A database with this schema holds the configuration for one Open vSwitch daemon. The top-level configuration for the daemon is the table, which must have exactly one record. Records in other tables are significant only when they can be reached directly or indirectly from the table. Records that are not reachable from the table are automatically deleted from the database, except for records in a few distinguished ``root set'' tables noted below.

Configuration for an Open vSwitch daemon. There must be exactly one record in the table. Set of bridges managed by the daemon. SSL used globally by the daemon. Key-value pairs for use by external frameworks that integrate with Open vSwitch, rather than by Open vSwitch itself. System integrators should either use the Open vSwitch development mailing list to coordinate on common key-value definitions, or choose key names that are likely to be unique. The currently defined common key-value pairs are:

system-id: A unique identifier for the Open vSwitch's physical host. The form of the identifier depends on the type of the host. On a Citrix XenServer, this will likely be the same as xs-system-uuid.
xs-system-uuid: The Citrix XenServer universally unique identifier for the physical host as displayed by xe host-list.

Sequence number for client to increment. When a client modifies any part of the database configuration and wishes to wait for Open vSwitch to finish applying the changes, it may increment this sequence number. Sequence number that Open vSwitch sets to the current value of after it finishes applying a set of configuration changes. Describes functionality supported by the hardware and software platform on which this Open vSwitch is based. Clients should not modify this column. See the description for defined capability categories and the meaning of associated records.

Key-value pairs that report statistics about a system running an Open vSwitch. These are updated periodically (currently, every 5 seconds). Key-value pairs that cannot be determined or that do not apply to a platform are omitted.

cpu

Number of CPU processors, threads, or cores currently online and available to the operating system on which Open vSwitch is running, as an integer. This may be less than the number installed, if some are not online or if they are not available to the operating system.

Open vSwitch userspace processes are not multithreaded, but the Linux kernel-based datapath is.

load_average

A comma-separated list of three floating-point numbers, representing the system load average over the last 1, 5, and 15 minutes, respectively.

memory

A comma-separated list of integers, each of which represents a quantity of memory in kilobytes that describes the operating system on which Open vSwitch is running. In respective order, these values are:

Total amount of RAM allocated to the OS.
RAM allocated to the OS that is in use.
RAM that can be flushed out to disk or otherwise discarded if that space is needed for another purpose. This number is necessarily less than or equal to the previous value.
Total disk space allocated for swap.
Swap space currently in use.

On Linux, all five values can be determined and are included. On other operating systems, only the first two values can be determined, so the list will only have two values.

process_name

One such key-value pair will exist for each running Open vSwitch daemon process, with name replaced by the daemon's name (e.g. process_ovs-vswitchd). The value is a comma-separated list of integers. The integers represent the following, with memory measured in kilobytes and durations in milliseconds:

The process's virtual memory size.
The process's resident set size.
The amount of user and system CPU time consumed by the process.
The number of times that the process has crashed and been automatically restarted by the monitor.
The duration since the process was started.
The duration for which the process has been running.

The interpretation of some of these values depends on whether the process was started with the . If it was not, then the crash count will always be 0 and the two durations will always be the same. If was given, then the crash count may be positive; if it is, the latter duration is the amount of time since the most recent crash and restart.

There will be one key-value pair for each file in Open vSwitch's ``run directory'' (usually /var/run/openvswitch) whose name ends in .pid, whose contents are a process ID, and which is locked by a running process. The name is taken from the pidfile's name.

Currently Open vSwitch is only able to obtain all of the above detail on Linux systems. On other systems, the same key-value pairs will be present but the values will always be the empty string.

file_systems

A space-separated list of information on local, writable file systems. Each item in the list describes one file system and consists in turn of a comma-separated list of the following:

Mount point, e.g. / or /var/log. Any spaces or commas in the mount point are replaced by underscores.
Total size, in kilobytes, as an integer.
Amount of storage in use, in kilobytes, as an integer.

This key-value pair is omitted if there are no local, writable file systems or if Open vSwitch cannot obtain the needed information.

These columns report the types and versions of the hardware and software running Open vSwitch. We recommend in general that software should test whether specific features are supported instead of relying on version number checks. These values are primarily intended for reporting to human administrators.

The Open vSwitch version number, e.g. 1.1.0. If Open vSwitch was configured with a build number, then it is also included, e.g. 1.1.0+build6579.

The database schema version number in the form major.minor.tweak, e.g. 1.2.3. Whenever the database schema is changed in a non-backward compatible way (e.g. deleting a column or a table), major is incremented. When the database schema is changed in a backward compatible way (e.g. adding a new column), minor is incremented. When the database schema is changed cosmetically (e.g. reindenting its syntax), tweak is incremented.

The schema version is part of the database schema, so it can also be retrieved by fetching the schema using the Open vSwitch database protocol.

An identifier for the type of system on top of which Open vSwitch runs, e.g. XenServer or KVM.

System integrators are responsible for choosing and setting an appropriate value for this column.

The version of the system identified by , e.g. 5.6.100-39265p on XenServer 5.6.100 build 39265.

System integrators are responsible for choosing and setting an appropriate value for this column.

These columns primarily configure the Open vSwitch database (ovsdb-server), not the Open vSwitch switch (ovs-vswitchd). The OVSDB database also uses the settings.

The Open vSwitch switch does read the database configuration to determine remote IP addresses to which in-band control should apply.

Database clients to which the Open vSwitch database server should connect or to which it should listen, along with options for how these connection should be configured. See the table for more information.

Configuration for a bridge within an .

A record represents an Ethernet switch with one or more ``ports,'' which are the records pointed to by the 's column.

Bridge identifier. Should be alphanumeric and no more than about 8 bytes long. Must be unique among the names of ports, interfaces, and bridges on a host. Ports included in the bridge. Port mirroring configuration. NetFlow configuration. sFlow configuration. VLAN IDs of VLANs on which MAC address learning should be disabled, so that packets are flooded instead of being sent to specific ports that are believed to contain packets' destination MACs. This should ordinarily be used to disable MAC learning on VLANs used for mirroring (RSPAN VLANs). It may also be useful for debugging. OpenFlow controller set. If unset, then no OpenFlow controllers will be used.

