vtep: Initial checkin of vtep schema.

[cascardo/ovs.git] / vtep / vtep.xml

<?xml version="1.0" encoding="utf-8"?>
<database title="Hardware VTEP Database">
  <p>
    This schema specifies relations that a VTEP can use to integrate
    physical ports into logical switches maintained by a network
    virtualization controller such as NSX.
  </p>
  
  <p>Glossary:</p>

  <dl>
    <dt>VTEP</dt>
    <dd>
      VXLAN Tunnel End Point, an entity which originates and/or terminates
      VXLAN tunnels.
    </dd>

    <dt>HSC</dt>
    <dd>
      Hardware Switch Controller.
    </dd>

    <dt>NVC</dt>
    <dd>
      Network Virtualization Controller, e.g. NSX.
    </dd>

    <dt>VRF</dt>
    <dd>
      Virtual Routing and Forwarding instance.
    </dd>
 </dl>

  <table name="Global" title="Top-level configuration.">
    Top-level configuration for a hardware VTEP.  There must be
    exactly one record in the <ref table="Global"/> table.

    <column name="switches">
      The physical switches managed by the VTEP.
    </column>

    <group title="Database Configuration">
      <p>
        These columns primarily configure the database server
        (<code>ovsdb-server</code>), not the hardware VTEP itself.
      </p>

      <column name="managers">
        Database clients to which the database server should connect or
        to which it should listen, along with options for how these
        connection should be configured.  See the <ref table="Manager"/>
        table for more information.
      </column>
    </group>
  </table>

  <table name="Manager" title="OVSDB management connection.">
    <p>
      Configuration for a database connection to an Open vSwitch Database
      (OVSDB) client.
    </p>

    <p>
      The database server can initiate and maintain active connections
      to remote clients.  It can also listen for database connections.
    </p>

    <group title="Core Features">
      <column name="target">
        <p>Connection method for managers.</p>
        <p>
          The following connection methods are currently supported:
        </p>
        <dl>
          <dt><code>ssl:<var>ip</var></code>[<code>:<var>port</var></code>]</dt>
          <dd>
            <p>
              The specified SSL <var>port</var> (default: 6632) on the host at
              the given <var>ip</var>, which must be expressed as an IP address
              (not a DNS name).
            </p>
            <p>
              SSL key and certificate configuration happens outside the
              database.
            </p>
          </dd>

          <dt><code>tcp:<var>ip</var></code>[<code>:<var>port</var></code>]</dt>
          <dd>
            The specified TCP <var>port</var> (default: 6632) on the host at
            the given <var>ip</var>, which must be expressed as an IP address
            (not a DNS name).
          </dd>
          <dt><code>pssl:</code>[<var>port</var>][<code>:<var>ip</var></code>]</dt>
          <dd>
            <p>
              Listens for SSL connections on the specified TCP <var>port</var>
              (default: 6632).  If <var>ip</var>, which must be expressed as an
              IP address (not a DNS name), is specified, then connections are
              restricted to the specified local IP address.
            </p>
          </dd>
          <dt><code>ptcp:</code>[<var>port</var>][<code>:<var>ip</var></code>]</dt>
          <dd>
            Listens for connections on the specified TCP <var>port</var>
            (default: 6632).  If <var>ip</var>, which must be expressed as an
            IP address (not a DNS name), is specified, then connections are
            restricted to the specified local IP address.
          </dd>
        </dl>
      </column>
    </group>

    <group title="Client Failure Detection and Handling">
      <column name="max_backoff">
        Maximum number of milliseconds to wait between connection attempts.
        Default is implementation-specific.
      </column>

      <column name="inactivity_probe">
        Maximum number of milliseconds of idle time on connection to the
        client before sending an inactivity probe message.  If the Open
        vSwitch database does not communicate with the client for the
        specified number of seconds, it will send a probe.  If a
        response is not received for the same additional amount of time,
        the database server assumes the connection has been broken
        and attempts to reconnect.  Default is implementation-specific.
        A value of 0 disables inactivity probes.
      </column>
    </group>

    <group title="Status">
      <column name="is_connected">
        <code>true</code> if currently connected to this manager,
        <code>false</code> otherwise.
      </column>

      <column name="status" key="last_error">
        A human-readable description of the last error on the connection
        to the manager; i.e. <code>strerror(errno)</code>.  This key
        will exist only if an error has occurred.
      </column>

