RFC 3868
Signalling Connection Control Part User Adaptation Layer (SUA)
Pages: 131
Proposed Standard
Proposed Standard
Part 4 of 5 – Pages 92 to 123
4. Procedures
The SUA layer needs to respond to various local primitives it receives from other layers as well as the messages that it receives from the peer SUA layer. This section describes the SUA procedures in response to these events.4.1. Procedures to Support the SUA-User Layer
4.1.1. Receipt of Primitives from SCCP
When an SCCP Subsystem Management (SCMG) message is received from the SS7 network, the SGP needs to determine whether there are concerned Application Servers interested in subsystem status changes. The SUA management function is informed with the N-State or N-Coord primitive upon which it formats and transfers the applicable SNMM message to the list of concerned ASPs using stream ID "0". When MTP-3 Management indications are received (MTP-PAUSE, MTP- RESUME, MTP-STATUS), SCCP Subsystem Management determines whether there are concerned local SCCP-users. When these local SCCP-users are in fact Application Servers, serviced by ASPs, SUA management is informed with the N-PCSTATE indication primitive upon which it formats and transfers the applicable SNM message (DUNA, DAVA, DRST or SCON) to the list of concerned ASPs using stream ID "0". The SUA message distribution function determines the Application Server (AS) based on comparing the information in the N-UNITDATA request primitive with a provisioned Routing Key. From the list of ASPs within the AS table, an ASP in the ASP-ACTIVE state is selected and a DATA message is constructed and issued on the corresponding SCTP association. If more than one ASP is in the ASP- ACTIVE state (i.e., traffic is to be load-shared across more than one ASP), one of the ASPs in the ASP_ACTIVE state is selected from the list. If the ASPs are in Broadcast Mode, all active ASPs will be selected and the message sent to each of the active ASPs. The selection algorithm is implementation dependent but could, for example, be round robin or based on the SLS. The appropriate
selection algorithm must be chosen carefully as it is dependent on application assumptions and understanding of the degree of state coordination between the ASP_ACTIVE ASPs in the AS. In addition, the message needs to be sent on the appropriate SCTP stream, again taking care to meet the message sequencing needs of the signalling application. DATA messages MUST be sent on an SCTP stream other than stream '0' when there is more than one stream. When there is no Routing Key match, or only a partial match, for an incoming SS7 message, a default treatment MAY be specified. Possible solutions are to provide a default Application Server at the SGP that directs all unallocated traffic to a (set of) default ASP(s), or to drop the message and provide a notification to Layer Management in an M-ERROR indication primitive. The treatment of unallocated traffic is implementation dependent.4.2. Receipt of Primitives from the Layer Management
On receiving primitives from the local Layer Management, the SUA layer will take the requested action and provide an appropriate response primitive to Layer Management. An M-SCTP_ESTABLISH request primitive from Layer Management at an ASP or IPSP will initiate the establishment of an SCTP association. The SUA layer will attempt to establish an SCTP association with the remote SUA peer by sending an SCTP-ASSOCIATE primitive to the local SCTP layer. When an SCTP association has been successfully established, the SCTP will send an SCTP-COMMUNICATION_UP notification primitive to the local SUA layer. At the ASP or IPSP that initiated the request, the SUA layer will send an M-SCTP_ESTABLISH confirm primitive to Layer Management when the association setup is complete. At the peer SUA layer, an M-SCTP_ESTABLISH indication primitive is sent to Layer Management upon successful completion of an incoming SCTP association setup. An M-SCTP_RELEASE request primitive from Layer Management initiates the shutdown of an SCTP association. The SUA layer accomplishes a graceful shutdown of the SCTP association by sending an SCTP-SHUTDOWN primitive to the SCTP layer. When the graceful shutdown of the SCTP association has been accomplished, the SCTP layer returns an SCTP-SHUTDOWN_COMPLETE notification primitive to the local SUA layer. At the SUA Layer that initiated the request, the SUA layer will send an M-SCTP_RELEASE confirm primitive to Layer Management when the association shutdown
is complete. At the peer SUA Layer, an M-SCTP_RELEASE indication primitive is sent to Layer Management upon abort or successful shutdown of an SCTP association. An M-SCTP_STATUS request primitive supports a Layer Management query of the local status of a particular SCTP association. The SUA layer simply maps the M-SCTP_STATUS request primitive to an SCTP-STATUS primitive to the SCTP layer. When the SCTP responds, the SUA layer maps the association status information to an M-SCTP_STATUS confirm primitive. No peer protocol is invoked. Similar LM-to-SUA-to-SCTP and/or SCTP-to-SUA-to-LM primitive mappings can be described for the various other SCTP Upper Layer primitives in RFC 2960 [2960] such as INITIALIZE, SET PRIMARY, CHANGE HEARTBEAT, REQUEST HEARTBEAT, GET SRTT REPORT, SET FAILURE THRESHOLD, SET PROTOCOL PARAMETERS, DESTROY SCTP INSTANCE, SEND FAILURE, AND NETWORK STATUS CHANGE. Alternatively, these SCTP Upper Layer primitives (and Status as well) can be considered for modeling purposes as a Layer Management interaction directly with the SCTP Layer. M-NOTIFY indication and M-ERROR indication primitives indicate to Layer Management the notification or error information contained in a received SUA Notify or Error message respectively. These indications can also be generated based on local SUA events. An M-ASP_STATUS request primitive supports a Layer Management query of the status of a particular local or remote ASP. The SUA layer responds with the status in an M-ASP_STATUS confirm primitive. No SUA peer protocol is invoked. An M-AS_STATUS request supports a Layer Management query of the status of a particular AS. The SUA responds with an M-AS_STATUS confirm primitive. No SUA peer protocol is invoked. M-ASP_UP request, M-ASP_DOWN request, M-ASP_ACTIVE request and M- ASP_INACTIVE request primitives allow Layer Management at an ASP to initiate state changes. Upon successful completion, a corresponding confirm primitive is provided by the SUA layer to Layer Management. If an invocation is unsuccessful, an Error indication primitive is provided in the primitive. These requests result in outgoing ASP Up, ASP Down, ASP Active and ASP Inactive messages to the remote SUA peer at an SGP or IPSP.
4.2.1. Receipt of SUA Peer Management Messages
Upon successful state changes resulting from reception of ASP Up, ASP Down, ASP Active and ASP Inactive messages from a peer SUA, the SUA layer MAY invoke corresponding M-ASP_UP, M-ASP_DOWN, M-ASP_ACTIVE and M-ASP_INACTIVE, M-AS_ACTIVE, M-AS_INACTIVE, and M-AS_DOWN indication primitives to the local Layer Management. M-NOTIFY indication and M-ERROR indication primitives indicate to Layer Management the notification or error information contained in a received SUA Notify or Error message. These indications can also be generated based on local SUA events. All non-Transfer and non-SSNM messages, except BEAT and BEAT Ack, SHOULD be sent with sequenced delivery to ensure ordering. All non- Transfer messages, with the exception of ASPTM, BEAT and BEAT Ack messages SHOULD be sent on SCTP stream '0'. ASPTM messages MAY be sent on one of the streams used to carry data traffic related to the Routing Context(s), to minimize possible message loss. BEAT and BEAT Ack messages MAY be sent using out-of-order delivery, and MAY be sent on any stream.4.3. AS and ASP State Maintenance
The SUA layer on the SGP maintains the state of each remote ASP, in each Application Server that the ASP is configured to receive traffic, as input to the SUA message distribution function. Similarly, where IPSPs use SUA in a point-to-point fashion, the SUA layer in an IPSP maintains the state of remote IPSPs. Two IPSP models are defined with regards to the number of messages that are needed to a IPSP state change. They are defined as follows: 1. IPSP Single Exchange (SE) model. Only a single exchange of ASPTM or ASPSM messages is needed to change the IPSP state. This means that a set of request from one end and acknowledge from the other will be enough. 2. IPSP Double Exchange (DE) model. Both IPSPs have to send request messages and both IPSPs have to acknowledge the request messages from the other end. This results in a double exchange of ASPTM and ASPSM message, one from each end. This configuration supports dynamic routing key configuration by using RKM messages in the same way as ASP-SGP scenario. To ensure interoperability, an SUA implementation supporting IPSP communication MUST support IPSP SE model and MAY implement IPSP DE model.
