Network Working Group L. Berger Request for Comments: 2961 LabN Consulting, LLC Category: Standards Track D. Gan Juniper Networks, Inc. G. Swallow Cisco Systems, Inc. P. Pan Juniper Networks, Inc. F. Tommasi S. Molendini University of Lecce April 2001
RSVP Refresh Overhead Reduction Extensions
Status of this Memo
This document specifies an Internet standards track protocol for the Internet community, and requests discussion and suggestions for improvements. Please refer to the current edition of the "Internet Official Protocol Standards" (STD 1) for the standardization state and status of this protocol. Distribution of this memo is unlimited.
Copyright Notice
Copyright (C) The Internet Society (2001). All Rights Reserved.
Abstract
This document describes a number of mechanisms that can be used to reduce processing overhead requirements of refresh messages, eliminate the state synchronization latency incurred when an RSVP (Resource ReserVation Protocol) message is lost and, when desired, refreshing state without the transmission of whole refresh messages. The same extensions also support reliable RSVP message delivery on a per hop basis. These extension present no backwards compatibility issues.
Berger, et al. Standards Track [Page 1]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
Standard RSVP [RFC2205] maintains state via the generation of RSVP refresh messages. Refresh messages are used to both synchronize state between RSVP neighbors and to recover from lost RSVP messages. The use of Refresh messages to cover many possible failures has resulted in a number of operational problems. One problem relates to scaling, another relates to the reliability and latency of RSVP Signaling.
Berger, et al. Standards Track [Page 2]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
The scaling problems are linked to the resource requirements (in terms of processing and memory) of running RSVP. The resource requirements increase proportionally with the number of sessions. Each session requires the generation, transmission, reception and processing of RSVP Path and Resv messages per refresh period. Supporting a large number of sessions, and the corresponding volume of refresh messages, presents a scaling problem.
The reliability and latency problem occurs when a non-refresh RSVP message is lost in transmission. Standard RSVP [RFC2205] recovers from a lost message via RSVP refresh messages. In the face of transmission loss of RSVP messages, the end-to-end latency of RSVP signaling is tied to the refresh interval of the node(s) experiencing the loss. When end-to-end signaling is limited by the refresh interval, the delay incurred in the establishment or the change of a reservation may be beyond the range of what is acceptable for some applications.
One way to address the refresh volume problem is to increase the refresh period, "R" as defined in Section 3.7 of [RFC2205]. Increasing the value of R provides linear improvement on transmission overhead, but at the cost of increasing the time it takes to synchronize state.
One way to address the reliability and latency of RSVP Signaling is to decrease the refresh period R. Decreasing the value of R increases the probability that state will be installed in the face of message loss, but at the cost of increasing refresh message rate and associated processing requirements.
An additional issue is the time to deallocate resources after a tear message is lost. RSVP does not retransmit ResvTear or PathTear messages. If the sole tear message transmitted is lost, then resources will only be deallocated once the "cleanup timer" interval has passed. This may result in resources being allocated for an unnecessary period of time. Note that even when the refresh period is adjusted, the "cleanup timer" must still expire since tear messages are not retransmitted.
The extensions defined in this document address both the refresh volume and the reliability issues with mechanisms other than adjusting refresh rate. The extensions are collectively referred to as the "Refresh Overhead Reduction" or the "Refresh Reduction" extensions. A Bundle message is defined to reduce overall message handling load. A MESSAGE_ID object is defined to reduce refresh message processing by allowing the receiver to more readily identify an unchanged message. A MESSAGE_ACK object is defined which can be used to detect message loss and support reliable RSVP message
Berger, et al. Standards Track [Page 3]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
delivery on a per hop basis. A summary refresh message is defined to enable refreshing state without the transmission of whole refresh messages, while maintaining RSVP's ability to indicate when state is lost and to adjust to changes in routing.
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119].
This document categorizes RSVP messages into two types: trigger and refresh messages. Trigger messages are those RSVP messages that advertise state or any other information not previously transmitted. Trigger messages include messages advertising new state, a route change that alters a reservation path, or a modification to an existing RSVP session or reservation. Trigger messages also include those messages that include changes in non-RSVP processed objects, such as changes in the Policy or ADSPEC objects.
Refresh messages represent previously advertised state and contain exactly the same objects and same information as a previously transmitted message, and are sent over the same path. Only Path and Resv messages can be refresh messages. Refresh messages are identical to the corresponding previously transmitted message, with some possible exceptions. Specifically, the checksum field, the flags field and the INTEGRITY object may differ in refresh messages.
To indicate support for the refresh overhead reduction extensions, an additional capability bit is added to the common RSVP header, which is defined in [RFC2205].
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
When set, indicates that this node is willing and capable of receiving all the messages and objects described in this document. This includes the Bundle message described in Section 3, the MESSAGE_ID objects and Ack messages described in Section 4, and the MESSAGE_ID LIST objects and Srefresh message described in Section 5. This bit is meaningful only between RSVP neighbors.
Nodes supporting the refresh overhead reduction extensions must also take care to recognize when a next hop stops sending RSVP messages with the Refresh-Reduction-Capable bit set. To cover this case, nodes supporting the refresh overhead reduction extensions MUST examine the flags field of each received RSVP message. If the flag changes from indicating support to indicating non-support then, unless configured otherwise, Srefresh messages (described in Section 5) MUST NOT be used for subsequent state refreshes to that neighbor and Bundle messages (Section 3) MUST NOT be sent to that neighbor. Note, a node that supports reliable RSVP message delivery (Section 4) but not Bundle and Srefresh messages, will not set the Refresh- Reduction-Capable bit.