      <column name="status" key="state"
              type='{"type": "string", "enum": ["set", ["VOID", "BACKOFF", "CONNECTING", "ACTIVE", "IDLE"]]}'>
        <p>
          The state of the connection to the manager:
        </p>
        <dl>
          <dt><code>VOID</code></dt>
          <dd>Connection is disabled.</dd>

          <dt><code>BACKOFF</code></dt>
          <dd>Attempting to reconnect at an increasing period.</dd>

          <dt><code>CONNECTING</code></dt>
          <dd>Attempting to connect.</dd>

          <dt><code>ACTIVE</code></dt>
          <dd>Connected, remote host responsive.</dd>

          <dt><code>IDLE</code></dt>
          <dd>Connection is idle.  Waiting for response to keep-alive.</dd>
        </dl>
        <p>
          These values may change in the future.  They are provided only for
          human consumption.
        </p>
      </column>

      <column name="status" key="sec_since_connect"
              type='{"type": "integer", "minInteger": 0}'>
        The amount of time since this manager last successfully connected
        to the database (in seconds). Value is empty if manager has never
        successfully connected.
      </column>

      <column name="status" key="sec_since_disconnect"
              type='{"type": "integer", "minInteger": 0}'>
        The amount of time since this manager last disconnected from the
        database (in seconds). Value is empty if manager has never
        disconnected.
      </column>

      <column name="status" key="locks_held">
        Space-separated list of the names of OVSDB locks that the connection
        holds.  Omitted if the connection does not hold any locks.
      </column>

      <column name="status" key="locks_waiting">
        Space-separated list of the names of OVSDB locks that the connection is
        currently waiting to acquire.  Omitted if the connection is not waiting
        for any locks.
      </column>

      <column name="status" key="locks_lost">
        Space-separated list of the names of OVSDB locks that the connection
        has had stolen by another OVSDB client.  Omitted if no locks have been
        stolen from this connection.
      </column>

      <column name="status" key="n_connections"
              type='{"type": "integer", "minInteger": 2}'>
        <p>
          When <ref column="target"/> specifies a connection method that
          listens for inbound connections (e.g. <code>ptcp:</code> or
          <code>pssl:</code>) and more than one connection is actually active,
          the value is the number of active connections.  Otherwise, this
          key-value pair is omitted.
        </p>
        <p>
          When multiple connections are active, status columns and key-value
          pairs (other than this one) report the status of one arbitrarily
          chosen connection.
        </p>
      </column>
    </group>

    <group title="Connection Parameters">
      <p>
        Additional configuration for a connection between the manager
        and the database server.
      </p>

      <column name="other_config" key="dscp"
                type='{"type": "integer"}'>
        The Differentiated Service Code Point (DSCP) is specified using 6 bits
        in the Type of Service (TOS) field in the IP header. DSCP provides a
        mechanism to classify the network traffic and provide Quality of
        Service (QoS) on IP networks.

        The DSCP value specified here is used when establishing the
        connection between the manager and the database server.  If no
        value is specified, a default value of 48 is chosen.  Valid DSCP
        values must be in the range 0 to 63.
      </column>
    </group>
  </table>

  <table name="Physical_Switch" title="A physical switch.">
    A physical switch that implements a VTEP.

    <column name="ports">
      The physical ports within the switch.
    </column>

    <group title="Network Status">
      <column name="management_ips">
        IPv4 or IPv6 addresses at which the switch may be contacted
        for management purposes.
      </column>

      <column name="tunnel_ips">
        <p>
          IPv4 or IPv6 addresses on which the switch may originate or
          terminate tunnels.
        </p>

        <p>
          This column is intended to allow a <ref table="Manager"/> to
          determine the <ref table="Physical_Switch"/> that terminates
          the tunnel represented by a <ref table="Physical_Locator"/>.
        </p>
      </column>
    </group>

    <group title="Identification">
      <column name="name">
        Symbolic name for the switch, such as its hostname.
      </column>
      
      <column name="description">
        An extended description for the switch, such as its switch login
        banner.
      </column>
    </group>
  </table>

  <table name="Physical_Port" title="A port within a physical switch.">
    A port within a <ref table="Physical_Switch"/>.