In section 4.3.1: ASP/IPSP States, only the SGP-ASP and the IPSP SE scenarios are described. For the IPSP DE model, both IPSPs MUST follow the SGP side of the SGP-ASP procedures. In section 4.3.2, only the SGP-ASP scenario is described. All of the procedures referring to an AS served by ASPs are also applicable to ASs served by IPSPs. In section 4.3.3, only the Management procedures for the SGP-ASP scenario are described. The corresponding Management procedures for IPSPs are directly inferred. The remaining sections contain specific IPSP Considerations subsections.4.3.1. ASP States
The state of each remote ASP/IPSP, in each AS that it is configured to operate, is maintained in the peer SUA layer (i.e., in the SGP or peer IPSP, respectively). The state of a particular ASP/IPSP in a particular AS changes due to events. The events include: * Reception of messages from the peer SUA layer at the ASP/IPSP; * Reception of some messages from the peer SUA layer at other ASPs/IPSPs in the AS (e.g., ASP Active message indicating "Override"); * Reception of indications from the SCTP layer; or * Local Management intervention. The ASP/IPSP state transition diagram is shown in Figure 1. The possible states of an ASP/IPSP are: ASP-DOWN: The remote SUA peer at the ASP/IPSP is unavailable and/or the related SCTP association is down. Initially all ASPs/IPSPs will be in this state. An ASP/IPSP in this state SHOULD NOT be sent any SUA messages, with the exception of Heartbeat, ASP Down Ack and Error messages. ASP-INACTIVE: The remote SUA peer at the ASP/IPSP is available (and the related SCTP association is up) but application traffic is stopped. In this state the ASP/IPSP SHOULD NOT be sent any DATA or SSNM messages for the AS for which the ASP/IPSP is inactive. ASP-ACTIVE: The remote SUA peer at the ASP/IPSP is available and application traffic is active (for a particular Routing Context or set of Routing Contexts).
Figure 1: ASP/IPSP State Transition Diagram, per AS +--------------+ | | +----------------------| ASP-ACTIVE | | Other ASP/ +-------| | | IPSP in AS | +--------------+ | Overrides | ^ | | | ASPAC/ | | ASPIA/ | |[ASPAC-Ack]| | [ASPIA-Ack] | | | v | | +--------------+ | | | | | +------>| ASP-INACTIVE | | | | | +--------------+ | ^ | ASPDN/ | | | ASPDN / [ASPDN-Ack/]| ASPUP/ | | [ASPDN-Ack /] SCTP CDI/ | [ASPUP-Ack] | | SCTP CDI/ SCTP RI | | | SCTP RI | | v | +--------------+ | | | +--------------------->| ASP-DOWN | | | +--------------+ The transitions in brackets are just valid for the IPSP SE model communication while the rest are valid for both ASPs and IPSPs. SCTP CDI: The SCTP CDI denotes the local SCTP layer's Communication Down Indication to the Upper Layer Protocol (SUA) on an SGP. The local SCTP layer will send this indication when it detects the loss of connectivity to the ASP's peer SCTP layer. SCTP CDI is understood as either a SHUTDOWN_COMPLETE notification or COMMUNICATION_LOST notification from the SCTP layer. SCTP RI: The local SCTP layer's Restart indication to the upper layer protocol (SUA) on an SG. The local SCTP will send this indication when it detects a restart from the ASP's peer SCTP layer.4.3.2. AS States
The state of the AS is maintained in the SUA layer on the SGP. The state of an AS changes due to events. These events include:
* ASP state transitions
* Recovery timer triggers
The possible states of an AS are:
AS-DOWN: The Application Server is unavailable. This state
implies that all related ASPs are in the ASP-DOWN state
for this AS. Initially the AS will be in this state.
An Application Server is in the AS-DOWN state before it
can be removed from a configuration.
AS-INACTIVE: The Application Server is available but no application
traffic is active (i.e., one or more related ASPs are in
the ASP-INACTIVE state, but none in the ASP-ACTIVE
state). The recovery timer T(r) is not running or has
expired.
AS-ACTIVE : The Application Server is available and application
traffic is active. This state implies that at least one
ASP is in the ASP-ACTIVE state.
AS-PENDING: An active ASP has transitioned to ASP-INACTIVE or ASP-
DOWN and it was the last remaining active ASP in the AS.
A recovery timer T(r) SHOULD be started and all incoming
signalling messages SHOULD be queued by the SGP. If an
ASP becomes ASP-ACTIVE before T(r) expires, the AS is
moved to the AS-ACTIVE state and all the queued messages
will be sent to the ASP.
If T(r) expires before an ASP becomes ASP-ACTIVE, and the SGP has no
alternative, the SGP may stop queueing messages and discard all
previously queued messages. The AS will move to the AS-INACTIVE
state if at least one ASP is in ASP-INACTIVE state, otherwise it will
move to AS-DOWN state.
Figure 2 shows an example AS state machine for the case where the
AS/ASP data is provisioned. For other cases where the AS/ASP
configuration data is created dynamically, there would be differences
in the state machine, especially at creation of the AS.
For example, where the AS/ASP configuration data is not created until
Registration of the first ASP, the AS-INACTIVE state is entered
directly upon the first successful REG REQ from an ASP. Another
example is where the AS/ASP configuration data is not created until
the first ASP successfully enters the ASP-ACTIVE state. In this case
the AS-ACTIVE state is entered directly.
Figure 2: AS State Transition Diagram
+----------+ one ASP trans to ACTIVE +-------------+
| AS- |---------------------------->| AS- |
| INACTIVE | | ACTIVE |
| |<--- | |
+----------+ \ +-------------+
^ | \ Tr Expiry, ^ |
| | \ at least one | |
| | \ ASP in ASP-INACTIVE | |
| | \ | |
| | \ | |
| | \ | |
one ASP | | all ASP \ one ASP | | Last ACTIVE
trans | | trans to \ trans to | | ASP trans to
to | | ASP-DOWN -------\ ASP- | | ASP-INACTIVE
ASP- | | \ ACTIVE | | or ASP-DOWN
INACTIVE| | \ | | (start Tr)
| | \ | |
| | \ | |
| v \ | v
+----------+ \ +-------------+
| | --| |
| AS-DOWN | | AS-PENDING |
| | | (queueing) |
| |<----------------------------| |
+----------+ Tr Expiry and no ASP +-------------+
in ASP-INACTIVE state
Tr = Recovery Timer
4.3.2.1. IPSP Considerations
The AS state diagram for the AS-SG case is applicable for IPSP
communication.