An RSVP Bundle message consists of a bundle header followed by a body consisting of a variable number of standard RSVP messages. A Bundle message is used to aggregate multiple RSVP messages within a single PDU. The term "bundling" is used to avoid confusion with RSVP reservation aggregation. The following subsections define the formats of the bundle header and the rules for including standard RSVP messages as part of the message.
The one's complement of the one's complement sum of the entire message, with the checksum field replaced by zero for the purpose of computing the checksum. An all-zero value means that no checksum was transmitted. Because individual sub- messages may carry their own checksum as well as the INTEGRITY object for authentication, this field MAY be set to zero. Note that when the checksum is not computed, the header of the bundle message will not be covered by any checksum. If the checksum is computed, individual sub-messages MAY set their own checksum to zero.
Send_TTL: 8 bits
The IP TTL value with which the message was sent. This is used by RSVP to detect a non-RSVP hop by comparing the Send_TTL with the IP TTL in a received message.
RSVP length: 16 bits
The total length of this RSVP Bundle message in bytes, including the bundle header and the sub-messages that follow.
Support for RSVP Bundle messages is optional. While message bundling helps in scaling RSVP, by reducing processing overhead and bandwidth consumption, a node is not required to transmit every standard RSVP message in a Bundle message. A node MUST always be ready to receive standard RSVP messages.
RSVP Bundle messages can only be sent to RSVP neighbors that support bundling. Methods for discovering such information include: (1) manual configuration and (2) observing the Refresh-Reduction-Capable bit (see Section 2) in the received RSVP messages. RSVP Bundle messages MUST NOT be used if the RSVP neighbor does not support RSVP Bundle messages.
RSVP Bundle messages are sent hop by hop between RSVP-capable nodes as "raw" IP datagrams with protocol number 46. The IP source address is an address local to the system that originated the Bundle message. The IP destination address is the RSVP neighbor for which the sub- messages are intended.
RSVP Bundle messages SHOULD NOT be sent with the Router Alert IP option in their IP headers. This is because Bundle messages are addressed directly to RSVP neighbors.
Each RSVP Bundle message MUST occupy exactly one IP datagram, which is approximately 64K bytes. If it exceeds the MTU, the datagram is fragmented by IP and reassembled at the recipient node. Implementations may choose to limit each RSVP Bundle message to the MTU size of the outgoing link, e.g., 1500 bytes. Implementations SHOULD also limit the amount of time that a message is delayed in order to be bundled. Different limits may be used for trigger and
Berger, et al. Standards Track [Page 7]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
standard refresh messages. Trigger messages SHOULD be delayed a minimal amount of time. Refresh messages may be delayed up to their refresh interval. Note that messages related to the same Resv or Path state should not be delayed at different intervals in order to preserve ordering.
If the RSVP neighbor is not known or changes in next hops cannot be identified via routing, Bundle messages MUST NOT be used. Note that when the routing next hop is not RSVP capable it will typically not be possible to identify changes in next hop.
Any message that will be handled by the RSVP neighbor indicated in a Bundle Message's destination address may be included in the same message. This includes all RSVP messages that would be sent out a point-to-point link. It includes any message, such as a Resv, addressed to the same destination address. It also includes Path and PathTear messages when the next hop is known to be the destination and changes in next hops can be detected. Path and PathTear messages for multicast sessions MUST NOT be sent in Bundle messages when the outgoing link is not a point-to-point link or when the next hop does not support the refresh overhead reduction extensions.
If the local system does not recognize or does not wish to accept a Bundle message, the received messages shall be discarded without further analysis.
The receiver next compares the Send_TTL with which a Bundle message is sent to the IP TTL with which it is received. If a non-RSVP hop is detected, the number of non-RSVP hops is recorded. It is used later in processing of sub-messages.
Next, the receiver verifies the version number and checksum of the RSVP Bundle message and discards the message if any mismatch is found.
The receiver then starts decapsulating individual sub-messages. Each sub-message has its own complete message length and authentication information. With the exception of using the Send_TTL from the header of the Bundle message, each sub-message is processed as if it was received individually.
Three new objects are defined as part of the MESSAGE_ID extension. The objects are the MESSAGE_ID object, the MESSAGE_ID_ACK object, and the MESSAGE_ID_NACK objects. The first two objects are used to
Berger, et al. Standards Track [Page 8]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
support acknowledgments and reliable RSVP message delivery. The last object is used to support the summary refresh extension described in Section 5. The MESSAGE_ID object can also be used to simply provide a shorthand indication of when the message carrying the object is a refresh message. Such information can be used by the receiving node to reduce refresh processing requirements.
Message identification and acknowledgment is done on a per hop basis. All types of MESSAGE_ID objects contain a message identifier. The identifier MUST be unique on a per object generator's IP address basis. No more than one MESSAGE_ID object may be included in an RSVP message. Each message containing a MESSAGE_ID object may be acknowledged via a MESSAGE_ID_ACK object, when so indicated. MESSAGE_ID_ACK and MESSAGE_ID_NACK objects may be sent piggy-backed in unrelated RSVP messages or in RSVP Ack messages. RSVP messages carrying any of the three object types may be included in a bundle message. When included, each object is treated as if it were contained in a standard, non-bundled, RSVP message.