When a controller is configured, it is, ordinarily, responsible for setting up all flows on the switch. Thus, if the connection to the controller fails, no new network connections can be set up. If the connection to the controller stays down long enough, no packets can pass through the switch at all. This setting determines the switch's response to such a situation. It may be set to one of the following:

standalone: If no message is received from the controller for three times the inactivity probe interval (see ), then Open vSwitch will take over responsibility for setting up flows. In this mode, Open vSwitch causes the bridge to act like an ordinary MAC-learning switch. Open vSwitch will continue to retry connecting to the controller in the background and, when the connection succeeds, it will discontinue its standalone behavior.
secure: Open vSwitch will not set up flows on its own when the controller connection fails or when no controllers are defined. The bridge will continue to retry connecting to any defined controllers forever.

If this value is unset, the default is implementation-specific.

When more than one controller is configured, is considered only when none of the configured controllers can be contacted.

Reports the OpenFlow datapath ID in use. Exactly 16 hex digits. (Setting this column will have no useful effect. Set :other-config instead.) Name of datapath provider. The kernel datapath has type system. The userspace datapath has type netdev. Key-value pairs for use by external frameworks that integrate with Open vSwitch, rather than by Open vSwitch itself. System integrators should either use the Open vSwitch development mailing list to coordinate on common key-value definitions, or choose key names that are likely to be unique. The currently defined key-value pairs are:

bridge-id: A unique identifier of the bridge. On Citrix XenServer this will commonly be the same as xs-network-uuids.
xs-network-uuids: Semicolon-delimited set of universally unique identifier(s) for the network with which this bridge is associated on a Citrix XenServer host. The network identifiers are RFC 4122 UUIDs as displayed by, e.g., xe network-list.

Key-value pairs for configuring rarely used bridge features. The currently defined key-value pairs are:

datapath-id: Exactly 16 hex digits to set the OpenFlow datapath ID to a specific value. May not be all-zero.
disable-in-band: If set to true, disable in-band control on the bridge regardless of controller and manager settings.
hwaddr: An Ethernet address in the form xx:xx:xx:xx:xx:xx to set the hardware address of the local port and influence the datapath ID.
in-band-queue: A queue ID as a nonnegative integer. This sets the OpenFlow queue ID that will be used by flows set up by in-band control on this bridge. If unset, or if the port used by an in-band control flow does not have QoS configured, or if the port does not have a queue with the specified ID, the default queue is used instead.

A port within a .

Most commonly, a port has exactly one ``interface,'' pointed to by its column. Such a port logically corresponds to a port on a physical Ethernet switch. A port with more than one interface is a ``bonded port'' (see ).

Some properties that one might think as belonging to a port are actually part of the port's members.

Port name. Should be alphanumeric and no more than about 8 bytes long. May be the same as the interface name, for non-bonded ports. Must otherwise be unique among the names of ports, interfaces, and bridges on a host. The port's interfaces. If there is more than one, this is a bonded Port.

A bridge port must be configured for VLANs in one of two mutually exclusive ways:

A ``trunk port'' has an empty value for . Its value may be empty or non-empty.
An ``implicitly tagged VLAN port'' or ``access port'' has an nonempty value for . Its value must be empty.

If and are both nonempty, the configuration is ill-formed.

If this is an access port (see above), the port's implicitly tagged VLAN. Must be empty if this is a trunk port.

Frames arriving on trunk ports will be forwarded to this port only if they are tagged with the given VLAN (or, if is 0, then if they lack a VLAN header). Frames arriving on other access ports will be forwarded to this port only if they have the same value. Frames forwarded to this port will not have an 802.1Q header.

When a frame with a 802.1Q header that indicates a nonzero VLAN is received on an access port, it is discarded.

If this is a trunk port (see above), the 802.1Q VLAN(s) that this port trunks; if it is empty, then the port trunks all VLANs. Must be empty if this is an access port.

Frames arriving on trunk ports are dropped if they are not in one of the specified VLANs. For this purpose, packets that have no VLAN header are treated as part of VLAN 0.

A port that has more than one interface is a ``bonded port.'' Bonding allows for load balancing and fail-over. Some kinds of bonding will work with any kind of upstream switch:

balance-slb: Balances flows among slaves based on source MAC address and output VLAN, with periodic rebalancing as traffic patterns change.
active-backup: Assigns all flows to one slave, failing over to a backup slave when the active slave is disabled.

The following mode requires the upstream switch to support 802.3ad with successful LACP negotiation. If LACP negotiation fails then balance-slb mode is used as a fallback:

balance-tcp: Balances flows among slaves based on L2, L3, and L4 protocol information such as destination MAC address, IP address, and TCP port.

These columns apply only to bonded ports. Their values are otherwise ignored.

The type of bonding used for a bonded port. Defaults to balance-slb if unset.

For a bonded port, the number of milliseconds for which carrier must stay up on an interface before the interface is considered to be up. Specify 0 to enable the interface immediately.

This setting is honored only when at least one bonded interface is already enabled. When no interfaces are enabled, then the first bond interface to come up is enabled immediately.

For a bonded port, the number of milliseconds for which carrier must stay down on an interface before the interface is considered to be down. Specify 0 to disable the interface immediately. For a bonded port, whether to create a fake internal interface with the name of the port. Use only for compatibility with legacy software that requires this.

Configures LACP on this port. LACP allows directly connected switches to negotiate which links may be bonded. LACP may be enabled on non-bonded ports for the benefit of any switches they may be connected to. active ports are allowed to initiate LACP negotiations. passive ports are allowed to participate in LACP negotiations initiated by a remote switch, but not allowed to initiate such negotiations themselves. If unset Open vSwitch will choose a reasonable default.

Quality of Service configuration for this port. The MAC address to use for this port for the purpose of choosing the bridge's MAC address. This column does not necessarily reflect the port's actual MAC address, nor will setting it change the port's actual MAC address. Does this port represent a sub-bridge for its tagged VLAN within the Bridge? See ovs-vsctl(8) for more information.

Key-value pairs for use by external frameworks that integrate with Open vSwitch, rather than by Open vSwitch itself. System integrators should either use the Open vSwitch development mailing list to coordinate on common key-value definitions, or choose key names that are likely to be unique.

No key-value pairs native to are currently defined. For fake bridges (see the column), external IDs for the fake bridge are defined here by prefixing a key with fake-bridge-, e.g. fake-bridge-xs-network-uuids.