4.3.3. SUA Management Procedures for Primitives
Before the establishment of an SCTP association the ASP state at both
the SGP and ASP is assumed to be in the state ASP-DOWN.
Once the SCTP association is established (see Section 4.2.1) and
assuming that the local SUA-User is ready, the local SUA ASP
Maintenance (ASPM) function will initiate the relevant procedures,
using the ASP Up/ASP Down/ASP Active/ASP Inactive messages to convey
the ASP state to the SGP (see Section 4.3.4).
If the SUA layer subsequently receives an SCTP-COMMUNICATION_DOWN or SCTP-RESTART indication primitive from the underlying SCTP layer, it will inform the Layer Management by invoking the M-SCTP_STATUS indication primitive. The state of the ASP will be moved to ASP- DOWN. In the case of SCTP-COMMUNICATION_DOWN, the SCTP client MAY try to reestablish the SCTP association. This MAY be done by the SUA layer automatically, or Layer Management MAY reestablish using the M- SCTP_ESTABLISH request primitive. In the case of an SCTP-RESTART indication at an ASP, the ASP is now considered by its SUA peer to be in the ASP-DOWN state. The ASP, if it is to recover, must begin any recovery with the ASP-Up procedure.4.3.4. ASPM Procedures for Peer-to-Peer Messages
4.3.4.1. ASP Up Procedures
After an ASP has successfully established an SCTP association to an SGP, the SGP waits for the ASP to send an ASP Up message, indicating that the ASP SUA peer is available. The ASP is always the initiator of the ASP Up message. This action MAY be initiated at the ASP by an M-ASP_UP request primitive from Layer Management or MAY be initiated automatically by an SUA management function. When an ASP Up message is received at an SGP and internally the remote ASP is in the ASP-DOWN state and not considered locked-out for local management reasons, the SGP marks the remote ASP in the state ASP-INACTIVE and informs Layer Management with an M-ASP_Up indication primitive. If the SGP is aware, via current configuration data, which Application Servers the ASP is configured to operate in, the SGP updates the ASP state to ASP-INACTIVE in each AS that it is a member. Alternatively, the SGP may move the ASP into a pool of Inactive ASPs available for future configuration within Application Server(s), determined in a subsequent Registration Request or ASP Active procedure. If the ASP Up message contains an ASP Identifier, the SGP should save the ASP Identifier for that ASP. The SGP MUST send an ASP Up Ack message in response to a received ASP Up message even if the ASP is already marked as ASP-INACTIVE at the SGP. If for any local reason (e.g., management lock-out) the SGP cannot respond with an ASP Up Ack message, the SGP responds to an ASP Up message with an Error message with Reason "Refused - Management Blocking".
At the ASP, the ASP Up Ack message received is not acknowledged. Layer Management is informed with an M-ASP_UP confirm primitive. When the ASP sends an ASP Up message it starts timer T(ack). If the ASP does not receive a response to an ASP Up message within T(ack), the ASP MAY restart T(ack) and resend ASP Up messages until it receives an ASP Up Ack message. T(ack) is provisioned, with a default of 2 seconds. Alternatively, retransmission of ASP Up messages MAY be put under control of Layer Management. In this method, expiry of T(ack) results in an M-ASP_UP confirm primitive carrying a negative indication. The ASP must wait for the ASP Up Ack message before sending any other SUA messages (e.g., ASP Active or REG REQ). If the SGP receives any other SUA messages before ASPUP message is received (other than ASPDN - see section 4.3.4.2), the SGP SHOULD discard them. If an ASP Up message is received and internally the remote ASP is in the ASP-ACTIVE state, an ASP Up Ack message is returned, as well as an Error message ("Unexpected Message), and the remote ASP state is changed to ASP-INACTIVE in all relevant Application Servers. If an ASP Up message is received and internally the remote ASP is already in the ASP-INACTIVE state, an ASP Up Ack message is returned and no further action is taken.4.3.4.1.1. SUA Version Control
If an ASP Up message with an unsupported version is received, the receiving end responds with an Error message, indicating the version the receiving node supports and notifies Layer Management. This is useful when protocol version upgrades are being performed in a network. A node upgraded to a newer version should support the older versions used on other nodes it is communicating with. Because ASPs initiate the ASP Up procedure it is assumed that the Error message would normally come from the SGP.4.3.4.1.2. IPSP Considerations
An IPSP may be considered in the ASP-INACTIVE state after and ASPUP or ASPUP Ack has been received from it. An IPSP can be considered in the ASP-DOWN state after an ASPDN or ASPDN Ack has been received from it. The IPSP may inform Layer Management of the change in state of the remote IPSP using M-ASP_UP or M-ASP_DN indication or confirmation primitives.
Alternatively, when using IPSP DE model, an interchange of ASP Up messages from each end MUST be performed. Four messages are needed for completion. If for any local reason (e.g., management lock-out) and IPSP cannot respond to an ASP Up message with an ASP Up Ack message, it responds to an ASP Up message with an Error message with Reason "Refused - Management Blocking" and leaves the remote IPSP in the ASP-DOWN state.4.3.4.2. ASP Down Procedures
The ASP will send an ASP Down message to an SGP when the ASP wishes to be removed from service in all Application Servers that it is a member and no longer receive any Connectionless or Connection - Oriented, SSNM or ASPTM messages. This action MAY be initiated at the ASP by an M-ASP_DOWN request primitive from Layer Management or MAY be initiated automatically by an SUA management function. Whether the ASP is permanently removed from any AS is a function of configuration management. In the case where the ASP previously used the Registration procedures (see Section 4.4.1) to register within Application Servers but has not deregistered from all of them prior to sending the ASP Down message, the SGP MUST consider the ASP as deregistered in all Application Servers that it is still a member. The SGP marks the ASP as ASP-DOWN, informs Layer Management with an M-ASP_Down indication primitive, and returns an ASP Down Ack message to the ASP. The SGP MUST send an ASP Down Ack message in response to a received ASP Down message from the ASP even if the ASP is already marked as ASP-DOWN at the SGP. At the ASP, the ASP Down Ack message received is not acknowledged. Layer Management is informed with an M-ASP_DOWN confirm primitive. If the ASP receives an ASP Down Ack without having sent an ASP Down message, the ASP should now consider itself as in the ASP-DOWN state. If the ASP was previously in the ASP-ACTIVE or ASP_INACTIVE state, the ASP should then initiate procedures to return itself to its previous state. When the ASP sends an ASP Down message it starts timer T(ack). If the ASP does not receive a response to an ASP Down message within T(ack), the ASP MAY restart T(ack) and resend ASP Down messages until it receives an ASP Down Ack message. T(ack) is provisioned, with a default of 2 seconds. Alternatively, retransmission of ASP Down
messages MAY be put under control of Layer Management. In this method, expiry of T(ack) results in an M-ASP_DOWN confirm primitive carrying a negative indication.4.3.4.3. ASP Active Procedures
Anytime after the ASP has received an ASP Up Ack message from the SGP or IPSP, the ASP MAY send an ASP Active message to the SGP indicating that the ASP is ready to start processing traffic. This action MAY be initiated at the ASP by an M-ASP_ACTIVE request primitive from Layer Management or MAY be initiated automatically by an SUA management function. In the case where an ASP wishes to process the traffic for more than one Application Server across a common SCTP association, the ASP Active message(s) SHOULD contain a list of one or more Routing Contexts to indicate for which Application Servers the ASP Active message applies. It is not necessary for the ASP to include all Routing Contexts of interest in a single ASP Active message, thus requesting to become active in all Routing Contexts at the same time. Multiple ASP Active messages MAY be used to activate within the Application Servers independently, or in sets. In the case where an ASP Active message does not contain a Routing Context parameter, the receiver must know, via configuration data, which Application Server(s) the ASP is a member. For the Application Servers that the ASP can be successfully activated, the SGP or IPSP responds with one or more ASP Active Ack messages, including the associated Routing Context(s) and reflecting any Traffic Mode Type value present in the related ASP Active message. The Routing Context parameter MUST be included in the ASP Active Ack message(s) if the received ASP Active message contained any Routing Contexts. Depending on any Traffic Mode Type request in the ASP Active message, or local configuration data if there is no request, the SGP moves the ASP to the correct ASP traffic state within the associated Application Server(s). Layer Management is informed with an M-ASP_Active indication. If the SGP or IPSP receives any Data messages before an ASP Active message is received, the SGP or IPSP MAY discard them. By sending an ASP Active Ack message, the SGP or IPSP is now ready to receive and send traffic for the related Routing Context(s). The ASP SHOULD NOT send Data or SSNM messages for the related Routing Context(s) before receiving an ASP Active Ack message, or it will risk message loss. Multiple ASP Active Ack messages MAY be used in response to an ASP Active message containing multiple Routing Contexts, allowing the SGP or IPSP to independently acknowledge the ASP Active message for different (sets of) Routing Contexts. The SGP or IPSP MUST send an Error message ("Invalid Routing Context") for each Routing Context value that cannot be successfully activated.