The MESSAGE_ID, MESSAGE_ID_ACK and MESSAGE_ID_NACK objects may be included in the standard RSVP messages, as defined in [RFC2205]. When included, one or more MESSAGE_ID_ACK or MESSAGE_ID_NACK objects MUST immediately follow the INTEGRITY object. When no INTEGRITY object is present, the MESSAGE_ID_ACK or MESSAGE_ID_NACK objects MUST immediately follow the message or sub-message header. Only one MESSAGE_ID object MAY be included in a message or sub-message and it MUST follow any present MESSAGE_ID_ACK or MESSAGE_ID_NACK objects. When no MESSAGE_ID_ACK or MESSAGE_ID_NACK objects are present, the MESSAGE_ID object MUST immediately follow the INTEGRITY object. When no INTEGRITY object is present, the MESSAGE_ID object MUST immediately follow the message or sub-message header.
The ordering of the ACK objects for all standard RSVP messages is: <Common Header> [ <INTEGRITY> ] [ [<MESSAGE_ID_ACK> | <MESSAGE_ID_NACK>] ... ] [ <MESSAGE_ID> ]
Berger, et al. Standards Track [Page 9]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
Indicates that the sender requests the receiver to send an acknowledgment for the message.
Epoch: 24 bits
A value that indicates when the Message_Identifier sequence has reset. SHOULD be randomly generated each time a node reboots or the RSVP agent is restarted. The value SHOULD NOT be the same as was used when the node was last operational. This value MUST NOT be changed during normal operation.
Message_Identifier: 32 bits
When combined with the message generator's IP address, the Message_Identifier field uniquely identifies a message. The values placed in this field change incrementally and only decrease when the Epoch changes or when the value wraps.
Berger, et al. Standards Track [Page 10]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
Ack messages carry one or more MESSAGE_ID_ACK or MESSAGE_ID_NACK objects. They MUST NOT contain any MESSAGE_ID objects. Ack messages are sent between neighboring RSVP nodes. The IP destination address of an Ack message is the unicast address of the node that generated the message(s) being acknowledged. For messages with RSVP_HOP objects, such as Path and Resv messages, the address is found in the RSVP_HOP object. For other messages, such as ResvConf, the associated IP address is the source address in the IP header. The IP source address is an address of the node that sends the Ack message.
Berger, et al. Standards Track [Page 11]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
The MESSAGE_ID object may be included in any RSVP message other than the Ack and Bundle messages. The MESSAGE_ID object is always generated and processed over a single hop between RSVP neighbors. The IP address of the object generator, i.e., the node that creates the object, is represented in a per RSVP message type specific fashion. For messages with RSVP_HOP objects, such as Path and Resv messages, the generator's IP address is found in the RSVP_HOP object. For other messages, such as ResvConf message, the generator's IP address is the source address in the IP header. Note that MESSAGE_ID objects can only be used in a Bundle sub-messages, but not in a Bundle message. As is always the case with the Bundle message, each sub-message is processed as if it was received individually. This includes processing of MESSAGE_ID objects.
The Epoch field contains a generator selected value. The value is used to indicate when the sender resets the values used in the Message_Identifier field. On startup, a node SHOULD randomly select a value to be used in the Epoch field. The node SHOULD ensure that the selected value is not the same as was used when the node was last operational. The value MUST NOT be changed unless the node or the RSVP agent is restarted.
The Message_Identifier field contains a generator selected value. This value, when combined with the generator's IP address, identifies a particular RSVP message and the specific state information it represents. The combination of Message_Identifier and Epoch can also be used to detect out of order messages. When a node is sending a refresh message with a MESSAGE_ID object, it SHOULD use the same Message_Identifier value that was used in the RSVP message that first advertised the state being refreshed. When a node is sending a trigger message, the Message_Identifier value MUST have a value that is greater than any other value previously used with the same Epoch field value. A value is considered to have been used when it has
Berger, et al. Standards Track [Page 12]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
been sent in any message using the associated IP address with the same Epoch field value.
The ACK_Desired flag is set when the MESSAGE_ID object generator wants a MESSAGE_ID_ACK object sent in response to the message. Such information can be used to ensure reliable delivery of RSVP messages in the face of network loss. Nodes setting the ACK_Desired flag SHOULD retransmit unacknowledged messages at a more rapid interval than the standard refresh period until the message is acknowledged or until a "rapid" retry limit is reached. Rapid retransmission rate MUST be based on the exponential exponential back-off procedures defined in section 6. The ACK_Desired flag will typically be set only in trigger messages. The ACK_Desired flag MAY be set in refresh messages. Issues relate to multicast sessions are covered in a later section.
Nodes processing incoming MESSAGE_ID objects SHOULD check to see if a newly received message is out of order and can be ignored. Out of order messages SHOULD be ignored, i.e., silently dropped. Out of order messages can be identified by examining the values in the Epoch and Message_Identifier fields. To determine ordering, the received Epoch value must match the value previously received from the message sender. If the values differ then the receiver MUST NOT treat the message as out of order. When the Epoch values match and the Message_Identifier value is less than the largest value previously received from the sender, then the receiver SHOULD check the value previously received for the state associated with the message. This check should be performed for any message that installs or changes state. (Includes at least: Path, Resv, PathTear, ResvTear, PathErr and ResvErr.) If no local state information can be associated with the message, the receiver MUST NOT treat the message as out of order. If local state can be associated with the message and the received Message_Identifier value is less than the most recently received value associated with the state, the message SHOULD be treated as being out of order.
