3 - Bridge compatibility support has been removed. Any uses that
4 rely on ovs-brcompatd will have to stick with Open vSwitch 1.9.x
5 or adapt to native Open vSwitch support (e.g. use ovs-vsctl instead
7 - The maximum size of the MAC learning table is now configurable.
8 - New support for the VXLAN tunnel protocol (see the IETF draft here:
9 http://tools.ietf.org/html/draft-mahalingam-dutt-dcops-vxlan-02).
10 - With the Linux datapath, packets for new flows are now queued
11 separately on a per-port basis, so it should no longer be
12 possible for a large number of new flows arriving on one port to
13 prevent new flows from being processed on other ports.
14 - Many "ovs-vsctl" database commands now accept an --if-exists option.
15 Please refer to the ovs-vsctl manpage for details.
16 - New "vlog/disable-rate-limit" and "vlog/enable-rate-limit" commands
17 available through ovs-appctl allow control over logging rate limits.
18 - The OpenFlow "dp_desc" may now be configured by setting the value of
19 other-config:dp-desc in the Bridge table.
20 - Path MTU discovery is no longer supported.
21 - Backward-incompatible changes:
22 - Earlier Open vSwitch versions treated ANY as a wildcard in flow
23 syntax. OpenFlow 1.1 adds a port named ANY, which introduces a
24 conflict. ANY was rarely used in flow syntax, so we chose to
25 retire that meaning of ANY in favor of the OpenFlow 1.1 meaning.
26 - Inheritance of the Don't Fragment bit in IP tunnels (df_inherit) is
28 - Patch ports are implemented in userspace.
29 - Tunneling requires the version of the kernel module paired with Open
30 vSwitch 1.9.0 or later.
35 - The tunneling code no longer assumes input and output keys are symmetric.
36 If they are not, PMTUD needs to be disabled for tunneling to work. Note
37 this only applies to flow-based keys.
39 - Support for ipv6 set action.
40 - SKB mark matching and setting.
41 - support for Linux kernels up to 3.8
42 - FreeBSD is now a supported platform, thanks to code contributions from
43 Gaetano Catalli, Ed Maste, and Giuseppe Lettieri.
44 - ovs-bugtool: New --ovs option to report only OVS related information.
45 - New %t and %T log escapes to identify the subprogram within a
46 cooperating group of processes or threads that emitted a log message.
47 The default log patterns now include this information.
49 - Allow bitwise masking for SHA and THA fields in ARP, SLL and TLL
50 fields in IPv6 neighbor discovery messages, and IPv6 flow label.
51 - Adds support for writing to the metadata field for a flow.
52 - It is possible to request the OpenFlow port number with the
53 "ofport_request" column in the Interface table.
55 - Commands and actions that accept port numbers now also accept keywords
56 that represent those ports (such as LOCAL, NONE, and ALL). This is
57 also the recommended way to specify these ports, for compatibility
58 with OpenFlow 1.1 and later (which use the OpenFlow 1.0 numbers
59 for these ports for different purposes).
61 - Support requesting the port number with the "port_no" option in
63 - The "dump-flows" and "del-flows" no longer require an argument
64 if only one datapath exists.
66 - New "dpif/dump-dps", "dpif/show", and "dpif/dump-flows" command
67 that mimic the equivalent ovs-dpctl commands.
68 - ovs-pki: The "online PKI" features have been removed, along with
69 the ovs-pki-cgi program that facilitated it, because of some
70 alarmist insecurity claims. We do not believe that these claims
71 are true, but because we do not know of any users for this
72 feature it seems better on balance to remove it. (The ovs-pki-cgi
73 program was not included in distribution packaging.)
74 - Tunnel Path MTU Discovery default value was set to 'disabled'. This
75 feature is deprecated and will be removed soon.
76 - ovsdb-server now enforces the immutability of immutable columns. This
77 was not enforced in earlier versions due to an oversight.
78 - New support for a nonstandard form of GRE that supports a 64-bit key.
