RFC 5352
Aggregate Server Access Protocol (ASAP)
Pages: 53
Experimental
Experimental
Part 2 of 3 – Pages 18 to 42
3. Procedures
This section will focus on the methods and procedures used by an internal ASAP Endpoint. Appropriate timers and recovery actions for failure detection and management are also discussed. Also, please note that ASAP messages sent between a PE and PU are identified by an SCTP Payload Protocol Identifier (PPID).3.1. Registration
When a PE wishes to initiate or join a server pool, it MUST use the procedures outlined in this section for registration. Often, the registration will be triggered by a user request primitive (discussed in Section 6.1). The PE MUST register using an SCTP association established between itself and the Home ENRP server. If the PE has not established its Home ENRP server, it MUST follow the procedures specified in Section 3.6. Once the PE's ASAP Endpoint has established its Home ENRP server, the following procedures MUST be followed to register: R1) The PE's SCTP endpoint used to communicate with the Home ENRP server MUST be bound to all IP addresses that will be used by the PE (regardless of which transport protocol will be used to service user requests to the PE). R2) The PE's ASAP Endpoint MUST formulate an ASAP_REGISTRATION message, as defined in Section 2.2.1. In formulating the message, the PE MUST: R2.1) Fill in the Pool Handle parameter to specify which server pool the ASAP Endpoint wishes to join. R2.2) Fill in the PE identifier using a good-quality randomly generated number ([RFC4086] provides some information on randomness guidelines).
R2.3) Fill in the Registration Lifetime parameter with the number
of seconds that this registration is valid for. Note that a PE
that wishes to continue service MUST re-register before the
registration expires.
R2.4) Fill in a User Transport parameter to specify the type of
transport and the data/control channel usage the PE is willing
to support. Note, in joining an existing server pool, the PE
MUST follow the overall transport type and overall data/control
channel usage of the pool. Otherwise, the registration may be
rejected by the ENRP server.
R2.5) Fill in the preferred Pool Member Selection Policy
parameter.
R3) Send the ASAP_REGISTRATION message to the Home ENRP server using
SCTP.
R4) Start a T2-registration timer.
Note: the PE does not need to fill in the optional ASAP transport
parameter. The ASAP transport parameter will be filled in and used
by the Home ENRP server.
If the T2-registration timer expires before receiving an
ASAP_REGISTRATION_RESPONSE message, or a SEND.FAILURE notification is
received from the SCTP layer, the PE shall start the Server Hunt
procedure (see Section 3.6) in an attempt to get service from a
different ENRP server. After establishing a new Home ENRP server,
the PE SHOULD restart the registration procedure.
At the reception of the registration response, the PE MUST stop the
T2-registration timer. If the response indicates success, the PE is
registered and will be considered an available member of the server
pool. If the registration response indicates a failure, the PE must
either re-attempt registration after correcting the error or return a
failure indication to the PE's upper layer. The PE MUST NOT re-
attempt registration without correcting the error condition.
At any time, a registered PE MAY wish to re-register to either update
its member selection Policy Value or registration expiration time.
When re-registering, the PE MUST use the same PE identifier.
After successful registration, the PE MUST start a T4-reregistration
timer. At its expiration, a re-registration SHOULD be made starting
at step R1, including (at completion) restarting the T4-
reregistration timer.
Note that an implementation SHOULD keep a record of the number of
registration (and re-registration) attempts it makes in a local
variable that gets set to zero before the initial registration
attempt to the Home ENRP server or after a successful re-
registration. If repeated registration timeouts or failures occurs
and the local count exceeds the Threshold 'MAX-REG-ATTEMPT', the
implementation SHOULD report the error to its upper layer and stop
attempting registration.
The ENRP server handles the ASAP_REGISTRATION message according to
the following rules:
1. If the named pool does not exist in the handlespace, the ENRP
server MUST create a new pool with that handle in the handlespace
and add the PE to the pool as its first PE.
When a new pool is created, the overall member selection policy
of the pool MUST be set to the policy type indicated by the first
PE, the overall pool transport type MUST be set to the transport
type indicated by the PE, and the overall pool data/control
channel configuration MUST be set to what is indicated in the
Transport Use field of the User Transport parameter by the
registering PE.
2. If the named pool already exists in the handlespace, but the
requesting PE is not currently a member of the pool, the ENRP
server will add the PE as a new member to the pool.
However, before adding the PE to the pool, the server MUST check
if the policy type, transport type, and transport usage indicated
by the registering PE is consistent with those of the pool. If
different, the ENRP server MUST reject the registration.
3. If the named pool already exists in the handlespace *and* the
requesting PE is already a member of the pool, the ENRP server
SHOULD consider this as a re-registration case. The ENRP server
MUST perform the same tests on policy, transport type, and
transport use, as described above. If the re-registration is
accepted after the test, the ENRP server SHOULD replace the
attributes of the existing PE with the information carried in the
received ASAP_REGISTRATION message.
4. After accepting the registration, the ENRP server MUST assign
itself the owner of this PE. If this is a re-registration, the
ENRP server MUST take over ownership of this PE, regardless of
whether the PE was previously owned by this server or by another
server. The ENRP server MUST also record the SCTP transport
address from which it received the ASAP_REGISTRATION in the ASAP
Transport parameter TLV inside the PE parameter of this PE.
5. The ENRP server may reject the registration due to other reasons
such as invalid values, lack of resource, authentication failure,
etc.