Key-value pairs for configuring rarely used port features. The currently defined key-value pairs are:

hwaddr: An Ethernet address in the form xx:xx:xx:xx:xx:xx.
bond-rebalance-interval: For an SLB bonded port, the number of milliseconds between successive attempts to rebalance the bond, that is, to move source MACs and their flows from one interface on the bond to another in an attempt to keep usage of each interface roughly equal. The default is 10000 (10 seconds), and the minimum is 1000 (1 second).
bond-detect-mode: Sets the method used to detect link failures in a bonded port. Options are carrier and miimon. Defaults to carrier which uses each interface's carrier to detect failures. When set to miimon, will check for failures by polling each interface's MII.
bond-miimon-interval: The number of milliseconds between successive attempts to poll each interface's MII. Only relevant on ports which use miimon to detect failures.
lacp-system-priority: The LACP system priority of this . In LACP negotiations, link status decisions are made by the system with the numerically lower priority. Must be a number between 1 and 65535.
lacp-time: The LACP timing which should be used on this . Possible values are fast and slow. By default slow is used. When configured to be fast more frequent LACP heartbeats will be requested causing connectivity problems to be detected more quickly.

An interface within a . Interface name. Should be alphanumeric and no more than about 8 bytes long. May be the same as the port name, for non-bonded ports. Must otherwise be unique among the names of ports, interfaces, and bridges on a host.

Ethernet address to set for this interface. If unset then the default MAC address is used:

For the local interface, the default is the lowest-numbered MAC address among the other bridge ports, either the value of the in its record, if set, or its actual MAC (for bonded ports, the MAC of its slave whose name is first in alphabetical order). Internal ports and bridge ports that are used as port mirroring destinations (see the table) are ignored.
For other internal interfaces, the default MAC is randomly generated.
External interfaces typically have a MAC address associated with their hardware.

Some interfaces may not have a software-controllable MAC address.

OpenFlow port number for this interface. Unlike most columns, this column's value should be set only by Open vSwitch itself. Other clients should set this column to an empty set (the default) when creating an .

Open vSwitch populates this column when the port number becomes known. If the interface is successfully added, will be set to a number between 1 and 65535 (generally either in the range 1 to 65279, inclusive, or 65534, the port number for the OpenFlow ``local port''). If the interface cannot be added then Open vSwitch sets this column to -1.

The interface type, one of:

system

An ordinary network device, e.g. eth0 on Linux. Sometimes referred to as ``external interfaces'' since they are generally connected to hardware external to that on which the Open vSwitch is running. The empty string is a synonym for system.

internal

A simulated network device that sends and receives traffic. An internal interface whose is the same as its bridge's is called the ``local interface.'' It does not make sense to bond an internal interface, so the terms ``port'' and ``interface'' are often used imprecisely for internal interfaces.

tap

A TUN/TAP device managed by Open vSwitch.

gre

An Ethernet over RFC 2890 Generic Routing Encapsulation over IPv4 tunnel. Each tunnel must be uniquely identified by the combination of remote_ip, local_ip, and in_key. Note that if two ports are defined that are the same except one has an optional identifier and the other does not, the more specific one is matched first. in_key is considered more specific than local_ip if a port defines one and another port defines the other. The following options may be specified in the column:

remote_ip: Required. The tunnel endpoint.

local_ip: Optional. The destination IP that received packets must match. Default is to match all addresses.

in_key: Optional. The GRE key that received packets must contain. It may either be a 32-bit number (no key and a key of 0 are treated as equivalent) or the word flow. If flow is specified then any key will be accepted and the key will be placed in the tun_id field for matching in the flow table. The ovs-ofctl manual page contains additional information about matching fields in OpenFlow flows. Default is no key.

out_key: Optional. The GRE key to be set on outgoing packets. It may either be a 32-bit number or the word flow. If flow is specified then the key may be set using the set_tunnel Nicira OpenFlow vendor extension (0 is used in the absence of an action). The ovs-ofctl manual page contains additional information about the Nicira OpenFlow vendor extensions. Default is no key.

key: Optional. Shorthand to set in_key and out_key at the same time.

tos: Optional. The value of the ToS bits to be set on the encapsulating packet. It may also be the word inherit, in which case the ToS will be copied from the inner packet if it is IPv4 or IPv6 (otherwise it will be 0). Note that the ECN fields are always inherited. Default is 0.

ttl: Optional. The TTL to be set on the encapsulating packet. It may also be the word inherit, in which case the TTL will be copied from the inner packet if it is IPv4 or IPv6 (otherwise it will be the system default, typically 64). Default is the system default TTL.

csum: Optional. Compute GRE checksums on outgoing packets. Checksums present on incoming packets will be validated regardless of this setting. Note that GRE checksums impose a significant performance penalty as they cover the entire packet. As the contents of the packet is typically covered by L3 and L4 checksums, this additional checksum only adds value for the GRE and encapsulated Ethernet headers. Default is disabled, set to true to enable.

pmtud: Optional. Enable tunnel path MTU discovery. If enabled ``ICMP destination unreachable - fragmentation'' needed messages will be generated for IPv4 packets with the DF bit set and IPv6 packets above the minimum MTU if the packet size exceeds the path MTU minus the size of the tunnel headers. It also forces the encapsulating packet DF bit to be set (it is always set if the inner packet implies path MTU discovery). Note that this option causes behavior that is typically reserved for routers and therefore is not entirely in compliance with the IEEE 802.1D specification for bridges. Default is enabled, set to false to disable.

header_cache: Optional. Enable caching of tunnel headers and the output path. This can lead to a significant performance increase without changing behavior. In general it should not be necessary to adjust this setting. However, the caching can bypass certain components of the IP stack (such as IP tables) and it may be useful to disable it if these features are required or as a debugging measure. Default is enabled, set to false to disable.

ipsec_gre

An Ethernet over RFC 2890 Generic Routing Encapsulation over IPv4 IPsec tunnel. Each tunnel (including those of type gre) must be uniquely identified by the combination of remote_ip and local_ip. Note that if two ports are defined that are the same except one has an optional identifier and the other does not, the more specific one is matched first. An authentication method of peer_cert or psk must be defined. The following options may be specified in the column:

remote_ip: Required. The tunnel endpoint.

local_ip: Optional. The destination IP that received packets must match. Default is to match all addresses.

peer_cert: Required for certificate authentication. A string containing the peer's certificate in PEM format. Additionally the host's certificate must be specified with the certificate option.

certificate: Required for certificate authentication. The name of a PEM file containing a certificate that will be presented to the peer during authentication.