79 - The ofproto library is now responsible for assigning OpenFlow port
80 numbers. An ofproto implementation should assign them when
81 port_construct() is called.
82 - All dpif-based bridges of a particular type share a common
83 datapath called "ovs-<type>", e.g. "ovs-system". The ovs-dpctl
84 commands will now return information on that shared datapath. To
85 get the equivalent bridge-specific information, use the new
86 "ovs-appctl dpif/*" commands.
87 - Tunnel header caching removed.
88 - The following features are now deprecated. They will be removed no
89 earlier than February 2013. Please email dev@openvswitch.org with
91 - Bridge compatibility.
93 - The autopath action.
94 - Interface type "null".
95 - Numeric values for reserved ports (see "ovs-ofctl" note above).
96 - Tunnel Path MTU Discovery.
97 - CAPWAP tunnel support.
98 - The data in the RARP packets can now be matched in the same way as the
103 ------------------------
104 - New FAQ. Please send updates and additions!
105 - Authors of controllers, please read the new section titled "Action
106 Reproduction" in DESIGN, which describes an Open vSwitch change in
107 behavior in corner cases that may affect some controllers.
109 - A new test utility that can create L3 tunnel between two Open
110 vSwitches and detect connectivity issues.
112 - New --sort and --rsort options for "dump-flows" command.
113 - "mod-port" command can now control all OpenFlow config flags.
115 - Allow general bitwise masking for IPv4 and IPv6 addresses in
116 IPv4, IPv6, and ARP packets. (Previously, only CIDR masks
118 - Allow support for arbitrary Ethernet masks. (Previously, only
119 the multicast bit in the destination address could be individually
121 - New field OXM_OF_METADATA, to align with OpenFlow 1.1.
122 - The OFPST_QUEUE request now reports an error if a specified port or
123 queue does not exist, or for requests for a specific queue on all
124 ports, if the specified queue does not exist on any port. (Previous
125 versions generally reported an empty set of results.)
126 - New "flow monitor" feature to allow controllers to be notified of
127 flow table changes as they happen.
128 - Additional protocols are not mirrored and dropped when forward-bpdu is
129 false. For a full list, see the ovs-vswitchd.conf.db man page.
130 - Open vSwitch now sends RARP packets in situations where it previously
131 sent a custom protocol, making it consistent with behavior of QEMU and
133 - All Open vSwitch programs and log files now show timestamps in UTC,
134 instead the local timezone, by default.
138 ------------------------
139 - kernel modules are renamed. openvswitch_mod.ko is now
140 openvswitch.ko and brcompat_mod.ko is now brcompat.ko.
141 - Increased the number of NXM registers to 8.
142 - Added ability to configure DSCP setting for manager and controller
143 connections. By default, these connections have a DSCP value of
144 Internetwork Control (0xc0).
145 - Added the granular link health statistics, 'cfm_health', to an
148 - Added support to mask nd_target for ICMPv6 neighbor discovery flows.
149 - Added support for OpenFlow 1.3 port description (OFPMP_PORT_DESC)
152 - Added the "dump-ports-desc" command to retrieve port
153 information using the new port description multipart messages.
155 - Added support for spawning ovs-test server from the client.
156 - Now ovs-test is able to automatically create test bridges and ports.
157 - "ovs-dpctl dump-flows" now prints observed TCP flags in TCP flows.
158 - Tripled flow setup performance.
159 - The "coverage/log" command previously available through ovs-appctl
160 has been replaced by "coverage/show". The new command replies with
161 coverage counter values, instead of logging them.
165 ------------------------
166 - Allow OFPP_CONTROLLER as the in_port for packet-out messages.
170 ------------------------
171 *** Internal only release ***
173 - LACP bonds no longer fall back to balance-slb when negotiations fail.
174 Instead they drop traffic.
175 - The default bond_mode changed from SLB to active-backup, to protect
176 unsuspecting users from the significant risks of SLB bonds (which are
177 documented in vswitchd/INTERNALS).