In the case where an "out-of-the-blue" ASP Active message is received
(i.e., the ASP has not registered with the SG or the SG has no static
configuration data for the ASP), the message MAY be silently
discarded.
The SGP MUST send an ASP Active Ack message in response to a received
ASP Active message from the ASP, if the ASP is already marked in the
ASP-ACTIVE state at the SGP.
At the ASP, the ASP Active Ack message received is not acknowledged.
Layer Management is informed with an M-ASP_ACTIVE confirm primitive.
It is possible for the ASP to receive Data message(s) before the ASP
Active Ack message as the ASP Active Ack and Data messages from an SG
or IPSP may be sent on different SCTP streams. Message loss is
possible, as the ASP does not consider itself in the ASP-ACTIVE state
until reception of the ASP Active Ack message.
When the ASP sends an ASP Active message it starts timer T(ack). If
the ASP does not receive a response to an ASP Active message within
T(ack), the ASP MAY restart T(ack) and resend ASP Active messages
until it receives an ASP Active Ack message. T(ack) is provisioned,
with a default of 2 seconds. Alternatively, retransmission of ASP
Active messages MAY be put under control of Layer Management. In
this method, expiry of T(ack) results in an M-ASP_ACTIVE confirm
primitive carrying a negative indication.
There are three modes of Application Server traffic handling in the
SGP SUA layer: Override, Loadshare and Broadcast. When included, the
Traffic Mode Type parameter in the ASP Active message indicates the
traffic-handling mode to be used in a particular Application Server.
If the SGP determines that the mode indicated in an ASP Active
message is unsupported or incompatible with the mode currently
configured for the AS, the SGP responds with an Error message
("Unsupported / Invalid Traffic Handling Mode"). If the traffic-
handling mode of the Application Server is not already known via
configuration data, then the traffic-handling mode indicated in the
first ASP Active message causing the transition of the Application
Server state to AS-ACTIVE MAY be used to set the mode.
In the case of an Override mode AS, reception of an ASP Active
message at an SGP causes the (re)direction of all traffic for the AS
to the ASP that sent the ASP Active message. Any previously active
ASP in the AS is now considered to be in state ASP-INACTIVE and
SHOULD no longer receive traffic from the SGP within the AS. The SGP
or IPSP then MUST send a Notify message ("Alternate ASP Active") to
the previously active ASP in the AS, and SHOULD stop traffic to/from
that ASP. The ASP receiving this Notify MUST consider itself now in the ASP-INACTIVE state, if it is not already aware of this via inter-ASP communication with the Overriding ASP. In the case of a Loadshare mode AS, reception of an ASP Active message at an SGP or IPSP causes the direction of traffic to the ASP sending the ASP Active message, in addition to all the other ASPs that are currently active in the AS. The algorithm at the SGP for loadsharing traffic within an AS to all the active ASPs is implementation dependent. The algorithm could, for example, be round robin or based on information in the Data message (e.g., the SLS or SSN). An SGP or IPSP, upon reception of an ASP Active message for the first ASP in a Loadshare AS, MAY choose not to direct traffic to a newly active ASP until it determines that there are sufficient resources to handle the expected load (e.g., until there are "n" ASPs in state ASP-ACTIVE in the AS). All ASPs within a load-sharing mode AS must be able to process any Data message received for the AS, to accommodate any potential fail- over or rebalancing of the offered load. In the case of a Broadcast mode AS, reception of an ASP Active message at an SGP or IPSP causes the direction of traffic to the ASP sending the ASP Active message, in addition to all the other ASPs that are currently active in the AS. The algorithm at the SGP for broadcasting traffic within an AS to all the active ASPs is a simple broadcast algorithm, where every message is sent to each of the active ASPs. An SGP or IPSP, upon reception of an ASP Active message for the first ASP in a Broadcast AS, MAY choose not to direct traffic to a newly active ASP until it determines that there are sufficient resources to handle the expected load (e.g., until there are "n" ASPs in state ASP-ACTIVE in the AS). Whenever an ASP in a Broadcast mode AS becomes ASP-ACTIVE, the SGP MUST tag the first DATA message broadcast in each traffic flow with a unique Correlation Id parameter. The purpose of this Correlation Id is to permit the newly active ASP to synchronize its processing of traffic in each traffic flow with the other ASPs in the broadcast group.
4.3.4.3.1. IPSP Considerations
Either of the IPSPs can initiate communication. When an IPSP receives an ASP Active, it should mark the peer as ASP-ACTIVE and return an ASP Active Ack message. An ASP receiving an ASP Active Ack message may mark the peer as ASP-Active, if it is not already in the ASP-ACTIVE state. Alternatively, when using IPSP DE model, an interchange of ASP Active messages from each end MUST be performed. Four messages are needed for completion.4.3.4.4. ASP Inactive Procedures
When an ASP wishes to withdraw from receiving traffic within an AS, or the ASP wants to initiate the process of deactivation, the ASP sends an ASP Inactive message to the SGP or IPSP. An ASP Inactive message MUST be always responded by the peer (although other messages may be sent in the middle): - If the corresponding RK is registered (statically or dynamically), the peer should respond with an ASP Inactive Ack message. - If the RK is not registered, or the RC information is not valid, the peer must respond with an ERROR message with Error Code = "Invalid Routing Context". - If the RC is missing and its specification is needed according to the used configuration, the peer must respond with an ERROR message with Error Code = "No Configured AS for ASP". The action of sending the ASP Inactive message MAY be initiated at the ASP by an M-ASP_INACTIVE request primitive from Layer Management or MAY be initiated automatically by an SUA management function. In the case where an ASP is processing the traffic for more than one Application Server across a common SCTP association, the ASP Inactive message contains one or more Routing Contexts to indicate for which Application Servers the ASP Inactive message applies. In the case where an ASP Inactive message does not contain a Routing Context parameter, the receiver must know, via configuration data, which Application Servers the ASP is a member and move the ASP to the ASP-INACTIVE state in each all Application Servers. In the case of an Override mode AS, where another ASP has already taken over the traffic within the AS with an ASP Active ("Override") message, the ASP that sends the ASP Inactive message is already
considered by the SGP to be in state ASP-INACTIVE. An ASP Inactive
Ack message is sent to the ASP, after ensuring that all traffic is
stopped to the ASP.