Note that the 32-bit Message_Identifier value MAY wrap. To cover the wrap case, the following expression may be used to test if a newly received Message_Identifier value is less than a previously received value:
if ((int) old_id - (int) new_id > 0) { new value is less than old value; }
MESSAGE_ID objects of messages that are not out of order SHOULD be used to aid in determining if the message represents new state or a state refresh. Note that state is only refreshed in Path and Resv
Berger, et al. Standards Track [Page 13]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
messages. If the received Epoch values differs from the value previously received from the message sender, the message is a trigger message and the receiver MUST fully process the message. If a Path or Resv message contains the same Message_Identifier value that was used in the most recently received message for the same session and, for Path messages, SENDER_TEMPLATE then the receiver SHOULD treat the message as a state refresh. If the Message_Identifier value is greater than the most recently received value, the receiver MUST fully processes the message. When fully processing a Path or Resv message, the receiver MUST store the received Message_Identifier value as part of the local Path or Resv state for future reference.
Nodes receiving a non-out of order message containing a MESSAGE_ID object with the ACK_Desired flag set, SHOULD respond with a MESSAGE_ID_ACK object. Note that MESSAGE_ID objects received in messages containing errors, i.e., are not syntactically valid, MUST NOT be acknowledged. PathErr and ResvErr messages SHOULD be treated as implicit acknowledgments.
4.6. MESSAGE_ID_ACK Object and MESSAGE_ID_NACK Object Usage
The MESSAGE_ID_ACK object is used to acknowledge receipt of messages containing MESSAGE_ID objects that were sent with the ACK_Desired flag set. A MESSAGE_ID_ACK object MUST NOT be generated in response to a received MESSAGE_ID object when the ACK_Desired flag is not set.
The MESSAGE_ID_NACK object is used as part of the summary refresh extension. The generation and processing of MESSAGE_ID_NACK objects is described in further detail in Section 5.4.
MESSAGE_ID_ACK and MESSAGE_ID_NACK objects MAY be sent in any RSVP message that has an IP destination address matching the generator of the associated MESSAGE_ID object. This means that the objects will not typically be included in the non hop-by-hop Path, PathTear and ResvConf messages. When no appropriate message is available, one or more objects SHOULD be sent in an Ack message. Implementations SHOULD include MESSAGE_ID_ACK and MESSAGE_ID_NACK objects in standard RSVP messages when possible.
Implementations SHOULD limit the amount of time that an object is delayed in order to be piggy-backed or sent in an Ack message. Different limits may be used for MESSAGE_ID_ACK and MESSAGE_ID_NACK objects. MESSAGE_ID_ACK objects are used to detect link transmission losses. If an ACK object is delayed too long, the corresponding message will be retransmitted. To avoid such retransmission, ACK objects SHOULD be delayed a minimal amount of time. A delay time equal to the link transit time MAY be used. MESSAGE_ID_NACK objects may be delayed an independent and longer time, although additional
Berger, et al. Standards Track [Page 14]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
delay increases the amount of time a desired reservation is not installed.
Path and PathTear messages may be sent to IP multicast destination addresses. When the destination is a multicast address, it is possible that a single message containing a single MESSAGE_ID object will be received by multiple RSVP next hops. When the ACK_Desired flag is set in this case, acknowledgment processing is more complex.
There are a number of issues to be addressed including ACK implosion, number of acknowledgments to be expected and handling of new receivers.
ACK implosion occurs when each receiver responds to the MESSAGE_ID object at approximately the same time. This can lead to a potentially large number of MESSAGE_ID_ACK objects being simultaneously delivered to the message generator. To address this case, the receiver MUST wait a random interval prior to acknowledging a MESSAGE_ID object received in a message destined to a multicast address. The random interval SHOULD be between zero (0) and a configured maximum time. The configured maximum SHOULD be set in proportion to the refresh and "rapid" retransmission interval, i.e, such that the maximum time before sending an acknowledgment does not result in retransmission. It should be noted that ACK implosion is being addressed by spreading acknowledgments out in time, not by ACK suppression.
A more fundamental issue is the number of acknowledgments that the upstream node, i.e., the message generator, should expect. The number of acknowledgments that should be expected is the same as the number of RSVP next hops. In the router-to-router case, the number of next hops can often be obtained from routing. When hosts are either the upstream node or the next hops, the number of next hops will typically not be readily available. Another case where the number of RSVP next hops will typically not be known is when there are non-RSVP routers between the message generator and the RSVP next hops.
When the number of next hops is not known, the message generator SHOULD only expect a single response. The result of this behavior will be special retransmission handling until the message is delivered to at least one next hop, then followed by standard RSVP refreshes. Refresh messages will synchronize state with any next hops that don't receive the original message.
Berger, et al. Standards Track [Page 15]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
When using the MESSAGE_ID extension with multicast sessions it is preferable for RSVP to obtain the number of next hops from routing and to be notified when that number changes. The interface between routing and RSVP is purely an implementation issue. Since RSVP [RFC2205] describes a reference routing interface, a version of the RSVP/routing interface updated to provide number of next hop information is presented. See [RFC2205] for previously defined parameters and function description.