178 - Load balancing can be disabled by setting the bond-rebalance-interval
181 - Added support for bitwise matching on TCP and UDP ports.
182 See ovs-ofctl(8) for more information.
183 - NXM flow dumps now include times elapsed toward idle and hard
185 - Added an OpenFlow extension NXT_SET_ASYNC_CONFIG that allows
186 controllers more precise control over which OpenFlow messages they
187 receive asynchronously.
188 - New "fin_timeout" action.
189 - Added "fin_timeout" support to "learn" action.
190 - New Nicira action NXAST_CONTROLLER that offers additional features
191 over output to OFPP_CONTROLLER.
192 - When QoS settings for an interface do not configure a default queue
193 (queue 0), Open vSwitch now uses a default configuration for that
194 queue, instead of dropping all packets as in previous versions.
196 - Logging to console and file will have UTC timestamp as a default for
197 all the daemons. An example of the default format is
198 2012-01-27T16:35:17Z. ovs-appctl can be used to change the default
200 - The syntax of commands and options to set log levels was simplified,
201 to make it easier to remember.
202 - New support for limiting the number of flows in an OpenFlow flow
203 table, with configurable policy for evicting flows upon
204 overflow. See the Flow_Table table in ovs-vswitch.conf.db(5)
205 for more information.
206 - New "enable-async-messages" column in the Controller table. If set to
207 false, OpenFlow connections to the controller will initially have all
208 asynchronous messages disabled, overriding normal OpenFlow behavior.
209 - ofproto-provider interface:
210 - "struct rule" has a new member "used" that ofproto implementations
211 should maintain by updating with ofproto_rule_update_used().
213 - The new option --timestamp causes the "monitor" command to print
214 a timestamp with every update.
215 - CFM module CCM broadcasts can now be tagged with an 802.1p priority.
219 ------------------------
221 - Added support for querying, modifying, and deleting flows
222 based on flow cookie when using NXM.
223 - Added new NXM_PACKET_IN format.
224 - Added new NXAST_DEC_TTL action.
226 - Added daemonization support to the monitor and snoop commands.
228 - The "find" command supports new set relational operators
229 {=}, {!=}, {<}, {>}, {<=}, and {>=}.
230 - ovsdb-tool now uses the typical database and schema installation
231 directories as defaults.
232 - The default MAC learning timeout has been increased from 60 seconds
233 to 300 seconds. The MAC learning timeout is now configurable.
237 ------------------------
238 - Compatible with Open vSwitch kernel module included in Linux 3.3.
239 - New "VLAN splinters" feature to work around buggy device drivers
240 in old Linux versions. (This feature is deprecated. When
241 broken device drivers are no longer in widespread use, we will
242 delete this feature.) See ovs-vswitchd.conf.db(5) for more
245 - Added ability to match on IPv6 flow label through NXM.
246 - Added ability to match on ECN bits in IPv4 and IPv6 through NXM.
247 - Added ability to match on TTL in IPv4 and IPv6 through NXM.
248 - Added ability to modify ECN bits in IPv4.
249 - Added ability to modify TTL in IPv4.
251 - Don't require the "normal" action to use mirrors. Traffic will
252 now be properly mirrored for any flows, regardless of their
254 - Track packet and byte statistics sent on mirrors.
255 - The sFlow implementation can now usually infer the correct agent
256 device instead of having to be told explicitly.
258 - New "fdb/flush" command to flush bridge's MAC learning table.
260 - A new distributed testing tool that allows one to diagnose performance
261 and connectivity issues. This tool currently is not included in RH or
263 - RHEL packaging now supports integration with Red Hat network scripts.
265 - Post 1.4.*, OVS will be changing the default bond mode from balance-slb
266 to active-backup. SLB bonds carry significant risks with them
267 (documented vswitchd/INTERNALS) which we want to prevent unsuspecting
268 users from running into. Users are advised to update any scripts or
269 configuration which may be negatively impacted by explicitly setting
270 the bond mode which they want to use.
