Instead of running, the VM
          shows a progress dialog, indicating that it is waiting for a
          teleport request to arrive.
        </para>
      </listitem>

      <listitem>
        <para>
          Start the VM on the <emphasis>source</emphasis> host as usual.
          When it is running and you want it to be teleported, issue the
          following command on the source host:
        </para>

<screen>VBoxManage controlvm <replaceable>sourcevmname</replaceable> teleport --host <replaceable>targethost</replaceable> --port <replaceable>port</replaceable></screen>

        <para>
          where <replaceable>sourcevmname</replaceable> is the name of
          the virtual machine on the source host, which is the machine
          that is currently running.
          <replaceable>targethost</replaceable> is the host or IP name
          of the target host on which the machine is waiting for the
          teleport request, and <replaceable>port</replaceable> must be
          the same number as specified in the command on the target
          host. See <xref linkend="vboxmanage-controlvm" />.
        </para>
      </listitem>

    </orderedlist>

    <para>
      For testing, you can also teleport machines on the same host. In
      that case, use localhost as the hostname on both the source and
      the target host.
    </para>

    <note>
      <para>
        In rare cases, if the CPUs of the source and the target are very
        different, teleporting can fail with an error message, or the
        target may hang. This may happen especially if the VM is running
        application software that is highly optimized to run on a
        particular CPU without correctly checking that certain CPU
        features are actually present. &product-name; filters what CPU
        capabilities are presented to the guest operating system.
        Advanced users can attempt to restrict these virtual CPU
        capabilities with the <command>VBoxManage modifyvm
        --cpuid-portability-level</command> command. See
        <xref linkend="vboxmanage-modifyvm" />.
      </para>
    </note>

  </sect1>

  <xi:include href="user_man_VBoxHeadless.xml" xpointer="element(/1)" xmlns:xi="http://www.w3.org/2001/XInclude" />

</chapter>