All nodes sending messages with the Refresh-Reduction-Capable bit set will support the MESSAGE_ID Extension. There are no backward compatibility issues raised by the MESSAGE_ID Class with nodes that do not set the Refresh-Reduction-Capable bit. The MESSAGE_ID Class has an assigned value whose form is 0bbbbbbb. Per RSVP [RFC2205], classes with values of this form must be rejected with an "Unknown Object Class" error by nodes not supporting the class. When the receiver of a MESSAGE_ID object does not support the class, a corresponding error message will be generated. The generator of the MESSAGE_ID object will see the error and then MUST re-send the original message without the MESSAGE_ID object. In this case, the message generator MAY still choose to retransmit messages at the "rapid" retransmission interval. Lastly, since the MESSAGE_ID_ACK class can only be issued in response to the MESSAGE_ID object, there are no possible issues with this class or Ack messages. A node MAY support the MESSAGE_ID Extension without supporting the other refresh overhead reduction extensions.
Berger, et al. Standards Track [Page 16]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
The summary refresh extension enables the refreshing of RSVP state without the transmission of standard Path or Resv messages. The benefits of the described extension are that it reduces the amount of information that must be transmitted and processed in order to maintain RSVP state synchronization. Importantly, the described extension preserves RSVP's ability to handle non-RSVP next hops and to adjust to changes in routing. This extension cannot be used with Path or Resv messages that contain any change from previously transmitted messages, i.e., are trigger messages.
The summary refresh extension builds on the previously defined MESSAGE_ID extension. Only state that was previously advertised in Path and Resv messages containing MESSAGE_ID objects can be refreshed via the summary refresh extension.
The summary refresh extension uses the objects and the ACK message previously defined as part of the MESSAGE_ID extension, and a new Srefresh message. The new message carries a list of Message_Identifier fields corresponding to the Path and Resv trigger messages that established the state. The Message_Identifier fields are carried in one of three Srefresh related objects. The three objects are the MESSAGE_ID LIST object, the MESSAGE_ID SRC_LIST object, and the MESSAGE_ID MCAST_LIST object.
The MESSAGE_ID LIST object is used to refresh all Resv state, and Path state of unicast sessions. It is made up of a list of Message_Identifier fields that were originally advertised in MESSAGE_ID objects. The other two objects are used to refresh Path state of multicast sessions. A node receiving a summary refresh for multicast path state will at times need source and group information. These two objects provide this information. The objects differ in the information they contain and how they are sent. Both carry Message_Identifier fields and corresponding source IP addresses. The MESSAGE_ID SRC_LIST is sent in messages addressed to the session's multicast IP address. The MESSAGE_ID MCAST_LIST object adds the group address and is sent in messages addressed to the RSVP next hop. The MESSAGE_ID MCAST_LIST is normally used on point-to-point links.
An RSVP node receiving an Srefresh message, matches each listed Message_Identifier field with installed Path or Resv state. All matching state is updated as if a normal RSVP refresh message has been received. If matching state cannot be found, then the Srefresh message sender is notified via a refresh NACK.
Berger, et al. Standards Track [Page 17]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
A refresh NACK is sent via the MESSAGE_ID_NACK object. As described in the previous section, the rules for sending a MESSAGE_ID_NACK object are the same as for sending a MESSAGE_ID_ACK object. This includes sending MESSAGE_ID_NACK object both piggy-backed in unrelated RSVP messages or in RSVP ACK messages.
5.1. MESSAGE_ID LIST, SRC_LIST and MCAST_LIST Objects
No flags are currently defined. This field MUST be zero on transmission and ignored on receipt.
Epoch: 24 bits
The Epoch field from the MESSAGE_ID object corresponding to the trigger message that advertised the state being refreshed.
Message_Identifier: 32 bits
The Message_Identifier field from the MESSAGE_ID object corresponding to the trigger message that advertised the state being refreshed. One or more Message_Identifiers may be included.
Berger, et al. Standards Track [Page 18]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
No flags are currently defined. This field MUST be zero on transmission and ignored on receipt.
Epoch: 24 bits
The Epoch field from the MESSAGE_ID object corresponding to the trigger message that advertised the state being refreshed.
Berger, et al. Standards Track [Page 20]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
Message_Identifier
The Message_Identifier field from the MESSAGE_ID object corresponding to the trigger message that advertised the Path state being refreshed. One or more Message_Identifiers may be included. Each Message_Identifier MUST be followed by the source IP address corresponding to the sender described in the Path state being refreshed.
Source_IP_Address
The IP address corresponding to the sender of the Path state being refreshed.
No flags are currently defined. This field MUST be zero on transmission and ignored on receipt.
Epoch: 24 bits
The Epoch field from the MESSAGE_ID object corresponding to the trigger message that advertised the state being refreshed.
Message_Identifier: 32 bits
The Message_Identifier field from the MESSAGE_ID object corresponding to the trigger message that advertised the Path state being refreshed. One or more Message_Identifiers may be included. Each Message_Identifier MUST be followed by the source IP address corresponding to the sender of the Path state being refreshed, and the destination IP address of the session.
Source_IP_Address
The IP address corresponding to the sender of the Path state being refreshed.
Destination_IP_Address
The destination IP address corresponding to the session of the Path state being refreshed.
Berger, et al. Standards Track [Page 23]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
Srefresh messages carry one or more MESSAGE_ID LIST, MESSAGE_ID SRC_LIST, and MESSAGE_ID MCAST_LIST objects. MESSAGE_ID LIST and MESSAGE_ID MCAST_LIST objects MAY be carried in the same Srefresh message. MESSAGE_ID SRC_LIST can not be combined in Srefresh messages with the other objects. A single Srefresh message MAY refresh both Path and Resv state.