274 ------------------------
276 - Added an OpenFlow extension which allows the "output" action to accept
278 - Added an OpenFlow extension for flexible learning.
279 - Bumped number of NXM registers from four to five.
281 - New "version" command to determine version of running daemon.
282 - If no argument is provided for "cfm/show", displays detailed
283 information about all interfaces with CFM enabled.
284 - If no argument is provided for "lacp/show", displays detailed
285 information about all ports with LACP enabled.
287 - New "set-if" command to modify a datapath port's configuration.
289 - The software switch now supports 255 OpenFlow tables, instead
290 of just one. By default, only table 0 is consulted, but the
291 new NXAST_RESUBMIT_TABLE action can look up in additional
292 tables. Tables 128 and above are reserved for use by the
293 switch itself; please use only tables 0 through 127.
294 - Add support for 802.1D spanning tree (STP).
295 - Fragment handling extensions:
296 - New OFPC_FRAG_NX_MATCH fragment handling mode, in which L4
297 fields are made available for matching in fragments with
299 - New NXM_NX_IP_FRAG match field for matching IP fragments (usable
300 via "ip_frag" in ovs-ofctl).
301 - New ovs-ofctl "get-frags" and "set-frags" commands to get and set
302 fragment handling policy.
303 - CAPWAP tunneling now supports an extension to transport a 64-bit key.
304 By default it remains compatible with the old version and other
305 standards-based implementations.
306 - Flow setups are now processed in a round-robin manner across ports
307 to prevent any single client from monopolizing the CPU and conducting
308 a denial of service attack.
309 - Added support for native VLAN tagging. A new "vlan_mode"
310 parameter can be set for "port". Possible values: "access",
311 "trunk", "native-tagged" and "native-untagged".
312 - test-openflowd has been removed. Please use ovs-vswitchd instead.
315 ------------------------
316 - New "ofproto" abstraction layer to ease porting to hardware
318 - Packaging for Red Hat Enterprise Linux 5.6 and 6.0.
319 - Datapath support for Linux kernels up to 3.0.
321 - New "bundle" and "bundle_load" action extensions.
323 - Implement table unique constraints.
324 - Support cooperative locking between callers.
326 - New "-s" option for "show" command prints packet and byte
327 counters for each port.
329 - New "--readd" option for "replace-flows".
331 - New "show" command to print an overview of configuration.
332 - New "comment" command to add remark that explains intentions.
333 - ovs-brcompatd has been rewritten to fix long-standing bugs.
334 - ovs-openflowd has been renamed test-openflowd and moved into the
335 tests directory. Its presence confused too many users. Please
336 use ovs-vswitchd instead.
337 - New ovs-benchmark utility to test flow setup performance.
338 - A new log level "off" has been added. Configuring a log facility
339 "off" prevents any messages from being logged to it. Previously,
340 "emer" was effectively "off" because no messages were ever logged at
341 level "emer". Now, errors that cause a process to exit are logged
343 - "configure" option --with-l26 has been renamed --with-linux, and
344 --with-l26-source has been renamed --with-linux-source. The old
345 names will be removed after the next release, so please update
347 - The "-2.6" suffix has been dropped from the datapath/linux-2.6 and
348 datapath/linux-2.6/compat-2.6 directories.
350 - Dropped support for "tun_id_from_cookie" OpenFlow extension.
351 Please use the extensible match extensions instead.
352 - Removed the Maintenance_Point and Monitor tables in an effort
353 to simplify 802.1ag configuration.
354 - Performance and scalability improvements
358 ------------------------
359 - Ability to define policies over IPv6
362 - Support for extensible match extensions to OpenFlow
364 - Support for HFSC qdisc.
365 - Queue used by in-band control can now be configured.
367 - Kernel<->userspace interface has been reworked and should be
368 close to a stable ABI now.