In the case of a Loadshare mode AS, the SGP moves the ASP to the
ASP-INACTIVE state and the AS traffic is reallocated across the
remaining ASPs in the state ASP-ACTIVE, as per the loadsharing
algorithm currently used within the AS. A Notify message
("Insufficient ASP resources active in AS") MAY be sent to all
inactive ASPs, if required. An ASP Inactive Ack message is sent to
the ASP after all traffic is halted and Layer Management is informed
with an M-ASP_INACTIVE indication primitive.
In the case of a Broadcast mode AS, the SGP moves the ASP to the
ASP-INACTIVE state and the AS traffic is broadcast only to the
remaining ASPs in the state ASP-ACTIVE. A Notify message
("Insufficient ASP resources active in AS") MAY be sent to all
inactive ASPs, if required. An ASP Inactive Ack message is sent to
the ASP after all traffic is halted and Layer Management is informed
with an M-ASP_INACTIVE indication primitive.
Multiple ASP Inactive Ack messages MAY be used in response to an ASP
Inactive message containing multiple Routing Contexts, allowing the
SGP or IPSP to independently acknowledge for different (sets of)
Routing Contexts. The SGP or IPSP sends an Error message ("Invalid
Routing Context") message for each invalid or not configured Routing
Context value in a received ASP Inactive message.
The SGP MUST send an ASP Inactive Ack message in response to a
received ASP Inactive message from the ASP and the ASP is already
marked as ASP-INACTIVE at the SGP.
At the ASP, the ASP Inactive Ack message received is not
acknowledged. Layer Management is informed with an M-ASP_INACTIVE
confirm primitive. If the ASP receives an ASP Inactive Ack without
having sent an ASP Inactive message, the ASP should now consider
itself as in the ASP-INACTIVE state. If the ASP was previously in
the ASP-ACTIVE state, the ASP should then initiate procedures to
return itself to its previous state. When the ASP sends an ASP
Inactive message it starts timer T(ack). If the ASP does not receive
a response to an ASP Inactive message within T(ack), the ASP MAY
restart T(ack) and resend ASP Inactive messages until it receives an
ASP Inactive Ack message. T(ack) is provisioned, with a default of 2
seconds. Alternatively, retransmission of ASP Inactive messages MAY
be put under control of Layer Management. In this method, expiry of
T(ack) results in a M-ASP_Inactive confirm primitive carrying a
negative indication.
If no other ASPs in the Application Server are in the state ASP-
ACTIVE, the SGP MUST send a Notify message ("AS-Pending") to all of
the ASPs in the AS which are in the state ASP-INACTIVE. The SGP
SHOULD start buffering the incoming messages for T(r) seconds, after
which messages MAY be discarded. T(r) is configurable by the network
operator. If the SGP receives an ASP Active message from an ASP in
the AS before expiry of T(r), the buffered traffic is directed to
that ASP and the timer is cancelled. If T(r) expires, the AS is
moved to the AS-INACTIVE state.
4.3.4.4.1. IPSP Considerations
An IPSP may be considered in the ASP-INACTIVE state by a remote IPSP
after an ASP Inactive or ASP Inactive Ack message has been received
from it.
Alternatively, when using IPSP DE model, an interchange of ASP
Inactive messages from each end MUST be performed. Four messages are
needed for completion.
4.3.4.5. Notify Procedures
A Notify message reflecting a change in the AS state MUST be sent to
all ASPs in the AS, except those in the ASP-DOWN state, with
appropriate Status Information and any ASP Identifier of the failed
ASP. At the ASP, Layer Management is informed with an M-NOTIFY
indication primitive. The Notify message must be sent whether the AS
state change was a result of an ASP failure or reception of an ASP
State management (ASPSM) / ASP Traffic Management (ASPTM) message.
In the second case, the Notify message MUST be sent after any ASP
State or Traffic Management related acknowledgement messages (e.g.,
ASP Up Ack, ASP Down Ack, ASP Active Ack, or ASP Inactive Ack).
In the case where a Notify ("AS-PENDING") message is sent by an SGP
that now has no ASPs active to service the traffic, or where a Notify
("Insufficient ASP resources active in AS") message MUST be sent in
the Loadshare or Broadcast mode, the Notify message does not
explicitly compel the ASP(s) receiving the message to become active.
The ASPs remain in control of what (and when) traffic action is
taken.
In the case where a Notify message does not contain a Routing Context
parameter, the receiver must know, via configuration data, of which
Application Servers the ASP is a member and take the appropriate
action in each AS.
4.3.4.5.1. IPSP Considerations (NTFY)
Notify works in the same manner as in the SG-AS case. One of the IPSPs can send this message to any remote IPSP that is not in the ASP-DOWN state.4.3.4.6. Heartbeat Procedures
The optional Heartbeat procedures MAY be used when operating over transport layers that do not have their own heartbeat mechanism for detecting loss of the transport association (i.e., other than SCTP). Either SUA peer may optionally send Heartbeat messages periodically, subject to a provisioned timer T(beat). Upon receiving a Heartbeat message, the SUA peer MUST respond with a Heartbeat Ack message. If no Heartbeat Ack message (or any other SUA message) is received from the SUA peer within 2*T(beat), the remote SUA peer is considered unavailable. Transmission of Heartbeat messages is stopped and the signalling process SHOULD attempt to reestablish communication if it is configured as the client for the disconnected SUA peer. The Heartbeat message may optionally contain an opaque Heartbeat Data parameter that MUST be echoed back unchanged in the related Heartbeat Ack message. The sender, upon examining the contents of the returned Heartbeat Ack message, MAY choose to consider the remote SUA peer as unavailable. The contents/format of the Heartbeat Data parameter is implementation-dependent and only of local interest to the original sender. The contents may be used, for example, to support a Heartbeat sequence algorithm (to detect missing Heartbeats), and/or a timestamp mechanism (to evaluate delays). Note: Heartbeat related events are not shown in Figure 2 "ASP state transition diagram".4.4. Routing Key Management Procedures
4.4.1. Registration
An ASP MAY dynamically register with an SGP as an ASP within an Application Server using the REG REQ message. A Routing Key parameter in the REG REQ message specifies the parameters associated with the Routing Key. The SGP examines the contents of the received Routing Key parameter and compares it with the currently provisioned Routing Keys. If the received Routing Key matches an existing SGP Routing Key entry, and the ASP is not currently included in the list of ASPs for the related
Application Server, the SGP MAY authorize the ASP to be added to the AS. Or, if the Routing Key does not currently exist and the received Routing Key data is valid and unique, an SGP supporting dynamic configuration MAY authorize the creation of a new Routing Key and related Application Server and add the ASP to the new AS. In either case, the SGP returns a Registration Response message to the ASP, containing the same Local-RK-Identifier as provided in the initial request, and a Registration Result "Successfully Registered". A unique Routing Context value assigned to the SGP Routing Key is included. The method of Routing Context value assignment at the SGP is implementation dependent but must be guaranteed to be unique for each Application Server or Routing Key supported by the SGP. If the SGP determines that the received Routing Key data is invalid, or contains invalid parameter values, the SGP returns a Registration Response message to the ASP, containing a Registration Result "Error - Invalid Routing Key", "Error - Invalid DPC", "Error - Invalid Network Appearance" as appropriate. If the SGP does not support the registration procedure, the SGP returns an Error message to the ASP, with an error code of "Unsupported Message Type". If the SGP determines that a unique Routing Key cannot be created, the SGP returns a Registration Response message to the ASP, with a Registration Status of "Error - Cannot Support Unique Routing". An incoming signalling message received at an SGP should not match against more than one Routing Key. If the SGP does not authorize the registration request, the SGP returns a REG RSP message to the ASP containing the Registration Result "Error - Permission Denied". If an SGP determines that a received Routing Key does not currently exist and the SGP does not support dynamic configuration, the SGP returns a Registration Response message to the ASP, containing a Registration Result "Error - Routing Key not Currently Provisioned". If an SGP determines that a received Routing Key does not currently exist and the SGP supports dynamic configuration but does not have the capacity to add new Routing Key and Application Server entries, the SGP returns a Registration Response message to the ASP, containing a Registration Result "Error - Insufficient Resources". If an SGP determines that one or more of the Routing Key parameters are not supported for the purpose of creating new Routing Key entries, the SGP returns a Registration Response message to the ASP,