Srefresh messages carrying Message_Identifier fields corresponding to Path state are normally sent with a destination IP address equal to the address carried in the corresponding SESSION objects. The destination IP address MAY be set to the RSVP next hop when the next hop is known to be RSVP capable and either (a) the session is unicast or (b) the outgoing interface is a point-to-point link. Srefresh messages carrying Message_Identifier fields corresponding to Resv state MUST be sent with a destination IP address set to the Resv state's previous hop.
Srefresh messages sent to a multicast session's destination IP address, MUST contain MESSAGE_ID SRC_LIST objects and MUST NOT include any MESSAGE_ID LIST or MESSAGE_ID MCAST_LIST objects. Srefresh messages sent to the RSVP next hop MAY contain either or both MESSAGE_ID LIST and MESSAGE_ID MCAST_LIST objects, but MUST NOT include any MESSAGE_ID SRC_LIST objects.
The source IP address of an Srefresh message is an address of the node that generates the message. The source IP address MUST match the address associate with the MESSAGE_ID objects when they were included in a standard RSVP message. As previously mentioned, the source address associated with a MESSAGE_ID object is represented in a per RSVP message type specific fashion. For messages with RSVP_HOP objects, such as Path and Resv messages, the address is found in the RSVP_HOP object. For other messages, such as ResvConf message, the associated IP address is the source address in the IP header.
Srefresh messages that are addressed to a session's destination IP address MUST be sent with the Router Alert IP option in their IP headers. Srefresh messages addressed directly to RSVP neighbors SHOULD NOT be sent with the Router Alert IP option in their IP headers.
Each Srefresh message MUST occupy exactly one IP datagram. If it exceeds the MTU, the datagram is fragmented by IP and reassembled at the recipient node. Srefresh messages MAY be sent within an RSVP Bundle messages. Although this is not expected since Srefresh
Berger, et al. Standards Track [Page 24]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
messages can carry a list of Message_Identifier fields within a single object. Implementations may choose to limit each Srefresh message to the MTU size of the outgoing link, e.g., 1500 bytes.
An Srefresh message may be generated to refresh Resv and Path state. If an Srefresh message is used to refresh some particular state, then the generation of a standard refresh message for that particular state SHOULD be suppressed. A state's refresh interval is not affected by the use of Srefresh message based refreshes.
When generating an Srefresh message, a node SHOULD refresh as much Path and Resv state as is possible by including the information from as many MESSAGE_ID objects in the same Srefresh message. Only the information from MESSAGE_ID objects that meet the source and destination IP address restrictions, as described in Sections 5.2, may be included in the same Srefresh message. Identifying Resv state that can be refreshed using the same Srefresh message is fairly straightforward. Identifying which Path state may be included is a little more complex.
Only state that was previously advertised in Path and Resv messages containing MESSAGE_ID objects can be refreshed via an Srefresh message. Srefresh message based refreshes must preserve the state synchronization properties of Path or Resv message based refreshes. Specifically, the use of Srefresh messages MUST NOT result in state being timed-out at the RSVP next hop. The period at which state is refreshed when using Srefresh messages MAY be shorter than the period that would be used when using Path or Resv message based refreshes, but it MUST NOT be longer.
Berger, et al. Standards Track [Page 25]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
The particular approach used to trigger Srefresh message based refreshes is implementation specific. Some possibilities are triggering Srefresh message generation based on each state's refresh period or, on a per interface basis, periodically generating Srefresh messages to refresh all state that has not been refreshed within the state's refresh interval. Other approaches are also possible. A default Srefresh message generation interval of 30 seconds is suggested for nodes that do not dynamically calculate a generation interval.
When generating an Srefresh message, there are two methods for identifying which Path state may be refreshed in a specific message. In both cases, the previously mentioned refresh interval and source IP address restrictions must be followed. The primary method is to include only those sessions that share the same destination IP address in the same Srefresh message.
The secondary method for identifying which Path state may be refreshed within a single Srefresh message is an optimization. This method MAY be used when the next hop is known to support RSVP and when either (a) the session is unicast or (b) the outgoing interface is a point-to-point link. This method MUST NOT be used when the next hop is not known to support RSVP or when the outgoing interface is to a multi-access network and the session is to a multicast address. The use of this method MAY be administratively configured. When using this method, the destination address in the IP header of the Srefresh message is usually the next hop's address. When the use of this method is administratively configured, the destination address should be the well known group address 224.0.0.14. When the outgoing interface is a point-to-point link, all Path state associated with sessions advertised out the interface SHOULD be included in the same Srefresh message. When the outgoing interface is not a point-to- point link, all unicast session Path state SHOULD be included in the same Srefresh message.
Identifying which Resv state may be refreshed within a single Srefresh message is based simply on the source and destination IP addresses. Any state that was previously advertised in Resv messages with the same IP addresses as an Srefresh message MAY be included.
After identifying the Path and Resv state that can be included in a particular Srefresh message, the message generator adds to the message MESSAGE_ID information matching each identified state's previously used object. For all Resv state and for Path state of unicast sessions, the information is added to the message in a MESSAGE_ID LIST object that has a matching Epoch value. (Note only one Epoch value will be in use during normal operation.) If no matching object exists, then a new MESSAGE_ID LIST object is created.