In all above cases, the ENRP server MUST reply to the requesting PE
with an ASAP_REGISTRATION_RESPONSE message. If the registration is
accepted, the ENRP server MUST set the R flag in the
ASAP_REGISTRATION_RESPONSE to '0'. If the registration is rejected,
the ENRP server MUST indicate the rejection by setting the R flag in
the ASAP_REGISTRATION_RESPONSE to '1'.
If the registration is rejected, the ENRP server SHOULD include the
proper error cause(s) in the ASAP_REGISTRATION_RESPONSE message.
If the registration is granted (either a new registration or a re-
registration case), the ENRP server MUST assign itself to be the Home
ENRP server of the PE, i.e., to "own" the PE.
Implementation note: For better performance, the ENRP server may
find it both efficient and convenient to internally maintain two
separate PE lists or tables -- one is for the PEs that are owned
by the ENRP server and the other is for all the PEs owned by their
peer(s).
Moreover, if the registration is granted, the ENRP server MUST take
the handlespace update action to inform its peers about the change
just made. If the registration is denied, no message will be sent to
its peers.
3.2. De-Registration
In the event a PE wishes to de-register from its server pool
(normally, via an upper-layer request, see Section 6.2), it SHOULD
use the following procedure. It should be noted that an alternate
method of de-registration is to NOT re-register and to allow the
registration life of the PE to expire. In this case, an
ASAP_DEREGISTRATION_RESPONSE message is sent to the PE's ASAP
Endpoint to indicate the removal of the PE from the pool it
registered.
When de-registering, the PE SHOULD use the SCTP association that was
used for registration with its Home ENRP server. To de-register, the
PE's ASAP Endpoint MUST take the following actions:
D1) Fill in the Pool Handle parameter of the ASAP_DEREGISTRATION
message (Section 2.2.2) using the same Pool Handle parameter sent
during registration.
D2) Fill in the PE Identifier parameter of the ASAP_DEREGISTRATION
message. The identifier MUST be the same as used during
registration. The use of the same Pool Handle and Pool Identifier
parameters used in registration allows the identity of the PE ASAP
Endpoint to be verified before de-registration can occur.
D3) Send the ASAP_DEREGISTRATION message to the Home ENRP server
using the PE's SCTP association.
D4) Start a T3-Deregistration timer.
If the T3-Deregistration timer expires before receiving either an
ASAP_REGISTRATION_RESPONSE message, or a SEND.FAILURE notification
from the PE's SCTP endpoint, the PE's ASAP Endpoint shall start the
ENRP Server Hunt procedure (see Section 3.6) in an attempt to get
service from another ENRP server. After establishing a new Home ENRP
server, the ASAP Endpoint SHOULD restart the de-registration
procedure.
At the reception of the ASAP_DEREGISTRATION_RESPONSE, the PE's ASAP
endpoint MUST stop the T3-Deregistration timer.
It should be noted that after a successful de-registration, the PE
MAY still receive requests for some period of time. The PE MAY wish
to remain active and service these requests or to exit and ignore
these requests.
Upon receiving the message, the ENRP server SHALL remove the PE from
its handlespace. Moreover, if the PE is the last one of the named
pool, the ENRP server will remove the pool from the handlespace as
well.
If the ENRP server fails to find any record of the PE in its
handlespace, it SHOULD consider the de-registration granted and
completed, and send an ASAP_DEREGISTRATION_RESPONSE message to the
PE.
The ENRP server may reject the de-registration request for various
reasons, such as invalid parameters, authentication failure, etc.
In response, the ENRP server MUST send an
ASAP_DEREGISTRATION_RESPONSE message to the PE. If the de-
registration is rejected, the ENRP server MUST indicate the rejection
by including the proper Operational Error parameter.
It should be noted that de-registration does not stop the PE from sending or receiving application messages. Once the de-registration request is granted *and* the PE removed from its local copy of the handlespace, the ENRP server MUST take the handlespace update action to inform its peers about the change just made. Otherwise, the ENRP server MUST NOT inform its peers.3.3. Handle Resolution
At any time, a PE or PU may wish to resolve a handle. This usually will occur when an ASAP Endpoint sends a Pool Handle (Section 6.5.1) to its Home ENRP server or requests a cache population (Section 6.3). It may also occur for other reasons (e.g., the internal ASAP PE wishes to know its peers to send a message to all of them). When an ASAP Endpoint (PE or PU) wishes to resolve a pool handle to a list of accessible transport addresses of the member PEs of the pool, it MUST take the following actions: NR1) Fill in an ASAP_HANDLE_RESOLUTION message (Section 2.2.5) with the Pool Handle to be resolved. NR2) If the endpoint does not have a Home ENRP server, start the ENRP Server Hunt procedures specified in Section 3.6 to obtain one. Otherwise, proceed to step NR3. NR3) If a PE, send the ASAP_HANDLE_RESOLUTION message to the Home ENRP server using SCTP; if a PU, send the ASAP_HANDLE_RESOLUTION message to the Home ENRP server using either TCP or SCTP. If sent from a PE, the SCTP association used for registration SHOULD be used. NR4) Start a T1-ENRPrequest timer. If the T1-ENRPrequest timer expires before receiving a response message, the ASAP Endpoint SHOULD take the steps described in Section 3.7.2. If a SEND.FAILURE notification is received from the SCTP or TCP layer, the ASAP Endpoint SHOULD start the Server Hunt procedure (see Section 3.6) in an attempt to get service from a different ENRP server. After establishing a new Home ENRP server, the ASAP Endpoint SHOULD restart the handle resolution procedure. At the reception of the ASAP_HANDLE_RESOLUTION_RESPONSE message, the ASAP Endpoint MUST stop its T1-ENRPrequest timer. After stopping the T1-ENRPrequest timer, the ASAP Endpoint SHOULD process the message as appropriate (e.g., populate a local cache, give the response to the ASAP User, and/or use the response to send the ASAP User's message).