private_key: Optional for certificate authentication. The name of a PEM file containing the private key associated with certificate. If certificate contains the private key, this option may be omitted.

psk: Required for pre-shared key authentication. Specifies a pre-shared key for authentication that must be identical on both sides of the tunnel.

in_key: Optional. The GRE key that received packets must contain. It may either be a 32-bit number (no key and a key of 0 are treated as equivalent) or the word flow. If flow is specified then any key will be accepted and the key will be placed in the tun_id field for matching in the flow table. The ovs-ofctl manual page contains additional information about matching fields in OpenFlow flows. Default is no key.

out_key: Optional. The GRE key to be set on outgoing packets. It may either be a 32-bit number or the word flow. If flow is specified then the key may be set using the set_tunnel Nicira OpenFlow vendor extension (0 is used in the absence of an action). The ovs-ofctl manual page contains additional information about the Nicira OpenFlow vendor extensions. Default is no key.

key: Optional. Shorthand to set in_key and out_key at the same time.

tos: Optional. The value of the ToS bits to be set on the encapsulating packet. It may also be the word inherit, in which case the ToS will be copied from the inner packet if it is IPv4 or IPv6 (otherwise it will be 0). Note that the ECN fields are always inherited. Default is 0.

ttl: Optional. The TTL to be set on the encapsulating packet. It may also be the word inherit, in which case the TTL will be copied from the inner packet if it is IPv4 or IPv6 (otherwise it will be the system default, typically 64). Default is the system default TTL.

csum: Optional. Compute GRE checksums on outgoing packets. Checksums present on incoming packets will be validated regardless of this setting. Note that GRE checksums impose a significant performance penalty as they cover the entire packet. As the contents of the packet is typically covered by L3 and L4 checksums, this additional checksum only adds value for the GRE and encapsulated Ethernet headers. Default is disabled, set to true to enable.

pmtud: Optional. Enable tunnel path MTU discovery. If enabled ``ICMP destination unreachable - fragmentation'' needed messages will be generated for IPv4 packets with the DF bit set and IPv6 packets above the minimum MTU if the packet size exceeds the path MTU minus the size of the tunnel headers. It also forces the encapsulating packet DF bit to be set (it is always set if the inner packet implies path MTU discovery). Note that this option causes behavior that is typically reserved for routers and therefore is not entirely in compliance with the IEEE 802.1D specification for bridges. Default is enabled, set to false to disable.

capwap

Ethernet tunneling over the UDP transport portion of CAPWAP (RFC 5415). This allows interoperability with certain switches where GRE is not available. Note that only the tunneling component of the protocol is implemented. Due to the non-standard use of CAPWAP, UDP ports 58881 and 58882 are used as the source and destination ports respectively. Each tunnel must be uniquely identified by the combination of remote_ip and local_ip. If two ports are defined that are the same except one includes local_ip and the other does not, the more specific one is matched first. CAPWAP support is not available on all platforms. Currently it is only supported in the Linux kernel module with kernel versions >= 2.6.25. The following options may be specified in the column:

remote_ip: Required. The tunnel endpoint.

local_ip: Optional. The destination IP that received packets must match. Default is to match all addresses.

tos: Optional. The value of the ToS bits to be set on the encapsulating packet. It may also be the word inherit, in which case the ToS will be copied from the inner packet if it is IPv4 or IPv6 (otherwise it will be 0). Note that the ECN fields are always inherited. Default is 0.

ttl: Optional. The TTL to be set on the encapsulating packet. It may also be the word inherit, in which case the TTL will be copied from the inner packet if it is IPv4 or IPv6 (otherwise it will be the system default, typically 64). Default is the system default TTL.

pmtud: Optional. Enable tunnel path MTU discovery. If enabled ``ICMP destination unreachable - fragmentation'' needed messages will be generated for IPv4 packets with the DF bit set and IPv6 packets above the minimum MTU if the packet size exceeds the path MTU minus the size of the tunnel headers. It also forces the encapsulating packet DF bit to be set (it is always set if the inner packet implies path MTU discovery). Note that this option causes behavior that is typically reserved for routers and therefore is not entirely in compliance with the IEEE 802.1D specification for bridges. Default is enabled, set to false to disable.

header_cache: Optional. Enable caching of tunnel headers and the output path. This can lead to a significant performance increase without changing behavior. In general it should not be necessary to adjust this setting. However, the caching can bypass certain components of the IP stack (such as IP tables) and it may be useful to disable it if these features are required or as a debugging measure. Default is enabled, set to false to disable.

patch

A pair of virtual devices that act as a patch cable. The column must have the following key-value pair:

peer: The of the for the other side of the patch. The named 's own peer option must specify this 's name. That is, the two patch interfaces must have reversed and peer values.

Configuration options whose interpretation varies based on .

Status information about interfaces attached to bridges, updated every 5 seconds. Not all interfaces have all of these properties; virtual interfaces don't have a link speed, for example. Non-applicable columns will have empty values.

The administrative state of the physical network link.

The observed state of the physical network link. This is ordinarily the link's carrier status. If the interface's is a bond configured for miimon monitoring, it is instead the network link's miimon status.

The negotiated speed of the physical network link. Valid values are positive integers greater than 0.

The duplex mode of the physical network link.

The MTU (maximum transmission unit); i.e. the largest amount of data that can fit into a single Ethernet frame. The standard Ethernet MTU is 1500 bytes. Some physical media and many kinds of virtual interfaces can be configured with higher MTUs.

This column will be empty for an interface that does not have an MTU as, for example, some kinds of tunnels do not.

Key-value pairs that report port status. Supported status values are type-dependent; some interfaces may not have a valid driver_name, for example.

The currently defined key-value pairs are:

driver_name: The name of the device driver controlling the network adapter.

driver_version: The version string of the device driver controlling the network adapter.

firmware_version: The version string of the network adapter's firmware, if available.

source_ip: The source IP address used for an IPv4 tunnel end-point, such as gre or capwap.

tunnel_egress_iface: Egress interface for tunnels. Currently only relevant for GRE and CAPWAP tunnels. On Linux systems, this column will show the name of the interface which is responsible for routing traffic destined for the configured remote_ip. This could be an internal interface such as a bridge port.

tunnel_egress_iface_carrier: Whether a carrier is detected on . Valid values are down and up.