Berger, et al. Standards Track [Page 26]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
Path state of multicast sessions may be added to the same message when the destination address of the Srefresh message is the RSVP next hop and the outgoing interface is a point-to-point link. In this case the information is added to the message in a MESSAGE_ID MCAST_LIST object that has a matching Epoch value. If no matching object exists, then a new MESSAGE_ID MCAST_LIST object is created. When the destination address of the message is a multicast address, then identified information is added to the message in a MESSAGE_ID SRC_LIST object that has a matching Epoch value. If no matching object exists, then a new MESSAGE_ID SRC_LIST object is created. Once the Srefresh message is composed, the message generator transmits the message out the proper interface.
Upon receiving an Srefresh message, the node MUST attempt to identify matching installed Path or Resv state. Matching is done based on the source address in the IP header of the Srefresh message, the object type and each Message_Identifier field. If matching state can be found, then the receiving node MUST update the matching state information as if a standard refresh message had been received. If matching state cannot be identified, then an Srefresh NACK MUST be generated corresponding to the unmatched Message_Identifier field. Message_Identifier fields received in MESSAGE_ID LIST objects may correspond to any Resv state or to Path state of unicast sessions. Message_Identifier fields received in MESSAGE_ID SRC_LIST or MCAST_LIST objects correspond to Path state of multicast sessions.
An additional check must be performed to determine if a NACK should be generated for unmatched Message_Identifier fields associated with Path state of multicast sessions, i.e., fields that were carried in MESSAGE_ID SRC_LIST or MCAST_LIST objects. The receiving node must check to see if the node would forward data packets originated from the source corresponding to the unmatched field. This check, commonly known as an RPF check, is performed based on the source and group information carried in the MESSAGE_ID SRC_LIST and MCAST_LIST objects. In both objects the IP address of the source is listed immediately after the corresponding Message_Identifier field. The group address is listed immediately after the source IP address in MESSAGE_ID MCAST_LIST objects. The group address is the message's destination IP address when MESSAGE_ID SRC_LIST objects are used. The receiving node only generates an Srefresh NACK when the node would forward packets to the identified group from the listed sender. If the node would forward multicast data packets from a listed sender and there is a corresponding unmatched Message_Identifier field, then an appropriate Srefresh NACK MUST be generated. If the node would not forward packets to the identified group from a listed sender, a corresponding unmatched Message_Identifier field is silently ignored.
Berger, et al. Standards Track [Page 27]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
Srefresh NACKs are used to indicate that a received Message_Identifier field carried in MESSAGE_ID LIST, SRC_LIST, or MCAST_LIST object does not match any installed state. This may occur for a number of reasons including, for example, a route change. An Srefresh NACK is encoded in a MESSAGE_ID_NACK object. When generating an Srefresh NACK, the epoch and Message_Identifier fields of the MESSAGE_ID_NACK object MUST have the same value as was received. MESSAGE_ID_NACK objects are transmitted as described in Section 4.6.
Received MESSAGE_ID_NACK objects indicate that the object generator does not have any installed state matching the object. Upon receiving a MESSAGE_ID_NACK object, the receiver performs an installed Path or Resv state lookup based on the Epoch and Message_Identifier values contained in the object. If matching state is found, then the receiver MUST transmit the matching state via a standard Path or Resv message. If the receiver cannot identify any installed state, then no action is required.
As discussed in [RFC2205], RSVP uses soft state to address a large class of potential errors. RSVP does this by periodically sending a full representation of installed state in Resv and Path messages. Srefresh messages are used in place of the periodic sending of standard Path and Resv refresh messages. While this provides scaling benefits and protects against common network events such as packet loss or routing change, it does not provide exactly the same error recovery properties. An example error that could potentially be recovered from via standard messages but not with Srefresh messages is internal corruption of state. This section recommends two methods that can be used to better preserve RSVP's soft state error recovery mechanism. Both mechanisms are supported using existing protocol messages.
The first mechanism uses a checksum or other algorithm to detect a previously unnoticed change in internal state. This mechanism does not protect against internal state corruption. It just covers the case where a trigger message should have been sent, but was not. When sending a Path or Resv trigger message, a node should run a checksum or other algorithm, such as [MD5], over the internal state and store the result. The choice of algorithm is an administrative decision. Periodically the node should rerun the algorithm and compare the new result with the stored result. If the values differ, then a corresponding standard Path or Resv refresh message should be
Berger, et al. Standards Track [Page 28]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
sent and the new value should be stored. The recomputation period should be set based on the computation resources of the node and the reliability requirements of the network.
The second mechanism is simply to periodically send standard Path and Resv refresh messages. Since this mechanism uses standard refresh messages, it can recover from the same set of errors as standard RSVP. When using this mechanism, the period that standard refresh messages are sent must be longer than the interval that Srefresh messages are generated in order to gain the benefits of using the summary refresh extension. When a standard refresh message is sent, a corresponding summary refresh SHOULD NOT be sent during the same refresh period. When a node supports the periodic generation of standard refresh messages while Srefreshes are being used, the frequency of generation of standard refresh messages relative to the generation of summary refreshes SHOULD be configurable by the network administrator.
Nodes supporting the summary refresh extension advertise their support via the Refresh-Reduction-Capable bit in the RSVP message header. This enables nodes supporting the extension to detect each other. When it is not known if a next hop supports the extension, standard Path and Resv message based refreshes MUST be used. Note that when the routing next hop does not support RSVP, it will not always be possible to detect if the RSVP next hop supports the summary refresh extension. Therefore, when the routing next hop is not RSVP capable the Srefresh message based refresh SHOULD NOT be used. A node MAY be administratively configured to use Srefresh messages in all cases when all RSVP nodes in a network are known to support the summary refresh extension. This is useful since when operating in this mode, the extension properly adjusts to the case of non-RSVP next hops and changes in routing.