Note that some ASAP Endpoints MAY use a cache to minimize the number
of handle resolutions sent. If a cache is used, it SHOULD:
C1) Be consulted before sending a handle resolution.
C2) Have a stale timeout timer associated with each cache entry. If
the cache entry is determined to be stale upon a cache hit, a
handle resolution message SHOULD be sent so the cache can be
updated.
C3) In the case of a stale cache entry, the implementation may, in
parallel, update the cache and answer the request, or it may block
the user and wait for an updated cache before proceeding with the
users request.
C4) If the cache entry is NOT stale, the endpoint SHOULD NOT send a
handle resolution request but instead SHOULD use the entry from
the cache.
It should be noted that the impact of using a cache depends on the
policy and the requirements of the application. For some
applications, cache-usage can increase the performance of the system;
for some, it can decrease it.
An ENRP server SHOULD be prepared to receive ASAP_HANDLE_RESOLUTION
requests from PUs, either over an SCTP association on the well-known
SCTP port, or over a TCP connection on the well-known TCP port.
Upon reception of the ASAP_HANDLE_RESOLUTION message, the ENRP server
MUST first look up the pool handle in its handlespace. If the pool
exists, the Home ENRP server MUST compose and send back an
ASAP_HANDLE_RESOLUTION_RESPONSE message to the requesting PU.
In the response message, the ENRP server SHOULD list all the PEs
currently registered in this pool, in a list of PE parameters. The
ENRP server MUST also include a pool member selection policy
parameter to indicate the overall member selection policy for the
pool, if the current pool member selection policy is not Round-Robin.
If the named pool does not exist in the handlespace, the ENRP server
MUST reject the handle resolution request by responding with an
ASAP_HANDLE_RESOLUTION_RESPONSE message carrying an Unknown Pool
Handle error.
3.4. Endpoint Keep Alive
The ASAP_ENDPOINT_KEEP_ALIVE message is sent by an ENRP server to a PE in order to verify it is reachable. If the transport level heartbeat mechanism is insufficient, this message can be used in a heartbeat mechanism for the ASAP level whose goal is determining the health status of the ASAP level in a timely fashion. (The transport level heartbeat mechanism may be insufficient due to either the timeouts or the heartbeat interval being set too long, or, that the transport level heartbeat mechanism's coverage is limited only to the transport level at the two ends.) Additionally, the ASAP_ENDPOINT_KEEP_ALIVE message has value in the reliability of fault detection if the SCTP stack is in the kernel. In such a case, while the SCTP-level heartbeat monitors the end-to-end connectivity between the two SCTP stacks, the ASAP-level heartbeat monitors the end-to-end liveliness of the ASAP layer above it. The use of the ASAP_ENDPOINT_KEEP_ALIVE message (Section 2.2.7) and the ASAP_ENDPOINT_KEEP_ALIVE_ACK (Section 2.2.8) is described below. Upon reception of an ASAP_ENDPOINT_KEEP_ALIVE message, the following actions MUST be taken: KA1) The PE must verify that the Pool Handle is correct and matches the Pool Handle sent in its earlier ASAP_REGISTRATION message. If the Pool Handle does not match, the PE MUST silently discard the message. KA2) Send an ASAP_ENDPOINT_KEEP_ALIVE_ACK (Section 2.2.8) as follows: KA2.1) Fill in the Pool Handle parameter with the PE's Pool Handle. KA2.2) Fill in the PE Identifier parameter using the PE identifier used by this PE for registration. KA2.3) Send the ASAP_ENDPOINT_KEEP_ALIVE_ACK message via the appropriate SCTP association for the ENRP server that sent the ASAP_ENDPOINT_KEEP_ALIVE message. KA2.4) If the H flag in the received ASAP_ENDPOINT_KEEP_ALIVE message is set, and the Server Identifier in the message is NOT the identity of your Home ENRP server (or it is not set, e.g., you have a no Home ENRP server) adopt the sender of the ASAP_ENDPOINT_KEEP_ALIVE message as the new Home ENRP server.