containing a Registration Result "Error - Unsupported RK parameter field". This result MAY be used if, for example, the SGP does not support RK Address parameter. A Registration Response "Error - Unsupported Traffic Handling Mode" is returned if the Routing Key in the REG REQ contains a Traffic Handling Mode that is inconsistent with the presently configured mode for the matching Application Server. An ASP MAY register multiple Routing Keys at once by including a number of Routing Key parameters in a single REG REQ message. The SGP MAY respond to each registration request in a single REG RSP message, indicating the success or failure result for each Routing Key in a separate Registration Result parameter. Alternatively the SGP MAY respond with multiple REG RSP messages, each with one or more Registration Result parameters. The ASP uses the Local-RK-Identifier parameter to correlate the requests with the responses. An ASP MAY modify an existing Routing Key by including a Routing Context parameter in the REG REQ. If the SGP determines that the Routing Context applies to an existing Routing Key, the SG MAY adjust the existing Routing Key to match the new information provided in the Routing Key parameter. A Registration Response "Routing Key Change Refused" is returned if the SGP does not accept the modification of the Routing Key. Upon successful registration of an ASP in an AS, the SGP can now send related SS7 Signalling Network Management messaging, if this did not previously start upon the ASP transitioning to state ASP-INACTIVE.4.4.2. Deregistration
An ASP MAY dynamically deregister with an SGP as an ASP within an Application Server using the DEREG REQ message. A Routing Context parameter in the DEREG REQ message specifies which Routing Keys to deregister. An ASP SHOULD move to the ASP-INACTIVE state for an Application Server before attempting to deregister the Routing Key (i.e., deregister after receiving an ASP Inactive Ack). Also, an ASP SHOULD deregister from all Application Servers that it is a member before attempting to move to the ASP-Down state. The SGP examines the contents of the received Routing Context parameter and validates that the ASP is currently registered in the Application Server(s) related to the included Routing Context(s). If validated, the ASP is deregistered as an ASP in the related Application Server.
The deregistration procedure does not necessarily imply the deletion of Routing Key and Application Server configuration data at the SGP. Other ASPs may continue to be associated with the Application Server, in which case the Routing Key data SHOULD NOT be deleted. If a Deregistration results in no more ASPs in an Application Server, an SGP MAY delete the Routing Key data. The SGP acknowledges the deregistration request by returning a DEREG RSP message to the requesting ASP. The result of the deregistration is found in the Deregistration Result parameter, indicating success or failure with cause. An ASP MAY deregister multiple Routing Contexts at once by including a number of Routing Contexts in a single DEREG REQ message. The SGP MAY respond to each deregistration request in a single DEREG RSP message, indicating the success or failure result for each Routing Context in a separate Deregistration Result parameter.4.4.3. IPSP Considerations (REG/DEREG)
The Registration/Deregistration procedures work in the IPSP cases in the same way as in AS-SG cases. An IPSP may register an RK in the remote IPSP. An IPSP is responsible for deregistering the RKs that it has registered.4.5. Availability and/or Congestion Status of SS7 Destination Support
4.5.1. At an SGP
On receiving a N-STATE, N-PCSTATE and N-INFORM indication primitive from the nodal interworking function at an SGP, the SGP SUA layer will send a corresponding SS7 Signalling Network Management (SNM) DUNA, DAVA, SCON, or DUPU message (see Section 3.4) to the SUA peers at concerned ASPs. The SUA layer must fill in various fields of the SNM messages consistently with the information received in the primitives. The SGP SUA layer determines the set of concerned ASPs to be informed based on the specific SS7 network for which the primitive indication is relevant. In this way, all ASPs configured to send/receive traffic within a particular network appearance are informed. If the SGP operates within a single SS7 network appearance, then all ASPs are informed. DUNA, DAVA, SCON, and DRST messages are sent sequentially and processed at the receiver in the order sent. SCTP stream 0 SHOULD NOT be used. The Unordered bit in the SCTP DATA chunk MAY be used for the SCON message.
Sequencing is not required for the DUPU or DAUD messages, which MAY be sent unordered. SCTP stream 0 is used, with optional use of the Unordered bit in the SCTP DATA chunk.4.5.2. At an ASP
4.5.2.1. Single SG Configurations
At an ASP, upon receiving an SS7 Signalling Network Management (SSNM) message from the remote SUA Peer, the SUA layer invokes the appropriate primitive indications to the resident SUA-Users. Local management is informed. In the case where a local event has caused the unavailability or congestion status of SS7 destinations, the SUA layer at the ASP SHOULD pass up appropriate indications in the primitives to the SUA User, as though equivalent SSNM messages were received. For example, the loss of an SCTP association to an SGP may cause the unavailability of a set of SS7 destinations. N-PCSTATE indication primitives to the SUA User are appropriate. Implementation Note: To accomplish this, the SUA layer at an ASP maintains the status of routes via the SG.4.5.2.2. Multiple SG Configurations
At an ASP, upon receiving a Signalling Network Management message from the remote SUA Peer, the SUA layer updates the status of the affected route(s) via the originating SG and determines, whether or not the overall availability or congestion status of the effected destination(s) has changed. If so, the SUA layer invokes the appropriate primitive indications to the resident SUA-Users. Local management is informed.4.5.3. ASP Auditing
An ASP may optionally initiate an audit procedure to inquire of an SGP the availability and, if the national congestion method with multiple congestion levels and message priorities is used, congestion status of an SS7 destination or set of destinations. A Destination Audit (DAUD) message is sent from the ASP to the SGP requesting the current availability and congestion status of one or more SS7 destinations or subsystems. The DAUD message MAY be sent unordered. The ASP MAY send the DAUD in the following cases:
- Periodic. A Timer originally set upon reception of a DUNA, SCON or
DRST message has expired without a subsequent DAVA,
DUNA, SCON or DRST message updating the
availability/congestion status of the affected
Destination Point Code. The Timer is reset upon issuing
a DAUD. In this case the DAUD is sent to the SGP that
originally sent the SSNM message.