    <column name="vlan_bindings">
      Identifies how VLANs on the physical port are bound to logical switches.
      If, for example, the map contains a (VLAN, logical switch) pair, a packet
      that arrives on the port in the VLAN is considered to belong to the
      paired logical switch.
    </column>

    <column name="vlan_stats">
      Statistics for VLANs bound to logical switches on the physical port.  An
      implementation that fully supports such statistics would populate this
      column with a mapping for every VLAN that is bound in <ref
      column="vlan_bindings"/>.  An implementation that does not support such
      statistics or only partially supports them would not populate this column
      or partially populate it, respectively.
    </column>

    <group title="Identification">
      <column name="name">
        Symbolic name for the port.  The name ought to be unique within a given
        <ref table="Physical_Switch"/>, but the database is not capable of
        enforcing this.
      </column>
      
      <column name="description">
        An extended description for the port.
      </column>
    </group>
  </table>

  <table name="Logical_Binding_Stats" title="Statistics for a VLAN on a physical port bound to a logical network.">
    Reports statistics for the <ref table="Logical_Switch"/> with which a VLAN
    on a <ref table="Physical_Port"/> is associated.

    <group title="Statistics">
      These statistics count only packets to which the binding applies.

      <column name="packets_from_local">
        Number of packets sent by the <ref table="Physical_Switch"/>.
      </column>

      <column name="bytes_from_local">
        Number of bytes in packets sent by the <ref table="Physical_Switch"/>.
      </column>

      <column name="packets_to_local">
        Number of packets received by the <ref table="Physical_Switch"/>.
      </column>

      <column name="bytes_to_local">
        Number of bytes in packets received by the <ref
        table="Physical_Switch"/>.
      </column>
    </group>
  </table>

  <table name="Logical_Switch" title="A layer-2 domain.">
    A logical Ethernet switch, whose implementation may span physical and
    virtual media, possibly crossing L3 domains via tunnels; a logical layer-2
    domain; an Ethernet broadcast domain.



    <group title="Per Logical-Switch Tunnel Key">
      <p>
        Tunnel protocols tend to have a field that allows the tunnel
        to be partitioned into sub-tunnels: VXLAN has a VNI, GRE and
        STT have a key, CAPWAP has a WSI, and so on.  We call these
        generically ``tunnel keys.''  Given that one needs to use a
        tunnel key at all, there are at least two reasonable ways to
        assign their values:
      </p>

      <ul>
        <li>
          <p>
            Per <ref table="Logical_Switch"/>+<ref table="Physical_Locator"/>
            pair.  That is, each logical switch may be assigned a different
            tunnel key on every <ref table="Physical_Locator"/>.  This model is
            especially flexible.
          </p>

          <p>
            In this model, <ref table="Physical_Locator"/> carries the tunnel
            key.  Therefore, one <ref table="Physical_Locator"/> record will
            exist for each logical switch carried at a given IP destination.
          </p>
        </li>

        <li>
          <p>
            Per <ref table="Logical_Switch"/>.  That is, every tunnel
            associated with a particular logical switch carries the same tunnel
            key, regardless of the <ref table="Physical_Locator"/> to which the
            tunnel is addressed.  This model may ease switch implementation
            because it imposes fewer requirements on the hardware datapath.
          </p>

          <p>
            In this model, <ref table="Logical_Switch"/> carries the tunnel
            key.  Therefore, one <ref table="Physical_Locator"/> record will
            exist for each IP destination.
          </p>
        </li>
      </ul>

      <column name="tunnel_key">
        <p>
          This column is used only in the tunnel key per <ref
          table="Logical_Switch"/> model (see above), because only in that
          model is there a tunnel key associated with a logical switch.
        </p>

        <p>
          For <code>vxlan_over_ipv4</code> encapsulation, this column
          is the VXLAN VNI that identifies a logical switch.  It must
          be in the range 0 to 16,777,215.
        </p>
      </column>
    </group>

    <group title="Identification">
      <column name="name">
        Symbolic name for the logical switch.
      </column>
      
      <column name="description">
        An extended description for the logical switch, such as its switch
        login banner.
      </column>
    </group>
  </table>

  <table name="Ucast_Macs_Local" title="Unicast MACs (local)">
    <p>
      Mapping of unicast MAC addresses to tunnels (physical
      locators). This table is written by the HSC, so it contains the
      MAC addresses that have been learned on physical ports by a
      VTEP. 
    </p>

    <column name="MAC">
      A MAC address that has been learned by the VTEP.
    </column>

    <column name="logical_switch">
      The Logical switch to which this mapping applies.
    </column>

    <column name="locator">
      The physical locator to be used to reach this MAC address. In
      this table, the physical locator will be one of the tunnel IP
      addresses of the appropriate VTEP.
    </column>

    <column name="ipaddr">
      The IP address to which this MAC corresponds. Optional field for
      the purpose of ARP supression.
    </column>