3.5. Unreachable Endpoints
Occasionally, an ASAP Endpoint may realize a PE is unreachable. This may occur by a specific SCTP error realized by the ASAP endpoint or via an ASAP User report via the Transport.Failure Primitive (Section 6.9.2). In either case, the ASAP Endpoint SHOULD report the unavailability of the PE by sending an ASAP_ENDPOINT_UNREACHABLE message to any ENRP server. Before sending the ASAP_ENDPOINT_UNREACHABLE message, the ASAP Endpoint should fill in the Pool Handle parameter and PE Identifier parameter of the unreachable endpoint. If the sender is a PE, the message MUST be sent via SCTP. It should be noted that an ASAP Endpoint MUST report no more than once each time it encounters such an event. Additionally, when processing a Transport.Failure Primitive (Section 6.9.2), the ASAP Endpoint MUST NOT send an ASAP_ENDPOINT_UNREACHABLE message unless the user has made a previous request to send data to the PE specified by the primitive. Upon the reception of an ASAP_ENDPOINT_UNREACHABLE message, an ENRP server MUST immediately send a point-to-point ASAP_ENDPOINT_KEEP_ALIVE message to the PE in question (the H flag in the message SHOULD be set to '0', in this case). If this ASAP_ENDPOINT_KEEP_ALIVE fails (e.g., it results in an SCTP SEND.FAILURE notification), the ENRP server MUST consider the PE as truly unreachable and MUST remove the PE from its handlespace. If the ASAP_ENDPOINT_KEEP_ALIVE message is transmitted successfully to the PE, the ENRP server MUST retain the PE in its handlespace. Moreover, the server SHOULD keep a counter to record how many ASAP_ENDPOINT_UNREACHABLE messages it has received reporting reachability problem relating to this PE. If the counter exceeds the protocol threshold MAX-BAD-PE-REPORT, the ENRP server SHOULD remove the PE from its handlespace. Optionally, an ENRP server may also periodically send point-to-point ASAP_ENDPOINT_KEEP_ALIVE (with the H flag set to '0') messages to each of the PEs owned by the ENRP server in order to check their reachability status. If the sending of ASAP_ENDPOINT_KEEP_ALIVE to a PE fails, the ENRP server MUST consider the PE as unreachable and MUST remove the PE from its handlespace. Note, if an ENRP server owns a large number of PEs, the implementation should pay attention not to flood the network with bursts of ASAP_ENDPOINT_KEEP_ALIVE messages. Instead, the implementation MUST distribute the ASAP_ENDPOINT_KEEP_ALIVE message traffic over a time period. This can be achieved by varying the time between two ASAP_ENDPOINT_KEEP_ALIVE messages to the same PE randomly by plus/ minus 50 percent.
3.6. ENRP Server Hunt Procedures
Each PU and PE manages a list of transport addresses of ENRP servers it knows about. If multicast capabilities are used within the operational scope, an ENRP server MUST send periodically every (N+1)*T6-Serverannounce an ASAP_SERVER_ANNOUNCE message (Section 2.2.10), which includes all the transport addresses available for ASAP communication on the multicast ENRP client channel, where N is the number of ENRP servers the server has found via receiving ASAP_SERVER_ANNOUNCE messages. This should result in a message rate of approximately 1 ASAP_SERVER_ANNOUNCE per T6-Serverannounce. If an ASAP_SERVER_ANNOUNCE message is received by a PU or PE, it SHOULD insert all new included transport addresses into its list of ENRP server addresses and start a T7-ENRPoutdate timer for each address. For all already-known, included transport addresses, the T7-ENRPoutdate timer MUST be restarted for each address. If no transport parameters are included in the ASAP_SERVER_ANNOUNCE message, the SCTP transport protocol is assumed to be used and the source IP address and the IANA-registered ASAP port number is used for communication with the ENRP server. If a T7-ENRPoutdate timer for a transport address expires, the corresponding address is deleted from the managed list of transport addresses of the PU or PE. If multicast capabilities are not used within the operational scope, each PU and PE MUST have a configured list of transport addresses of ENRP servers. At its startup, or when it fails to communicate with its Home ENRP server (i.e., timed out on an ENRP request), a PE or PU MUST establish a new Home ENRP server (i.e., set up a TCP connection or SCTP association with a different ENRP server). To establish a Home ENRP server, the following rules MUST be followed: SH1) The PE or PU SHOULD try to establish an association or connection, with no more than three ENRP servers. An ASAP Endpoint MUST NOT establish more than three associations or connections. SH2) The ASAP Endpoint shall start a T5-Serverhunt timer.
SH3) If the ASAP Endpoint establishes an association or connection
it MUST stop its T5-Serverhunt timer. The ASAP Endpoint SHOULD
also reset the T5-Serverhunt timer to its initial value and then
proceed to step SH6.
SH4) If an association or connection establishment fails, the ASAP
Endpoint SHOULD try to establish an association or connection
using a different transport address.
SH5) If the T5-Serverhunt timer expires, the following should be
performed:
SH5.1) The ASAP Endpoint MUST double the value of the T5-
Serverhunt timer. Note that this doubling is capped at the
value RETRAN.max.
SH5.2) The ASAP Endpoint SHOULD stop the establishment of
associations and connections with the transport addresses
selected in step SH1.
SH5.2) The ASAP Endpoint SHOULD repeat trying to establish an
association or connection by proceeding to step SH1. It SHOULD
attempt to select a different set of transport addresses with
which to connect.
SH6) The PE or PU shall pick one of the ENRP servers with which it
was able to establish an association or connection, and send all
subsequent ENRP request messages to this new Home ENRP server.
3.7. Handling ASAP Endpoint to ENRP Server Communication Failures
Three types of failure may occur when the ASAP Endpoint at either the
PE or PU tries to communicate with an ENRP server:
A) SCTP send failure
B) T1-ENRPrequest timer expiration
C) Registration failure
3.7.1. SCTP Send Failure
This communication failure indicates that the SCTP layer was unable
to deliver a message sent to an ENRP server. In other words, the
ENRP server is unreachable.