These settings control ingress policing for packets received on this interface. On a physical interface, this limits the rate at which traffic is allowed into the system from the outside; on a virtual interface (one connected to a virtual machine), this limits the rate at which the VM is able to transmit.

Policing is a simple form of quality-of-service that simply drops packets received in excess of the configured rate. Due to its simplicity, policing is usually less accurate and less effective than egress QoS (which is configured using the and tables).

Policing is currently implemented only on Linux. The Linux implementation uses a simple ``token bucket'' approach:

The size of the bucket corresponds to . Initially the bucket is full.
Whenever a packet is received, its size (converted to tokens) is compared to the number of tokens currently in the bucket. If the required number of tokens are available, they are removed and the packet is forwarded. Otherwise, the packet is dropped.
Whenever it is not full, the bucket is refilled with tokens at the rate specified by .

Policing interacts badly with some network protocols, and especially with fragmented IP packets. Suppose that there is enough network activity to keep the bucket nearly empty all the time. Then this token bucket algorithm will forward a single packet every so often, with the period depending on packet size and on the configured rate. All of the fragments of an IP packets are normally transmitted back-to-back, as a group. In such a situation, therefore, only one of these fragments will be forwarded and the rest will be dropped. IP does not provide any way for the intended recipient to ask for only the remaining fragments. In such a case there are two likely possibilities for what will happen next: either all of the fragments will eventually be retransmitted (as TCP will do), in which case the same problem will recur, or the sender will not realize that its packet has been dropped and data will simply be lost (as some UDP-based protocols will do). Either way, it is possible that no forward progress will ever occur.

Maximum rate for data received on this interface, in kbps. Data received faster than this rate is dropped. Set to 0 (the default) to disable policing.

Maximum burst size for data received on this interface, in kb. The default burst size if set to 0 is 1000 kb. This value has no effect if is 0.

Specifying a larger burst size lets the algorithm be more forgiving, which is important for protocols like TCP that react severely to dropped packets. The burst size should be at least the size of the interface's MTU. Specifying a value that is numerically at least as large as 10% of helps TCP come closer to achieving the full rate.

Connectivity monitor configuration for this interface. Key-value pairs for use by external frameworks that integrate with Open vSwitch, rather than by Open vSwitch itself. System integrators should either use the Open vSwitch development mailing list to coordinate on common key-value definitions, or choose key names that are likely to be unique. The currently defined common key-value pairs are:

attached-mac: The MAC address programmed into the ``virtual hardware'' for this interface, in the form xx:xx:xx:xx:xx:xx. For Citrix XenServer, this is the value of the MAC field in the VIF record for this interface.
iface-id: A system-unique identifier for the interface. On XenServer, this will commonly be the same as xs-vif-uuid.

Additionally the following key-value pairs specifically apply to an interface that represents a virtual Ethernet interface connected to a virtual machine. These key-value pairs should not be present for other types of interfaces. Keys whose names end in -uuid have values that uniquely identify the entity in question. For a Citrix XenServer hypervisor, these values are UUIDs in RFC 4122 format. Other hypervisors may use other formats.

The currently defined key-value pairs for XenServer are:

xs-vif-uuid: The virtual interface associated with this interface.
xs-network-uuid: The virtual network to which this interface is attached.
xs-vm-uuid: The VM to which this interface belongs.

Key-value pairs for rarely used interface features.

lacp-port-priority: The LACP port priority of this . In LACP negotiations s with numerically lower priorities are preferred for aggregation. Must be a number between 1 and 65535.

Key-value pairs that report interface statistics. The current implementation updates these counters periodically. In the future, we plan to, instead, update them when an interface is created, when they are queried (e.g. using an OVSDB select operation), and just before an interface is deleted due to virtual interface hot-unplug or VM shutdown, and perhaps at other times, but not on any regular periodic basis.

The currently defined key-value pairs are listed below. These are the same statistics reported by OpenFlow in its struct ofp_port_stats structure. If an interface does not support a given statistic, then that pair is omitted.

Successful transmit and receive counters:

rx_packets

Number of received packets.

rx_bytes

Number of received bytes.

tx_packets

Number of transmitted packets.

tx_bytes

Number of transmitted bytes.
Receive errors:

rx_dropped

Number of packets dropped by RX.

rx_frame_err

Number of frame alignment errors.

rx_over_err

Number of packets with RX overrun.

rx_crc_err

Number of CRC errors.

rx_errors

Total number of receive errors, greater than or equal to the sum of the above.
Transmit errors:

tx_dropped

Number of packets dropped by TX.

collisions

Number of collisions.

tx_errors

Total number of transmit errors, greater than or equal to the sum of the above.

Quality of Service (QoS) configuration for each Port that references it.

The type of QoS to implement. The column in the table identifies the types that a switch actually supports. The currently defined types are listed below:

linux-htb: Linux ``hierarchy token bucket'' classifier. See tc-htb(8) (also at http://linux.die.net/man/8/tc-htb) and the HTB manual (http://luxik.cdi.cz/~devik/qos/htb/manual/userg.htm) for information on how this classifier works and how to configure it.

linux-hfsc: Linux "Hierarchical Fair Service Curve" classifier. See http://linux-ip.net/articles/hfsc.en/ for information on how this classifier works.

A map from queue numbers to records. The supported range of queue numbers depend on . The queue numbers are the same as the queue_id used in OpenFlow in struct ofp_action_enqueue and other structures. Queue 0 is used by OpenFlow output actions that do not specify a specific queue.

Key-value pairs for configuring QoS features that depend on .

The linux-htb and linux-hfsc classes support the following key-value pairs:

max-rate: Maximum rate shared by all queued traffic, in bit/s. Optional. If not specified, for physical interfaces, the default is the link rate. For other interfaces or if the link rate cannot be determined, the default is currently 100 Mbps.

A configuration for a port output queue, used in configuring Quality of Service (QoS) features. May be referenced by column in table.

Key-value pairs for configuring the output queue. The supported key-value pairs and their meanings depend on the of the records that reference this row.

The key-value pairs defined for of min-rate are:

min-rate: Minimum guaranteed bandwidth, in bit/s. Required. The floor value is 1500 bytes/s (12,000 bit/s).