- Isolation. The ASP is newly ASP-ACTIVE or has been isolated from an
SGP for an extended period. The ASP MAY request the
availability/congestion status of one or more SS7
destinations to which it expects to communicate.
Implementation Note:
In the first of the cases above, the auditing procedure must not
be invoked for the case of a received SCON message containing a
congestion level value of "no congestion" or undefined" (i.e.,
congestion Level = "0"). This is because the value indicates
either congestion abatement or that the ITU MTP3 international
congestion method is being used. In the international congestion
method, the MTP3 layer at the SGP does not maintain the congestion
status of any destinations and therefore the SGP cannot provide
any congestion information in response to the DAUD. For the same
reason, in the second of the cases above a DAUD message cannot
reveal any congested destination(s).
The SGP SHOULD respond to a DAUD message with the availability and
congestion status of the subsystem. The status of each SS7
destination or subsystem requested is indicated in a DUNA message (if
unavailable), a DAVA message (if available), or a DRST (if restricted
and the SGP supports this feature). If the SS7 destination or
subsystem is available and congested, the SGP responds with an SCON
message in addition to the DAVA message. If the SS7 destination or
subsystem is restricted and congested, the SGP responds with an SCON
message in addition to the DRST. If the SGP has no information on
the availability / congestion status of the SS7 destination or
subsystem, the SGP responds with a DUNA message, as it has no routing
information to allow it to route traffic to this destination or
subsystem.
An SG MAY refuse to provide the availability or congestion status of
a destination or subsystem if, for example, the ASP is not authorized
to know the status of the destination or subsystem. The SG MAY
respond with an Error Message (Error Code = "Destination Status
Unknown") or Error Message (Error Code = "Subsystem Status Unknown").
4.6. MTP3 Restart
In the case where the MTP3 in the SG undergoes an MTP restart, event communication SHOULD be handled as follows: When the SG discovers SS7 network isolation, the SGPs send an indication to all concerned available ASPs (i.e., ASPs in the ASP- ACTIVE state) using DUNA messages for the concerned destinations. When the SG has completed the MTP Restart procedure, the SUA layer at the SGPs inform all concerned ASPs in the ASP-ACTIVE state of any available/restricted SS7 destinations using the DAVA/DRST message. No message is necessary for those destinations still unavailable after the restart procedure. When the SUA layer at an ASP receives a DUNA message indicating SS7 destination unavailability at an SG, SCCP Users will receive an N- PCSTATE indication and will stop any affected traffic to this destination. When the SUA receives a DAVA/DRST message, SCCP Users will receive an N-PCSTATE indication and can resume traffic to the newly available SS7 destination via this SGP, provided the ASP is in the ASP-ACTIVE state toward this SGP. The ASP MAY choose to audit the availability of unavailable destinations by sending DAUD messages. This would be for example the case when an AS becomes active at an ASP and does not have current destination statuses. If MTP restart is in progress at the SG, the SGP returns a DUNA message for that destination, even if it received an indication that the destination became available or restricted.4.7. SCCP - SUA Interworking at the SG
4.7.1. Segmenting / Reassembly
When it is expected that signalling messages will not fit into a PDU of the most restrictive transport technology used (e.g., 272-SIF of MTP3), then segmenting/reassembly could be performed at the SG, ASP or IPSP. If the SG, ASP or IPSP is incapable of performing a necessary segmentation/reassembly, it can inform the peer of the failure using the appropriate error in a CLDR or RESRE/COERR message.4.7.2. Support for Loadsharing
Within an AS (identified by RK/RC parameters) several loadsharing ASPs may be active.
However, to assure the correct processing of TCAP transactions or SCCP connections, the loadsharing scheme used at the SG must make sure that messages continuing or ending the transactions/connections arrive at the same ASP where the initial message (TC_Query, TC_Begin, CR) was sent to/received from. When the ASP can be identified uniquely based on RK parameters (e.g., unique DPC or GT), loadsharing is not required. When the ASPs in the AS share state or use an internal distribution mechanism, the SG must only take into account the in-sequence-delivery requirement. In case of SCCP CO traffic, when the coupled approach is used, loadsharing of messages other than CR is not required. If these assumptions cannot be made, both SG and ASP should support the following general procedure in a loadsharing environment.4.7.2.1. Association Setup, ASP going active
After association setup and registration, an ASP normally goes active for each AS it registered for. In the ASPAC message, the ASP includes a TID and/or DRN Label Parameter, if applicable for the AS in question. All the ASPs within the AS must specify a unique label at a fixed position in the TID or DRN parameter. The same ASPAC message is sent to each SG used for interworking with the SS7 network. The SG builds, per RK, a list of ASPs that have registered for it. The SG can now build up and update a distribution table for a certain Routing Context, any time the association is (re-)established and the ASP goes active. The SG has to perform some trivial plausibility checks on the parameters: - Start and End parameters values are between 0 and 31 for TID. - Start and End parameters values are between 0 and 23 for DRN - 0 < (Start - End + 1) <= 16 (label length maximum 16-bit) - Start values are the same for each ASP within a RC - End values are the same for each ASP within a RC - TID and DRN Label values must be unique across the RC If any of these checks fail, the SG refuses the ASPAC request, with an error, "Invalid loadsharing label."
4.7.3. Routing and message distribution at the SG
4.7.3.1. TCAP traffic
Messages not containing a destination (or "responding") TID, i.e., Query, Begin, Unidirectional, are loadshared among the available ASPs. Any scheme permitting a fair load distribution among the ASPs is allowed (e.g., round robin). When a destination TID is present, the SG extracts the label and selects the ASP that corresponds with it. If an ASP is not available, the SG may generate (X)UDTS "routing failure", if the return option is used.4.7.3.2. SCCP Connection Oriented traffic
Messages not containing a destination reference number (DRN), i.e., a Connection Request, MAY be loadshared among the available ASPs. The load distribution mechanism is an implementation issue. When a DRN is present, the SG extracts the label and selects the ASP that corresponds with it. If an ASP is not available, the SG discards the message.4.7.4. Multiple SGs, SUA Relay Function
It is important that each ASP send its unique label (within the AS) to each SGP. For a better robustness against association failures, the SGs MAY cooperate to provide alternative routes toward an ASP. Mechanisms for SG cooperation/coordination are outside of the scope of this document.5. Examples of SUA Procedures
The following sequence charts overview the procedures of SUA. These are meant as examples, they do not, in and of themselves, impose additional requirements upon an instance of SUA.5.1. SG Architecture
The sequences below outline logical steps for a variety of scenarios within a SG architecture. Please note that these scenarios cover a Primary/Backup configuration. Where there is a load-sharing configuration then the SGP can declare availability when 1 ASP issues ASPAC but can only declare unavailability when all ASPs have issued ASPIA.