In such a case, the ASAP Endpoint MUST NOT re-send the undeliverable message. Instead, it SHOULD discard the message and start the ENRP Server Hunt procedure as described in Section 3.6. After finding a new Home ENRP server, the ASAP Endpoint should re-send the request. Note that an ASAP Endpoint MAY also choose to NOT discard the message, but to queue it for retransmission after a new Home ENRP server is found. If an ASAP Endpoint does choose to discard the message, after a new Home ENRP server is found, the ASAP Endpoint MUST be capable of reconstructing the original request.3.7.2. T1-ENRPrequest Timer Expiration
When the T1-ENRPrequest timer expires, the ASAP Endpoint should re- send the original request to the ENRP server and restart the T1- ENRPrequest timer. In parallel, the ASAP Endpoint should begin the ENRP server hunt procedures described in Section 3.6. This should be repeated up to MAX-REQUEST-RETRANSMIT times. After that, an Error.Report notification should be generated to inform the ASAP User, and the ENRP request message associated with the T1- ENRPrequest timer should be discarded. It should be noted that if an alternate ENRP server responds, the ASAP Endpoint SHOULD adopt the responding ENRP server as its new Home ENRP server and re-send the request to the new Home ENRP server.3.7.3. Registration Failure
Registration failure is discussed in Section 3.1.3.8. Cookie Handling Procedures
Whenever a PE wants, and a control channel exists, it can send an ASAP_COOKIE message to a PU via the control channel. The PU's ASAP endpoint stores the Cookie parameter and discards an older cookie if it is previously stored. Note: A control channel is a communication channel between a PU and PE that does not carry data passed to the user. This is accomplished with SCTP by using a PPID to separate the ASAP messages (Cookie and Business Card) from normal data messages. If the PU's ASAP Endpoint detects a failure and initiates a failover to a different PE, it SHOULD send the latest received cookie parameter in an ASAP_COOKIE_ECHO message to the new PE as the first message on the control channel. Upper layers may be involved in the failover procedure.
The cookie handling procedure can be used for state sharing. Therefore, a cookie should be signed by the sending PE ASAP Endpoint and the cookie should be verified by the receiving PE's ASAP Endpoint. The details of the verification procedure are out of scope for this document. It is only important that the PU always stores the last received Cookie parameter and sends that back unmodified in case of a PE failure.3.9. Business Card Handling Procedures
When communication begins between a PU and a PE, either of which could be part of a PU/PE combination (i.e., a message is sent between the entities), a PE should always send an ASAP_BUSINESS_CARD message to a PU. A PU should send an ASAP_BUSINESS_CARD message to a PE only if it is part of a PU/PE combination. An ASAP_BUSINESS_CARD message MUST ONLY be sent if a control channel exists between a PU and PE. After communication has been established between a PE and PU, a new ASAP_BUSINESS_CARD message may be sent at any time by either entity to update its failover order. The ASAP_BUSINESS_CARD message serves two purposes. First, it lists the pool handle. For a PU that is part of a PU/PE combination that is contacting a PE, this is essential so that the PE learns the pool handle of the PU/PE combination requesting service. Secondly, the ASAP_BUSINESS_CARD message tells the receiving entity a failover order that is recommended to follow. This should facilitate rendezvous between entities that have been working together, as well as to control the load redistribution upon the failure of any PE. Upon receipt of an ASAP_BUSINESS_CARD message (see Section 2.2.13), the receiving ASAP Endpoint SHOULD: BC1) Unpack the message, and if no entry exists in the translation cache of the receiving ASAP Endpoint for the pool handle listed within the ASAP_BUSINESS_CARD message, perform an ASAP_HANDLE_RESOLUTION for that pool handle. If the translation cache does hold an entry for the pool handle, then it may be necessary to update the peer endpoint. BC2) Unpack the message and populate a preferred list for failover order. If the peer's PE should fail, this preferred list will be used to guide the ASAP Endpoint in the selection of an alternate PE.
4. Roles of Endpoints
A PU MUST implement the handling of ASAP_HANDLE_RESOLUTION and ASAP_HANDLE_RESOLUTION_RESPONSE messages. Furthermore, it MUST support the handling of ASAP_ERROR messages. It MAY implement the handling of ASAP_COOKIE, ASAP_COOKIE_ECHO, and ASAP_BUSINESS_CARD messages. It MAY also implement the handling of ASAP_SERVER_ANNOUNCE messages. A PE MUST implement the handling of ASAP_REGISTRATION, ASAP_DEREGISTRATION, ASAP_REGISTRATION_RESPONSE, and ASAP_DEREGISTRATION_RESPONSE messages. Furthermore, it MUST support the handling of ASAP_ENDPOINT_KEEP_ALIVE, ASAP_ENDPOINT_KEEP_ALIVE_ACK, ASAP_ENDPOINT_UNREACHABLE, and ASAP_ERROR messages. It SHOULD support the handling of ASAP_COOKIE, ASAP_COOKIE_ECHO, and ASAP_BUSINESS_CARD messages. Furthermore, it MAY support the handling of ASAP_SERVER_ANNOUNCE messages. An ENRP server MUST implement the handling of ASAP_REGISTRATION, ASAP_DEREGISTRATION, ASAP_REGISTRATION_RESPONSE, and ASAP_DEREGISTRATION_RESPONSE messages. Furthermore, it MUST support the handling of ASAP_ENDPOINT_KEEP_ALIVE, ASAP_ENDPOINT_KEEP_ALIVE_ACK, ASAP_ENDPOINT_UNREACHABLE, and ASAP_ERROR messages. Furthermore, it MAY support the handling of ASAP_SERVER_ANNOUNCE messages. If a node acts as a PU and a PE, it MUST fulfill both roles.5. SCTP Considerations
Each ASAP message is considered as an SCTP user message. The PPID registered for ASAP SHOULD be used. The SCTP port used at the ENRP server might be preconfigured or announced in the ASAP_SERVER_ANNOUNCE message or the well-known ASAP port. ASAP messages belonging to the control channel MUST be sent using the PPID registered for ASAP. Messages belonging to the data channel MUST NOT use the PPID registered for ASAP.6. The ASAP Interfaces
This chapter will focus primarily on the primitives and notifications that form the interface between the ASAP User and ASAP and that between ASAP and its lower-layer transport protocol (e.g., SCTP).