The key-value pairs defined for of linux-htb are:

min-rate: Minimum guaranteed bandwidth, in bit/s.
max-rate: Maximum allowed bandwidth, in bit/s. Optional. If specified, the queue's rate will not be allowed to exceed the specified value, even if excess bandwidth is available. If unspecified, defaults to no limit.
burst: Burst size, in bits. This is the maximum amount of ``credits'' that a queue can accumulate while it is idle. Optional. Details of the linux-htb implementation require a minimum burst size, so a too-small burst will be silently ignored.
priority: A nonnegative 32-bit integer. Defaults to 0 if unspecified. A queue with a smaller priority will receive all the excess bandwidth that it can use before a queue with a larger value receives any. Specific priority values are unimportant; only relative ordering matters.

The key-value pairs defined for of linux-hfsc are:

min-rate: Minimum guaranteed bandwidth, in bit/s.
max-rate: Maximum allowed bandwidth, in bit/s. Optional. If specified, the queue's rate will not be allowed to exceed the specified value, even if excess bandwidth is available. If unspecified, defaults to no limit.

A attaches to an to implement 802.1ag Connectivity Fault Management (CFM). CFM allows a group of Maintenance Points (MPs) called a Maintenance Association (MA) to detect connectivity problems with each other. MPs within a MA should have complete and exclusive interconnectivity. This is verified by occasionally broadcasting Continuity Check Messages (CCMs) at a configurable transmission interval. A is responsible for collecting data about other MPs in its MA and broadcasting CCMs.

A Maintenance Point ID (MPID) uniquely identifies each endpoint within a Maintenance Association (see ). The MPID is used to identify this to other endpoints in the MA. A set of which this should have connectivity to. If this does not have connectivity to any MPs in this set, or has connectivity to any MPs not in this set, a fault is signaled. A Maintenance Association (MA) name pairs with a Maintenance Domain (MD) name to uniquely identify a MA. A MA is a group of endpoints who have complete and exclusive interconnectivity. Defaults to ovs if unset. A Maintenance Domain name pairs with a Maintenance Association name to uniquely identify a MA. Defaults to ovs if unset. The transmission interval of CCMs in milliseconds. Three missed CCMs indicate a connectivity fault. Defaults to 1000ms. Indicates a Connectivity Fault caused by a configuration error, a down remote MP, or unexpected connectivity to a remote MAID or remote MP.

A represents a MP which a has or should have connectivity to.

A Maintenance Point ID (MPID) uniquely identifies each endpoint within a Maintenance Association. All MPs within a MA should have a unique MPID. Indicates a connectivity fault.

A port mirror within a .

A port mirror configures a bridge to send selected frames to special ``mirrored'' ports, in addition to their normal destinations. Mirroring traffic may also be referred to as SPAN or RSPAN, depending on the mechanism used for delivery.

Arbitrary identifier for the .

To be selected for mirroring, a given packet must enter or leave the bridge through a selected port and it must also be in one of the selected VLANs.

If true, every packet arriving or departing on any port is selected for mirroring. Ports on which departing packets are selected for mirroring. Ports on which arriving packets are selected for mirroring. VLANs on which packets are selected for mirroring. An empty set selects packets on all VLANs.

These columns are mutually exclusive. Exactly one of them must be nonempty.

Output port for selected packets, if nonempty.

Specifying a port for mirror output reserves that port exclusively for mirroring. No frames other than those selected for mirroring will be forwarded to the port, and any frames received on the port will be discarded.

This type of mirroring is sometimes called SPAN.

Output VLAN for selected packets, if nonempty.

The frames will be sent out all ports that trunk , as well as any ports with implicit VLAN . When a mirrored frame is sent out a trunk port, the frame's VLAN tag will be set to , replacing any existing tag; when it is sent out an implicit VLAN port, the frame will not be tagged. This type of mirroring is sometimes called RSPAN.

The following destination MAC addresses will not be mirrored to a VLAN to avoid confusing switches that interpret the protocols that they represent:

01:80:c2:00:00:00: IEEE 802.1D Spanning Tree Protocol (STP).
01:80:c2:00:00:01: IEEE Pause frame.
01:80:c2:00:00:0x: Other reserved protocols.
01:00:0c:cc:cc:cc: Cisco Discovery Protocol (CDP), VLAN Trunking Protocol (VTP), Dynamic Trunking Protocol (DTP), Port Aggregation Protocol (PAgP), and others.
01:00:0c:cc:cc:cd: Cisco Shared Spanning Tree Protocol PVSTP+.
01:00:0c:cd:cd:cd: Cisco STP Uplink Fast.
01:00:0c:00:00:00: Cisco Inter Switch Link.

Please note: Mirroring to a VLAN can disrupt a network that contains unmanaged switches. Consider an unmanaged physical switch with two ports: port 1, connected to an end host, and port 2, connected to an Open vSwitch configured to mirror received packets into VLAN 123 on port 2. Suppose that the end host sends a packet on port 1 that the physical switch forwards to port 2. The Open vSwitch forwards this packet to its destination and then reflects it back on port 2 in VLAN 123. This reflected packet causes the unmanaged physical switch to replace the MAC learning table entry, which correctly pointed to port 1, with one that incorrectly points to port 2. Afterward, the physical switch will direct packets destined for the end host to the Open vSwitch on port 2, instead of to the end host on port 1, disrupting connectivity. If mirroring to a VLAN is desired in this scenario, then the physical switch must be replaced by one that learns Ethernet addresses on a per-VLAN basis. In addition, learning should be disabled on the VLAN containing mirrored traffic. If this is not done then intermediate switches will learn the MAC address of each end host from the mirrored traffic. If packets being sent to that end host are also mirrored, then they will be dropped since the switch will attempt to send them out the input port. Disabling learning for the VLAN will cause the switch to correctly send the packet out all ports configured for that VLAN. If Open vSwitch is being used as an intermediate switch, learning can be disabled by adding the mirrored VLAN to in the appropriate table or tables.

An OpenFlow controller.

Open vSwitch supports two kinds of OpenFlow controllers:

Primary controllers

This is the kind of controller envisioned by the OpenFlow 1.0 specification. Usually, a primary controller implements a network policy by taking charge of the switch's flow table.

Open vSwitch initiates and maintains persistent connections to primary controllers, retrying the connection each time it fails or drops. The column in the table applies to primary controllers.

Open vSwitch permits a bridge to have any number of primary controllers. When multiple controllers are configured, Open vSwitch connects to all of them simultaneously. Because OpenFlow 1.0 does not specify how multiple controllers coordinate in interacting with a single switch, more than one primary controller should be specified only if the controllers are themselves designed to coordinate with each other. (The Nicira-defined NXT_ROLE OpenFlow vendor extension may be useful for this.)