5.1.1. Establishment of SUA connectivity
The following is established before traffic can flow. Each node is configured (via MIB, for example) with the connections that need to be setup. ASP-a1 ASP-a2 SG SEP (Primary) (Backup) |------Establish SCTP Association------| |--Estab. SCTP Ass--| |--Align SS7 link---| +----------------ASP Up----------------> <--------------ASP Up Ack--------------+ +------ASP Up-------> <---ASP Up Ack------+ +-------------ASP Active---------------> <----------ASP Active Ack--------------+ <----------NTFY (ASP Active)-----------+ <-NTFY (ASP Active)-+ +--------SSA--------> <--------SSA--------+ <-----------------DAVA-----------------+ +-----------------CLDT-----------------> +--------UDT-------->5.1.2. Fail-over scenarios
The following sequences address fail-over of SEP and ASP.5.1.2.1. SEP Fail-over
The SEP knows that the SGP is 'concerned' about its availability. Similarly, the SGP knows that ASP-a1 is concerned about the SEPs availability. ASP-a1 ASP-a2 SG SEP (Primary) (Backup) <--------SSP--------+ <-----------------DUNA-----------------+ +-----------------DAUD-----------------> +--------SST-------->
5.1.2.2. Successful ASP Fail-over scenario
The following is an example of a successful fail-over scenario, where there is a fail-over from ASP-a1 to ASP-a2, i.e., Primary to Backup. During the fail-over, the SGP buffers any incoming data messages from the SEP, forwarding them when the Backup becomes available. ASP-a1 ASP-a2 SG SEP (Primary) (Backup) +-------------ASP Inactive-------------> <-----------ASP Inactive ACK-----------+ <--------------------NTFY (AS Pending)-+ <-NTFY (AS Pending)-+ +----ASP Active-----> <--ASP Active Ack---+ <-NTFY (AS Active)--+ <----------NTFY (AS Active)------------+5.1.2.3. Unsuccessful ASP Fail-over scenario
ASP-a1 ASP-a2 SG SEP (Primary) (Backup) +-------------ASP Inactive-------------> <-----------ASP Inactive ACK-----------+ <--------------------NTFY (AS Pending)-+ <--NTFY (AS Pending)-+ After some time elapses (i.e., timeout). +--------SSP--------> <--------SST--------+ <-------------------NTFY (AS Inactive)-+ <-NTFY (AS Inactive)-+5.2. IPSP Examples
The sequences below outline logical steps for a variety of scenarios within an IP-IP architecture. Please note that these scenarios cover a Primary/Backup configuration. Where there is a load-sharing configuration then the AS can declare availability when 1 ASP issues ASPAC but can only declare unavailability when all ASPs have issued ASPIA.5.2.1. Establishment of SUA connectivity
The following shows an example establishment of SUA connectivity. In this example, each IPSP consists of an Application Server and two ASPs. The following is established before SUA traffic can flow. A connectionless flow is shown for simplicity.
Establish SCTP Connectivity - as per RFC 2960. Note that SCTP connections are bidirectional. The endpoint that establishes SCTP connectivity MUST also establish UA connectivity (see RFC 2960, section 5.2.1 for handling collisions) [2960]. IP SEP A IP SEP B AS A AS B ASP-a1 ASP-a2 ASP-b2 ASP-b1 [All ASPs are in the ASP-DOWN state] +-------------------------------ASP Up--------------------------> <-----------------------------ASP Up Ack------------------------+ +--------------ASP Up---------------> <------------ASP Up Ack-------------+ +---------------------------ACTIVE-------------------------------> <-------------------------ACTIVE Ack-----------------------------+ [Traffic can now flow directly between ASPs] +-----------------------------CLDT------------------------------->5.2.2. Fail-over scenarios
The following sequences address fail-over of ASP.5.2.2.1. Successful ASP Fail-over scenario
The following is an example of a successful fail-over scenario, where there is a fail-over from ASP-a1 to ASP-a2, i.e., Primary to Backup. Since data transfer passes directly between peer ASPs, ASP-b1 is notified of the fail-over of ASP-a1 and buffers outgoing data messages until ASP-a2 becomes available. IP SEP A IP SEP B ASP-a1 ASP-a2 ASP-b2 ASP-b1 +-----------------------------ASP Inact------------------------> <---------------------------ASP Inact Ack----------------------+ <---------------NTFY (ASP-a1 Inactive)--------------+ +---------------------ASP Act-----------------------> <-------------------ASP Act Ack---------------------+
5.2.2.2. Unsuccessful ASP Fail-over scenario
The sequence is the same as 5.2.2.1 except that, since the backup fails to come in then, the Notify messages declaring the availability of the backup are not sent.6. Security Considerations
The security considerations discussed for the 'Security Considerations for SIGTRAN Protocols' [3788] document apply to this document.7. IANA Considerations
7.1. SCTP Payload Protocol ID
IANA has assigned a SUA value for the Payload Protocol Identifier in the SCTP DATA chunk. The following SCTP Payload Protocol Identifier is registered: SUA "4" The SCTP Payload Protocol Identifier value "4" SHOULD be included in each SCTP DATA chunk, to indicate that the SCTP is carrying the SUA protocol. The value "0" (unspecified) is also allowed but any other values MUST not be used. This Payload Protocol Identifier is not directly used by SCTP but MAY be used by certain network entities to identify the type of information being carried in a DATA chunk. The User Adaptation peer MAY use the Payload Protocol Identifier, as a way of determining additional information about the data being presented to it by SCTP.7.2. Port Number
IANA has registered SCTP Port Number 14001 for SUA. It is recommended that SGPs use this SCTP port number for listening for new connections. SGPs MAY also use statically configured SCTP port numbers instead.7.3. Protocol Extensions
This protocol may also be extended through IANA in three ways: - Through definition of additional message classes. - Through definition of additional message types. - Through definition of additional message parameters.
The definition and use of new message classes, types and parameters is an integral part of SIGTRAN adaptation layers. Thus, these extensions are assigned by IANA through an IETF Consensus action as defined in [2434]. The proposed extension MUST in no way adversely affect the general working of the protocol. A new registry has been created by IANA to allow the protocol to be extended.7.3.1. IETF Defined Message Classes
The documentation for a new message class MUST include the following information: (a) A long and short name for the message class; (b) A detailed description of the purpose of the message class.7.3.2. IETF Defined Message Types
Documentation of the message type MUST contain the following information: (a) A long and short name for the new message type; (b) A detailed description of the structure of the message. (c) A detailed definition and description of intended use of each field within the message. (d) A detailed procedural description of the use of the new message type within the operation of the protocol. (e) A detailed description of error conditions when receiving this message type. When an implementation receives a message type which it does not support, it MUST respond with an Error (ERR) message, with an Error Code = Unsupported Message Type.7.3.4. IETF-defined TLV Parameter Extension
Documentation of the message parameter MUST contain the following information: (a) Name of the parameter type. (b) Detailed description of the structure of the parameter field. This structure MUST conform to the general type-length-value format described earlier in the document. (c) Detailed definition of each component of the parameter value.
(d) Detailed description of the intended use of this parameter type,
and an indication of whether and under what circumstances
multiple instances of this parameter type may be found within the
same message type.
8. Timer Values
Ta 2 seconds
Tr 2 seconds
T(ack) 2 seconds
T(ias) Inactivity Send timer 7 minutes
T(iar) Inactivity Receive timer 15 minutes
T(beat) Heartbeat Timer 30 seconds
9. Acknowledgements
The authors would like to thank (in alphabetical order) Richard
Adams, Javier Pastor-Balbas, Andrew Booth, Martin Booyens, F.
Escobar, S. Furniss Klaus Gradischnig, Miguel A. Garcia, Marja-Liisa
Hamalainen, Sherry Karl, S. Lorusso, Markus Maanoja, Sandeep Mahajan,
Ken Morneault, Guy Mousseau, Chirayu Patel, Michael Purcell, W.
Sully, Michael Tuexen, Al Varney, Tim Vetter, Antonio Villena, Ben
Wilson, Michael Wright and James Yu for their insightful comments and
suggestions.
(page 123 continued on part 5)