369 - "Port group" concept has been dropped.
370 - GRE over IPSEC tunnels
372 - New active backup bonding mode.
373 - New L4 hashing support when LACP is enabled.
374 - Source MAC hash now includes VLAN field also.
376 - Greatly improved handling of large flow tables
378 - "show" command now prints full vport configuration.
379 - "dump-groups" command removed since kernel support for
380 port groups was dropped.
382 - New commands for working with the new Managers table.
383 - "list" command enhanced with new formatting options and --columns
385 - "get" command now accepts new --id option.
386 - New "find" command.
388 - New "queue-stats" command for printing queue stats.
389 - New commands "replace-flows" and "diff-flows".
390 - Commands to add and remove flows can now read from files.
391 - New --flow-format option to enable or disable NXM.
392 - New --more option to increase OpenFlow message verbosity.
393 - Removed "tun-cookie" command, which is no longer useful.
394 - ovs-controller enhancements for testing various features.
395 - New ovs-vlan-test command for testing for Linux kernel driver VLAN
396 bugs. New ovs-vlan-bug-workaround command for enabling and
397 disabling a workaround for these driver bugs.
399 - "Resubmit" actions now update flow statistics.
400 - New "register" extension for use in matching and actions, via NXM.
401 - New "multipath" experimental action extension.
402 - New support for matching multicast Ethernet frames, via NXM.
403 - New extension for OpenFlow vendor error codes.
404 - New extension to set the QoS output queue without actually
405 sending to an output port.
406 - Open vSwitch now reports a single flow table, instead of
407 separate hash and wildcard tables. This better models the
408 current implementation.
409 - New experimental "note" action.
410 - New "ofproto/trace" ovs-appctl command and associated utilities
411 to ease debugging complex flow tables.
413 - Schema documentation now includes an entity-relationship diagram.
414 - The database is now garbage collected. In most tables,
415 unreferenced rows will be deleted automatically.
416 - Many tables now include statistics updated periodically by
417 ovs-vswitchd or ovsdb-server.
418 - Every table now has an "external-ids" column for use by OVS
420 - There is no default controller anymore. Each bridge must have its
421 controller individually specified.
422 - The "fail-mode" is now a property of a Bridge instead of a Controller.
423 - New versioning and checksum features.
424 - New Managers table and manager_options column in Open_vSwitch table
425 for specifying managers. The old "managers" column in the
426 Open_vSwitch table has been removed.
427 - Many "name" columns are now immutable.
429 - Dropped support for XenServer pre-5.6.100.
430 - Dropped support for Linux pre-2.6.18.
431 - Dropped controller discovery support.
432 - Dropped "ovs-ofctl status" and the OpenFlow extension that it used.
433 Statistics reporting in the database is a rough equivalent.
434 - Dropped the "corekeeper" package (now separate, at
435 http://openvswitch.org/cgi-bin/gitweb.cgi?p=corekeeper).
436 - Performance and scalability improvements
439 v1.1.0pre2 - 13 Sep 2010
440 ------------------------
443 v1.1.0pre1 - 31 Aug 2010
444 ------------------------
445 - OpenFlow 1.0 slicing (QoS) functionality
446 - Python bindings for configuration database (no write support)
447 - Performance and scalability improvements
452 - New "patch" interface type
457 - Configuration database with remote management
460 - Support for XenServer 5.5 and 5.6
461 - Performance and scalability improvements
464 v0.99.2 - 18 Feb 2010
465 ---------------------
468 v0.99.1 - 25 Jan 2010
469 ---------------------
470 - Add support for sFlow(R)
471 - Make headers compatible with C++
474 v0.99.0 - 14 Jan 2010
475 ---------------------
476 - User-space forwarding engine
479 v0.90.7 - 29 Nov 2009
480 ---------------------
481 - Add support for NetFlow active timeouts
488 v0.90.5 - 21 Sep 2009
489 ---------------------
490 - Generalize in-band control to more diverse network setups