Service controllers

These kinds of OpenFlow controller connections are intended for occasional support and maintenance use, e.g. with ovs-ofctl. Usually a service controller connects only briefly to inspect or modify some of a switch's state.

Open vSwitch listens for incoming connections from service controllers. The service controllers initiate and, if necessary, maintain the connections from their end. The column in the table does not apply to service controllers.

Open vSwitch supports configuring any number of service controllers.

The determines the type of controller.

Connection method for controller.

The following connection methods are currently supported for primary controllers:

ssl:ip[:port]

The specified SSL port (default: 6633) on the host at the given ip, which must be expressed as an IP address (not a DNS name). The column in the table must point to a valid SSL configuration when this form is used.

SSL support is an optional feature that is not always built as part of Open vSwitch.

tcp:ip[:port]

The specified TCP port (default: 6633) on the host at the given ip, which must be expressed as an IP address (not a DNS name).

The following connection methods are currently supported for service controllers:

pssl:[port][:ip]

Listens for SSL connections on the specified TCP port (default: 6633). If ip, which must be expressed as an IP address (not a DNS name), is specified, then connections are restricted to the specified local IP address.

The column in the table must point to a valid SSL configuration when this form is used.

SSL support is an optional feature that is not always built as part of Open vSwitch.

ptcp:[port][:ip]

Listens for connections on the specified TCP port (default: 6633). If ip, which must be expressed as an IP address (not a DNS name), is specified, then connections are restricted to the specified local IP address.

When multiple controllers are configured for a single bridge, the values must be unique. Duplicate values yield unspecified results.

If it is specified, this setting must be one of the following strings that describes how Open vSwitch contacts this OpenFlow controller over the network:

in-band: In this mode, this controller's OpenFlow traffic travels over the bridge associated with the controller. With this setting, Open vSwitch allows traffic to and from the controller regardless of the contents of the OpenFlow flow table. (Otherwise, Open vSwitch would never be able to connect to the controller, because it did not have a flow to enable it.) This is the most common connection mode because it is not necessary to maintain two independent networks.
out-of-band: In this mode, OpenFlow traffic uses a control network separate from the bridge associated with this controller, that is, the bridge does not use any of its own network devices to communicate with the controller. The control network must be configured separately, before or after ovs-vswitchd is started.

If not specified, the default is implementation-specific.

Maximum number of milliseconds to wait between connection attempts. Default is implementation-specific. Maximum number of milliseconds of idle time on connection to controller before sending an inactivity probe message. If Open vSwitch does not communicate with the controller for the specified number of seconds, it will send a probe. If a response is not received for the same additional amount of time, Open vSwitch assumes the connection has been broken and attempts to reconnect. Default is implementation-specific. A value of 0 disables inactivity probes.

The maximum rate at which packets in unknown flows will be forwarded to the OpenFlow controller, in packets per second. This feature prevents a single bridge from overwhelming the controller. If not specified, the default is implementation-specific.

In addition, when a high rate triggers rate-limiting, Open vSwitch queues controller packets for each port and transmits them to the controller at the configured rate. The number of queued packets is limited by the value. The packet queue is shared fairly among the ports on a bridge.

Open vSwitch maintains two such packet rate-limiters per bridge. One of these applies to packets sent up to the controller because they do not correspond to any flow. The other applies to packets sent up to the controller by request through flow actions. When both rate-limiters are filled with packets, the actual rate that packets are sent to the controller is up to twice the specified rate.

In conjunction with , the maximum number of unused packet credits that the bridge will allow to accumulate, in packets. If not specified, the default is implementation-specific.

These values are considered only in in-band control mode (see ).

When multiple controllers are configured on a single bridge, there should be only one set of unique values in these columns. If different values are set for these columns in different controllers, the effect is unspecified.

The IP address to configure on the local port, e.g. 192.168.0.123. If this value is unset, then and are ignored. The IP netmask to configure on the local port, e.g. 255.255.255.0. If is set but this value is unset, then the default is chosen based on whether the IP address is class A, B, or C. The IP address of the gateway to configure on the local port, as a string, e.g. 192.168.0.1. Leave this column unset if this network has no gateway. Key-value pairs for use by external frameworks that integrate with Open vSwitch, rather than by Open vSwitch itself. System integrators should either use the Open vSwitch development mailing list to coordinate on common key-value definitions, or choose key names that are likely to be unique. No common key-value pairs are currently defined. true if currently connected to this controller, false otherwise.

The level of authority this controller has on the associated bridge. Possible values are:

other: Allows the controller access to all OpenFlow features.
master: Equivalent to other, except that there may be at most one master controller at a time. When a controller configures itself as master, any existing master is demoted to the slaverole.
slave: Allows the controller read-only access to OpenFlow features. Attempts to modify the flow table will be rejected with an error. Slave controllers do not receive OFPT_PACKET_IN or OFPT_FLOW_REMOVED messages, but they do receive OFPT_PORT_STATUS messages.

Key-value pairs that report controller status.

last_error: A human-readable description of the last error on the connection to the controller; i.e. strerror(errno). This key will exist only if an error has occurred.
state: The state of the connection to the controller. Possible values are: VOID (connection is disabled), BACKOFF (attempting to reconnect at an increasing period), CONNECTING (attempting to connect), ACTIVE (connected, remote host responsive), and IDLE (remote host idle, sending keep-alive). These values may change in the future. They are provided only for human consumption.
sec_since_connect: The amount of time since this controller last successfully connected to the switch (in seconds). Value is empty if controller has never successfully connected.
sec_since_disconnect: The amount of time since this controller last disconnected from the switch (in seconds). Value is empty if controller has never disconnected.

Configuration for a database connection to an Open vSwitch database (OVSDB) client.

This table primarily configures the Open vSwitch database (ovsdb-server), not the Open vSwitch switch (ovs-vswitchd). The switch does read the table to determine what connections should be treated as in-band.

The Open vSwitch database server can initiate and maintain active connections to remote clients. It can also listen for database connections.

Connection method for managers.

The following connection methods are currently supported:

ssl:ip[:port]

The specified SSL port (default: 6632) on the host at the given ip, which must be expressed as an IP address (not a DNS name). The column in the table must point to a valid SSL configuration when this form is used.