Note, the following primitive and notification descriptions are shown for illustrative purposes. We believe that including these descriptions in this document is important to the understanding of the operation of many aspects of ASAP; but an ASAP implementation is not required to use the exact syntax described in this section. An ASAP User passes primitives to the ASAP sub-layer to request certain actions. Upon the completion of those actions or upon the detection of certain events, the ASAP layer will notify the ASAP User.6.1. Registration.Request Primitive
Format: registration.request(Pool Handle, User Transport parameter(s)) The Pool Handle parameter contains a NULL terminated ASCII string of fixed length. The optional User Transport parameter(s) indicates specific transport parameters and types with which to register. If this optional parameter is left off, then the SCTP endpoint used to communicate with the ENRP server is used as the default User Transport parameter. Note that any IP address contained within a User Transport parameter MUST be a bound IP address in the SCTP endpoint used to communicate with the ENRP server. The ASAP User invokes this primitive to add itself to the handlespace, thus becoming a Pool Element of a pool. The ASAP User must register itself with the ENRP server by using this primitive before other ASAP Users using the handlespace can send message(s) to this ASAP User by Pool Handle or by PE handle (see Sections 6.5.1 and 6.5.3). In response to the registration primitive, the ASAP Endpoint will send an ASAP_REGISTRATION message to the Home ENRP server (see Sections 2.2.1 and 3.1), and start a T2-registration timer.6.2. Deregistration.Request Primitive
Format: deregistration.request(Pool Handle) The ASAP PE invokes this primitive to remove itself from the Server Pool. This should be used as a part of the graceful shutdown process by the application. An ASAP_DEREGISTRATION message will be sent by the ASAP Endpoint to the Home ENRP server (see Sections 2.2.2 and 3.2).
6.3. CachePopulateRequest Primitive
Format: cache_populate_request([Pool-Handle | Pool-Element-Handle]) If the address type is a Pool Handle and a local handle translation cache exists, the ASAP Endpoint should initiate a mapping information query by sending an ASAP_HANDLE_RESOLUTION message on the Pool handle and updating its local cache when the response comes back from the ENRP server. If a Pool-Element-Handle is passed, then the Pool Handle is unpacked from the Pool-Element-Handle and the ASAP_HANDLE_RESOLUTION message is sent to the ENRP server for resolution. When the response message returns from the ENRP server, the local cache is updated. Note that if the ASAP service does NOT support a local cache, this primitive performs NO action.6.4. CachePurgeRequest Primitive
Format: cache_purge_request([Pool-Handle | Pool-Element-Handle]) If the user passes a Pool Handle and local handle translation cache exists, the ASAP Endpoint should remove the mapping information on the Pool Handle from its local cache. If the user passes a Pool- Element-Handle, then the Pool Handle within is used for the cache_purge_request. Note that if the ASAP service does NOT support a local cache, this primitive performs NO action.6.5. DataSendRequest Primitive
Format: data_send_request(destinationAddress, typeOfAddress, message, sizeOfMessage, Options); This primitive requests ASAP to send a message to some specified Pool or Pool Element within the current Operational scope. Depending on the address type used for the send request, the sender's ASAP Endpoint may perform address translation and Pool Element selection before sending the message out. This MAY also dictate the creation of a local transport endpoint in order to meet the required transport type. The data_send_request primitive can take different forms of address types, as described in the following sections.
6.5.1. Sending to a Pool Handle
In this case, the destinationAddress and typeOfAddress together indicate a pool handle. This is the simplest form of send_data_request primitive. By default, this directs ASAP to send the message to one of the Pool Elements in the specified pool. Before sending the message out to the pool, the sender's ASAP endpoint MUST first perform a pool handle to address translation. It may also need to perform Pool Element selection if multiple Pool Elements exist in the pool. If the sender's ASAP implementation does not support a local cache of the mapping information, or if it does not have the mapping information on the pool in its local cache, it will transmit an ASAP_HANDLE_RESOLUTION message (see Sections 2.2.5 and 3.3) to the current Home ENRP server and MUST hold the outbound message in queue while awaiting the response from the ENRP server (any further send request to this pool before the ENRP server responds SHOULD also be queued). Once the necessary mapping information arrives from the ENRP server, the sender's ASAP will: A) map the pool handle into a list of transport addresses of the destination PE(s); B) if multiple PEs exist in the pool, choose one of them and transmit the message to it. In that case, the choice of the PE is made by the ASAP Endpoint of the sender based on the server pooling policy, as discussed in Section 6.5.2; C) optionally create any transport endpoint that may be needed to communicate with the PE selected; D) if no transport association or connection exists towards the destination PE, establish any needed transport state; E) send out the queued message(s) to the appropriate transport connection using the appropriate send mechanism (e.g., for SCTP, the SEND primitive in [RFC4960] would be used); and, F) if the local cache is implemented, append/update the local cache with the mapping information received in the ENRP server's response. Also, record the local transport information (e.g., the SCTP association id) if any new transport state was created.