  </table>

 <table name="Ucast_Macs_Remote" title="Unicast MACs (remote)">
    <p>
      Mapping of unicast MAC addresses to tunnels (physical
      locators). This table is written by the NVC, so it contains the
      MAC addresses that the NVC has learned. These include VM MAC
      addresses, in which case the physical locators will be
      hypervisor IP addresses. The NVC will also report MACs that it
      has learned from other HSCs in the network, in which case the
      physical locators will be tunnel IP addresses of the
      corresponding VTEPs.
    </p>

    <column name="MAC">
      A MAC address that has been learned by the NSC.
    </column>

    <column name="logical_switch">
      The Logical switch to which this mapping applies.
    </column>

    <column name="locator">
      The physical locator to be used to reach this MAC address. In
      this table, the physical locator will be either a hypervisor IP
      address or a tunnel IP addresses of another VTEP.
    </column>

    <column name="ipaddr">
      The IP address to which this MAC corresponds. Optional field for
      the purpose of ARP supression.
    </column>

  </table>

  <table name="Mcast_Macs_Local" title="Multicast MACs (local)">
    <p>
      Mapping of multicast MAC addresses to tunnels (physical
      locators). This table is written by the HSC, so it contains the
      MAC addresses that have been learned on physical ports by a
      VTEP. These may be learned by IGMP snooping, for example. This
      table also specifies how to handle unknown unicast and broadcast packets.
    </p>

    <column name="MAC">
      <p>
        A MAC address that has been learned by the VTEP. 
      </p>
      <p>
        The keyword <code>unknown-dst</code> is used as a special
        ``Ethernet address'' that indicates the locations to which
        packets in a logical switch whose destination addresses do not
        otherwise appear in <ref table="Ucast_Macs_Local"/> (for
        unicast addresses) or <ref table="Mcast_Macs_Local"/> (for
        multicast addresses) should be sent.
      </p>
    </column>

    <column name="logical_switch">
      The Logical switch to which this mapping applies.
    </column>

    <column name="locator_set">
      The physical locator set to be used to reach this MAC address. In
      this table, the physical locator set will be contain one or more tunnel IP
      addresses of the appropriate VTEP(s).
    </column>

  </table>

  <table name="Mcast_Macs_Remote" title="Multicast MACs (remote)">
    <p>
      Mapping of multicast MAC addresses to tunnels (physical
      locators). This table is written by the NVC, so it contains the
      MAC addresses that the NVC has learned. This
      table also specifies how to handle unknown unicast and broadcast
      packets.
    </p>
    <p>
      Multicast packet replication may be handled by a service node,
      in which case the physical locators will be IP addresses of
      service nodes. If the VTEP supports replication onto multiple
      tunnels, then this may be used to replicate directly onto
      VTEP-hyperisor tunnels.
    </p>

    <column name="MAC">
      <p>
        A MAC address that has been learned by the NSC.
      </p>
      <p>
        The keyword <code>unknown-dst</code> is used as a special
        ``Ethernet address'' that indicates the locations to which
        packets in a logical switch whose destination addresses do not
        otherwise appear in <ref table="Ucast_Macs_Remote"/> (for
        unicast addresses) or <ref table="Mcast_Macs_Remote"/> (for
        multicast addresses) should be sent.
      </p>
    </column>

    <column name="logical_switch">
      The Logical switch to which this mapping applies.
    </column>

    <column name="locator_set">
      The physical locator set to be used to reach this MAC address. In
      this table, the physical locator set will be either a service node IP
      address or a set of tunnel IP addresses of hypervisors (and
      potentially other VTEPs).
    </column>

    <column name="ipaddr">
      The IP address to which this MAC corresponds. Optional field for
      the purpose of ARP supression.
    </column>

  </table>

  <table name="Logical_Router" title="A logical L3 router.">
    <p>
      A logical router, or VRF. A logical router may be connected to one or more
      logical switches. Subnet addresses and interface addresses may be configured on the 
      interfaces.
    </p>
    
    <column name="switch_binding">
      Maps from an IPv4 or IPv6 address prefix in CIDR notation to a
      logical switch. Multiple prefixes may map to the same switch. By
      writing a 32-bit (or 128-bit for v6) address with a /N prefix
      length, both the router's interface address and the subnet
      prefix can be configured. For example, 192.68.1.1/24 creates a
      /24 subnet for the logical switch attached to the interface and
      assigns the address 192.68.1.1 to the router interface.
    </column>

    <column name="static_routes">
      One or more static routes, mapping IP prefixes to next hop IP addresses.
    </column>