SSL support is an optional feature that is not always built as part of Open vSwitch.

tcp:ip[:port]

The specified TCP port (default: 6632) on the host at the given ip, which must be expressed as an IP address (not a DNS name).

pssl:[port][:ip]

Listens for SSL connections on the specified TCP port (default: 6632). If ip, which must be expressed as an IP address (not a DNS name), is specified, then connections are restricted to the specified local IP address.

The column in the table must point to a valid SSL configuration when this form is used.

SSL support is an optional feature that is not always built as part of Open vSwitch.

ptcp:[port][:ip]

Listens for connections on the specified TCP port (default: 6632). If ip, which must be expressed as an IP address (not a DNS name), is specified, then connections are restricted to the specified local IP address.

When multiple managers are configured, the values must be unique. Duplicate values yield unspecified results.

If it is specified, this setting must be one of the following strings that describes how Open vSwitch contacts this OVSDB client over the network:

in-band: In this mode, this connection's traffic travels over a bridge managed by Open vSwitch. With this setting, Open vSwitch allows traffic to and from the client regardless of the contents of the OpenFlow flow table. (Otherwise, Open vSwitch would never be able to connect to the client, because it did not have a flow to enable it.) This is the most common connection mode because it is not necessary to maintain two independent networks.
out-of-band: In this mode, the client's traffic uses a control network separate from that managed by Open vSwitch, that is, Open vSwitch does not use any of its own network devices to communicate with the client. The control network must be configured separately, before or after ovs-vswitchd is started.

If not specified, the default is implementation-specific.

Maximum number of milliseconds to wait between connection attempts. Default is implementation-specific. Maximum number of milliseconds of idle time on connection to the client before sending an inactivity probe message. If Open vSwitch 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, Open vSwitch assumes the connection has been broken and attempts to reconnect. Default is implementation-specific. A value of 0 disables inactivity probes. Key-value pairs for use by external frameworks that integrate with Open vSwitch, rather than by Open vSwitch itself. System integrators should either use the Open vSwitch development mailing list to coordinate on common key-value definitions, or choose key names that are likely to be unique. No common key-value pairs are currently defined. true if currently connected to this manager, false otherwise.

Key-value pairs that report manager status.

last_error: A human-readable description of the last error on the connection to the manager; i.e. strerror(errno). This key will exist only if an error has occurred.

state: The state of the connection to the manager. Possible values are: VOID (connection is disabled), BACKOFF (attempting to reconnect at an increasing period), CONNECTING (attempting to connect), ACTIVE (connected, remote host responsive), and IDLE (remote host idle, sending keep-alive). These values may change in the future. They are provided only for human consumption.

sec_since_connect: The amount of time since this manager last successfully connected to the database (in seconds). Value is empty if manager has never successfully connected.

sec_since_disconnect: The amount of time since this manager last disconnected from the database (in seconds). Value is empty if manager has never disconnected.

A NetFlow target. NetFlow is a protocol that exports a number of details about terminating IP flows, such as the principals involved and duration. NetFlow targets in the form ip:port. The ip must be specified numerically, not as a DNS name. Engine ID to use in NetFlow messages. Defaults to datapath index if not specified. Engine type to use in NetFlow messages. Defaults to datapath index if not specified. The interval at which NetFlow records are sent for flows that are still active, in seconds. A value of 0 requests the default timeout (currently 600 seconds); a value of -1 disables active timeouts.

If this column's value is false, the ingress and egress interface fields of NetFlow flow records are derived from OpenFlow port numbers. When it is true, the 7 most significant bits of these fields will be replaced by the least significant 7 bits of the engine id. This is useful because many NetFlow collectors do not expect multiple switches to be sending messages from the same host, so they do not store the engine information which could be used to disambiguate the traffic.

When this option is enabled, a maximum of 508 ports are supported.

SSL configuration for an Open_vSwitch. Name of a PEM file containing the private key used as the switch's identity for SSL connections to the controller. Name of a PEM file containing a certificate, signed by the certificate authority (CA) used by the controller and manager, that certifies the switch's private key, identifying a trustworthy switch. Name of a PEM file containing the CA certificate used to verify that the switch is connected to a trustworthy controller. If set to true, then Open vSwitch will attempt to obtain the CA certificate from the controller on its first SSL connection and save it to the named PEM file. If it is successful, it will immediately drop the connection and reconnect, and from then on all SSL connections must be authenticated by a certificate signed by the CA certificate thus obtained. This option exposes the SSL connection to a man-in-the-middle attack obtaining the initial CA certificate. It may still be useful for bootstrapping. Key-value pairs for use by external frameworks that integrate with Open vSwitch, rather than by Open vSwitch itself. System integrators should either use the Open vSwitch development mailing list to coordinate on common key-value definitions, or choose key names that are likely to be unique. No common key-value pairs are currently defined.

An sFlow(R) target. sFlow is a protocol for remote monitoring of switches.

Name of the network device whose IP address should be reported as the ``agent address'' to collectors. If not specified, the IP address defaults to the in the collector's . If an agent IP address cannot be determined either way, sFlow is disabled. Number of bytes of a sampled packet to send to the collector. If not specified, the default is 128 bytes. Polling rate in seconds to send port statistics to the collector. If not specified, defaults to 30 seconds. Rate at which packets should be sampled and sent to the collector. If not specified, defaults to 400, which means one out of 400 packets, on average, will be sent to the collector. sFlow targets in the form ip:port. Key-value pairs for use by external frameworks that integrate with Open vSwitch, rather than by Open vSwitch itself. System integrators should either use the Open vSwitch development mailing list to coordinate on common key-value definitions, or choose key names that are likely to be unique. No common key-value pairs are currently defined.

Records in this table describe functionality supported by the hardware and software platform on which this Open vSwitch is based. Clients should not modify this table.

A record in this table is meaningful only if it is referenced by the column in the table. The key used to reference it, called the record's ``category,'' determines the meanings of the column. The following general forms of categories are currently defined:

qos-type: type is supported as the value for in the table.

Key-value pairs that describe capabilities. The meaning of the pairs depends on the category key that the column in the table uses to reference this record, as described above.

The presence of a record for category qos-type indicates that the switch supports type as the value of the column in the table. The following key-value pairs are defined to further describe QoS capabilities:

n-queues: Number of supported queues, as a positive integer. Keys in the column for records whose value equals type must range between 0 and this value minus one, inclusive.