For more on the ENRP server request procedures see [RFC5353]. Optionally, the ASAP Endpoint of the sender may return a Pool Element handle of the selected PE to the application after sending the message. This PE handle can then be used for future transmissions to that same PE (see Section 6.5.3). Section 3.7 defines the failover procedures for cases where the selected PE is found unreachable.6.5.2. Pool Element Selection
Each time an ASAP User sends a message to a pool that contains more than one PE, the sender's ASAP Endpoint must select one of the PEs in the pool as the receiver of the current message. The selection is made according to the current server pooling policy of the pool to which the message is sent. Note, no selection is needed if the ASAP_SEND_TOALL option is set (see Section 6.5.5). Together with the server pooling policy, each PE can also specify a Policy Value for itself at the registration time. The meaning of the Policy Value depends on the current server pooling policy of the group. A PE can also change its Policy Value whenever it desires, by re-registering itself with the handlespace with a new Policy Value. Re-registration shall be done by simply sending another ASAP_REGISTRATION to its Home ENRP server (see Section 2.2.1). One basic policy is defined in this document; others can be found in [RFC5356]6.5.2.1. Round-Robin Policy
When an ASAP Endpoint sends messages by Pool Handle and Round-Robin is the current policy of that Pool, the ASAP Endpoint of the sender will select the receiver for each outbound message by Round-Robining through all the registered PEs in that Pool, in an attempt to achieve an even distribution of outbound messages. Note that in a large server pool, the ENRP server might not send back all PEs to the ASAP client. In this case, the client or PU will be performing a Round- Robin policy on a subset of the entire Pool.6.5.3. Sending to a Pool Element Handle
In this case, the destinationAddress and typeOfAddress together indicate an ASAP Pool Element handle.
This requests that the ASAP Endpoint deliver the message to the PE
identified by the Pool Element handle.
The Pool Element handle should contain the Pool Handle and a
destination transport address of the destination PE or the Pool
Handle and the transport type. Other implementation dependent
elements may also be cached in a Pool Element handle.
The ASAP Endpoint shall use the transport address and transport type
to identify the endpoint with which to communicate. If no
communication state exists with the peer endpoint (and is required by
the transport protocol), the ASAP Endpoint MAY set up the needed
state and then invoke the SEND primitive for the particular transport
protocol to send the message to the PE.
In addition, if a local translation cache is supported, the endpoint
will:
A) send out the message to the transport address (or association id)
designated by the PE handle.
B) determine if the Pool Handle is in the local cache.
If it is *not*, the endpoint will:
i) ask the Home ENRP server for handle resolution on the pool
handle by sending an ASAP_HANDLE_RESOLUTION message (see
Section 2.2.5), and
ii) use the response to update the local cache.
If the pool handle is in the cache, the endpoint will only
update the pool handle if the cache is stale. A stale cache is
indicated by it being older than the protocol parameter
'stale.cache.value' (see Section 7.2).
Sections 3.5 and 6.9 define the failover procedures for cases where
the PE pointed to by the Pool Element handle is found to be
unreachable.
Optionally, the ASAP Endpoint may return the actual Pool Element
handle to which the message was sent (this may be different from the
Pool Element handle specified when the primitive is invoked, due to
the possibility of automatic failover).
6.5.4. Send by Transport Address
In this case, the destinationAddress and typeOfAddress together indicate a transport address and transport type. This directs the sender's ASAP Endpoint to send the message out to the specified transport address. No endpoint failover is supported when this form of send request is used. This form of send request effectively bypasses the ASAP endpoint.6.5.5. Message Delivery Options
The Options parameter passed in the various forms of the above data_send_request primitive gives directions to the sender's ASAP endpoint on special handling of the message delivery. The value of the Options parameter is generated by bit-wise "OR"ing of the following pre-defined constants: ASAP_USE_DEFAULT: 0x0000 Use default setting. ASAP_SEND_FAILOVER: 0x0001 Enables PE failover on this message. In the case where the first selected PE or the PE pointed to by the PE handle is found unreachable, the sender's ASAP Endpoint SHOULD re-select an alternate PE from the same pool if one exists, and silently re-send the message to this newly selected endpoint. Note that this is a best-effort service. Applications should be aware that messages can be lost during the failover process, even if the underlying transport supports retrieval of unacknowledged data (e.g., SCTP). (Example: messages acknowledged by the SCTP layer at a PE, but not yet read by the PE when a PE failure occurs.) In the case where the underlying transport does not support such retrieval (e.g., TCP), any data already submitted by ASAP to the transport layer may be lost upon failover. ASAP_SEND_NO_FAILOVER: 0x0002 This option prohibits the sender's ASAP Endpoint from re-sending the message to any alternate PE in case that the first selected PE, or the PE pointed to by the PE handle, is found to be unreachable. Instead, the sender's ASAP Endpoint shall notify its upper layer about the unreachability with an Error.Report and return any unsent data. ASAP_SEND_TO_LAST: 0x0004 This option requests that the sender's ASAP Endpoint send the message to the same PE in the pool to which the previous message destined to this pool was sent.
ASAP_SEND_TO_ALL: 0x0008 When sending by Pool Handle, this option
directs the sender's ASAP endpoint to send a copy of the message
to all the PEs, except for the sender itself if the sender is a PE
in that pool.
ASAP_SEND_TO_SELF: 0x0010 This option only applies in combination
with the ASAP_SEND_TO_ALL option. It permits the sender's ASAP
Endpoint to also deliver a copy of the message to itself if the
sender is a PE of the pool (i.e., loop-back).