    <group title="Identification">
      <column name="name">
        Symbolic name for the logical router.
      </column>
      
      <column name="description">
        An extended description for the logical router.
      </column>
    </group>
  </table>

  <table name="Physical_Locator_Set">
    <p>
      A set of one or more <ref table="Physical_Locator"/>s.
    </p>

    <p>
      This table exists only because OVSDB does not have a way to
      express the type ``map from string to one or more <ref
      table="Physical_Locator"/> records.''
    </p>

    <column name="locators"/>    
  </table>

  <table name="Physical_Locator">
    <p>
      Identifies an endpoint to which logical switch traffic may be
      encapsulated and forwarded.
    </p>

    <p>
      For the <code>vxlan_over_ipv4</code> encapsulation, the only
      encapsulation defined so far, all endpoints associated with a given <ref
      table="Logical_Switch"/> must use a common tunnel key, which is carried
      in the <ref table="Logical_Switch" column="tunnel_key"/> column of <ref
      table="Logical_Switch"/>.
    </p>

    <p>
      For some encapsulations yet to be defined, we expect <ref
      table="Physical_Locator"/> to identify both an endpoint and a tunnel key.
      When the first such encapsulation is defined, we expect to add a
      ``tunnel_key'' column to <ref table="Physical_Locator"/> to allow the
      tunnel key to be defined.
    </p>

    <p>
      See the ``Per Logical-Switch Tunnel Key'' section in the <ref
      table="Logical_Switch"/> table for further discussion of the model.
    </p>

    <column name="encapsulation_type">
      The type of tunneling encapsulation.
    </column>

    <column name="dst_ip">
      <p>
        For <code>vxlan_over_ipv4</code> encapsulation, the IPv4 address of the
        VXLAN tunnel endpoint.
      </p>

      <p>
        We expect that this column could be used for IPv4 or IPv6 addresses in
        encapsulations to be introduced later.
      </p>
    </column>
    <group title="Bidirectional Forwarding Detection (BFD)">
      <p>
        BFD, defined in RFC 5880, allows point to point detection of
        connectivity failures by occasional transmission of BFD control
        messages.
      </p>
      
      <p>
        BFD operates by regularly transmitting BFD control messages at a
        rate negotiated independently in each direction.  Each endpoint
        specifies the rate at which it expects to receive control messages,
        and the rate at which it's willing to transmit them.  An endpoint
        which fails to receive BFD control messages for a period of three
        times the expected reception rate, will signal a connectivity
        fault.  In the case of a unidirectional connectivity issue, the
        system not receiving BFD control messages will signal the problem
        to its peer in the messages is transmists.
      </p>
      
      <column name="bfd" key="min_rx">
        The minimum rate, in milliseconds, at which this BFD session is
        willing to receive BFD control messages.  The actual rate may slower
        if the remote endpoint isn't willing to transmit as quickly as
        specified.  Defaults to <code>1000</code>.
      </column>
      
      <column name="bfd" key="min_tx">
        The minimum rate, in milliseconds, at which this BFD session is
        willing to transmit BFD control messages.  The actual rate may be
        slower if the remote endpoint isn't willing to receive as quickly as
        specified.  Defaults to <code>100</code>.
      </column>
      
      <column name="bfd" key="cpath_down">
        Concatenated path down may be used when the local system should not
        have traffic forwarded to it for some reason other than a connectivty
        failure on the interface being monitored.  The local BFD session will
        notify the remote session of the connectivity problem, at which time
        the remote session may choose not to forward traffic.  Defaults to
        <code>false</code>.
      </column>
      
      <column name="bfd_status" key="state">
        State of the BFD session.  One of <code>ADMIN_DOWN</code>,
        <code>DOWN</code>, <code>INIT</code>, or <code>UP</code>.
      </column>
      
      <column name="bfd_status" key="forwarding">
        True if the BFD session believes this <ref table="Physical_Locator"/> may be
        used to forward traffic.  Typically this means the local session is
        up, and the remote system isn't signalling a problem such as
        concatenated path down.
      </column>
      
      <column name="bfd_status" key="diagnostic">
        A short message indicating what the BFD session thinks is wrong in
        case of a problem.
      </column>
      
      <column name="bfd_status" key="remote state">
        State of the remote endpoint's BFD session.
      </column>
      
      <column name="bfd_status" key="remote diagnostic">
        A short message indicating what the remote endpoint's BFD session
        thinks is wrong in case of a problem.
      </column>
    </group>
  </table>

</database>