Per section 2, nodes supporting the summary refresh extension must also take care to recognize when a next hop stops sending RSVP messages with the Refresh-Reduction-Capable bit set.
This section is based on [Pan] and provides procedures to implement exponential back-off for retransmission of messages awaiting acknowledgment, see Section 4.5. Implementations MUST use the described procedures or their equivalent.
Berger, et al. Standards Track [Page 29]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
The following is one possible mechanism for exponential back-off retransmission of an unacknowledged RSVP message: When sending such a message, a node inserts a MESSAGE_ID object with the ACK_Desired flag set. The sending node will retransmit the message until a message acknowledgment is received or the message has been transmitted a maximum number of times. Upon reception, a receiving node acknowledges the arrival of the message by sending back a message acknowledgment (that is, a corresponding MESSAGE_ID_ACK object.) When the sending node receives the acknowledgment retransmission of the message is stopped. The interval between retransmissions is governed by a rapid retransmission timer. The rapid retransmission timer starts at a small interval and increases exponentially until it reaches a threshold.
The described procedures make use of the following time parameters. All parameters are per interface.
Rapid retransmission interval Rf:
Rf is the initial retransmission interval for unacknowledged messages. After sending the message for the first time, the sending node will schedule a retransmission after Rf seconds. The value of Rf could be as small as the round trip time (RTT) between a sending and a receiving node, if known.
Rapid retry limit Rl:
Rl is the maximum number of times a message will be transmitted without being acknowledged.
Increment value Delta:
Delta governs the speed with which the sender increases the retransmission interval. The ratio of two successive retransmission intervals is (1 + Delta).
Suggested default values are an initial retransmission timeout (Rf) of 500ms, a power of 2 exponential back-off (Delta = 1) and a retry limit (Rl) of 3.
Berger, et al. Standards Track [Page 30]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
After a sending node transmits a message containing a MESSAGE_ID object with the ACK_Desired flag set, it should immediately schedule a retransmission after Rf seconds. If a corresponding MESSAGE_ID_ACK object is received earlier than Rf seconds, then retransmission SHOULD be canceled. Otherwise, it will retransmit the message after (1 + Delta)*Rf seconds. The staged retransmission will continue until either an appropriate MESSAGE_ID_ACK object is received, or the rapid retry limit, Rl, has been reached.
A sending node can use the following algorithm when transmitting a message containing a MESSAGE_ID object with the ACK_Desired flag set:
Prior to initial transmission initialize: Rk = Rf and Rn = 0
while (Rn++ < Rl) { transmit the message; wake up after Rk seconds; Rk = Rk * (1 + Delta); } /* acknowledged or no reply from receiver for too long: */ do any needed clean up; exit;
Asynchronously, when a sending node receives a corresponding MESSAGE_ID_ACK object, it will change the retry count, Rn, to Rl.
Note that the transmitting node does not advertise the use of the described exponential back-off procedures via the TIME_VALUE object.
The use of exponential back-off retransmission is a new and significant addition to RSVP. It will be important to review related operations and performance experience before this document advances to Draft Standard. It will be particularly important to review experience with multicast, and any ACK implosion problems actually encountered.
This document represents ideas and comments from the MPLS-TE design team and participants in the RSVP Working Group's interim meeting. Thanks to Bob Braden, Lixia Zhang, Fred Baker, Adrian Farrel, Roch Guerin, Kireeti Kompella, David Mankins, Henning Schulzrinne, Andreas Terzis, Lan Wang and Masanobu Yuhara for specific feedback on the various versions of the document.
Berger, et al. Standards Track [Page 31]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
Portions of this work are based on work done by Masanobu Yuhara and Mayumi Tomikawa [Yuhara].
[MD5] Rivest, R., "The MD5 Message-Digest Algorithm", RFC 1321, April 1992.
[RFC2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997.
[RFC2205] Braden, R., Ed., Zhang, L., Berson, S., Herzog, S. and S. Jamin , "Resource ReserVation Protocol -- Version 1 Functional Specification", RFC 2205, September 1997.
[Yuhara] Yuhara, M., and M Tomikawa, "RSVP Extensions for ID-based Refreshes", Work in Progress.
Berger, et al. Standards Track [Page 32]
RFC 2961 RSVP Refresh Overhead Reduction Extensions April 2001
Copyright (C) The Internet Society (2001). All Rights Reserved.
This document and translations of it may be copied and furnished to others, and derivative works that comment on or otherwise explain it or assist in its implementation may be prepared, copied, published and distributed, in whole or in part, without restriction of any kind, provided that the above copyright notice and this paragraph are included on all such copies and derivative works. However, this document itself may not be modified in any way, such as by removing the copyright notice or references to the Internet Society or other Internet organizations, except as needed for the purpose of developing Internet standards in which case the procedures for copyrights defined in the Internet Standards process must be followed, or as required to translate it into languages other than English.
The limited permissions granted above are perpetual and will not be revoked by the Internet Society or its successors or assigns.
This document and the information contained herein is provided on an "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
Acknowledgement
Funding for the RFC Editor function is currently provided by the Internet Society.