ASAP_SCTP_UNORDER: 0x1000 This option requests that the transport
layer send the current message using un-ordered delivery (note the
underlying transport must support un-ordered delivery for this
option to be effective).
6.6. Data.Received Notification
Format: data.received(messageReceived, sizeOfMessage,
senderAddress, typeOfAddress)
When a new user message is received, the ASAP Endpoint of the
receiver uses this notification to pass the message to its upper
layer.
Along with the message being passed, the ASAP Endpoint of the
receiver should also indicate to its upper layer the message senders
address. The sender's address can be in the form of either an SCTP
association id, TCP transport address, UDP transport address, or an
ASAP Pool Element handle.
A) If the handle translation local cache is implemented at the
receiver's ASAP Endpoint, a reverse mapping from the sender's IP
address to the pool handle should be performed, and if the mapping
is successful, the sender's ASAP Pool Element handle should be
constructed and passed in the senderAddress field.
B) If there is no local cache or the reverse mapping is not
successful, the SCTP association id or other transport specific
identification (if SCTP is not being used) should be passed in the
senderAddress field.
6.7. Error.Report Notification
Format: error.report(destinationAddress, typeOfAddress, failedMessage, sizeOfMessage) An error.report should be generated to notify the ASAP User about failed message delivery as well as other abnormalities. The destinationAddress and typeOfAddress together indicate to whom the message was originally sent. The address type can be either an ASAP Pool Element handle, association id, or a transport address. The original message (or the first portion of it if the message is too big) and its size should be passed in the failedMessage and sizeOfMessage fields, respectively.6.8. Examples
These examples assume an underlying SCTP transport between the PE and PU. Other transports are possible, but SCTP is utilized in the examples for illustrative purposes. Note that all communication between the PU and ENRP server and the PE and ENRP servers would be using SCTP.6.8.1. Send to a New Pool
This example shows the event sequence when a Pool User sends the message "hello" to a pool that is not in the local translation cache (assuming local caching is supported). ENRP Server PU new-handle:PEx | | | | +---+ | | | 1 | | |2. ASAP_HANDLE_RESOLUTION +---+ | |<-------------------------------| | | +---+ | | | 3 | | |4. ASAP_HANDLE_RESOLUTION_RSP +---+ | |------------------------------->| | | +---+ | | | 5 | | | +---+ 6. "hello1" | | |---------------->| | | |
1) The user at PU invokes:
data_send_request("new-handle", handle-type, "hello1", 6, 0);
The ASAP Endpoint, in response, looks up the pool "new-handle" in
its local cache, but fails to find it.
2) The ASAP Endpoint of the PU queues the message and sends an
ASAP_HANDLE_RESOLUTION request to the ENRP server asking for all
information about pool "new-handle".
3) A T1-ENRPrequest timer is started while the ASAP Endpoint is
waiting for the response from the ENRP server.
4) The ENRP server responds to the query with an
ASAP_HANDLE_RESOLUTION_RESPONSE message that contains all the
information about pool "new-handle".
5) ASAP at PU cancels the T1-ENRPrequest timer and populate its local
cache with information on pool "new-handle".
6) Based on the server pooling policy of pool "new-handle", ASAP at
PU selects the destination PE (PEx), sets up, if necessary, an
SCTP association towards PEx (explicitly or implicitly), and sends
out the queued "hello1" user message.
6.8.2. Send to a Cached Pool Handle
This shows the event sequence when the ASAP User PU sends another
message to the pool "new-handle" after what happened in
Section 6.8.1.
ENRP Server PU new-handle:PEx
| | |
| +---+ |
| | 1 | |
| +---+ 2. "hello2" |
| |---------------->|
| | |
1) The user at PU invokes:
data_send_request("new-handle", handle-type, "hello2", 6, 0);
The ASAP Endpoint, in response, looks up the pool "new-handle" in
its local cache and finds the mapping information.
2) Based on the server pooling policy of "new-handle", ASAP at PU
selects the PE (assuming EPx is selected again), and sends out
"hello2" message (assuming the SCTP association is already set
up).
6.9. PE Send Failure
When the ASAP Endpoint in a PE or PU attempts to send a message to a
PE and fails, the failed sender will report the event as described in
Section 3.5.
Additional primitives are also defined in this section to support
those user applications that do not wish to use ASAP as the actual
transport.
6.9.1. Translation.Request Primitive
Format: translation.request(Pool-Handle)
If the address type is a Pool Handle and a local handle translation
cache exists, the ASAP Endpoint should look within its translation
cache and return the current known transport types, ports, and
addresses to the caller.
If the Pool Handle does not exist in the local handle cache or no
handle cache exists, the ASAP Endpoint will send an
ASAP_HANDLE_RESOLUTION request using the Pool Handle. Upon
completion of the handle resolution, the ASAP Endpoint should
populate the local handle cache (if a local handle cache is
supported) and return the transport types, ports, and addresses to
the caller.
6.9.2. Transport.Failure Primitive
Format: transport.failure(Pool-Handle, Transport-address) If an external user encounters a failure in sending to a PE and is *not* using ASAP, it can use this primitive to report the failure to the ASAP endpoint. ASAP will send an ASAP_ENDPOINT_UNREACHABLE to the "Home" ENRP server in response to this primitive. Note ASAP SHOULD NOT send an ASAP_ENDPOINT_UNREACHABLE *unless* the user has actually made a previous request to send data to the PE.
(page 42 continued on part 3)