RFC 2273
SNMPv3 Applications
Network Working Group D. Levi Request for Comments: 2273 SNMP Research, Inc. Obsoletes: 2263 P. Meyer Category: Standards Track Secure Computing Corporation B. Stewart Cisco Systems January 1998 SNMPv3 Applications Status of this Memo This document specifies an Internet standards track protocol for the Internet community, and requests discussion and suggestions for improvements. Please refer to the current edition of the "Internet Official Protocol Standards" (STD 1) for the standardization state and status of this protocol. Distribution of this memo is unlimited. Copyright Notice Copyright (C) The Internet Society (1998). All Rights Reserved. IANA Note Due to a clerical error in the assignment of the snmpModules in this memo, this RFC provides the corrected number assignments for this protocol. This memo obsoletes RFC 2263. Abstract This memo describes five types of SNMP applications which make use of an SNMP engine as described in [RFC2271]. The types of application described are Command Generators, Command Responders, Notification Originators, Notification Receivers, and Proxy Forwarders. This memo also defines MIB modules for specifying targets of management operations, for notification filtering, and for proxy forwarding. Table of Contents 1 Overview ..................................................... 2 1.1 Command Generator Applications ............................. 3 1.2 Command Responder Applications ............................. 3 1.3 Notification Originator Applications ....................... 3 1.4 Notification Receiver Applications ......................... 3 1.5 Proxy Forwarder Applications ............................... 3 2 Management Targets ........................................... 5
3 Elements Of Procedure ........................................ 6 3.1 Command Generator Applications ............................. 6 3.2 Command Responder Applications ............................. 8 3.3 Notification Originator Applications ....................... 13 3.4 Notification Receiver Applications ......................... 16 3.5 Proxy Forwarder Applications ............................... 18 3.5.1 Request Forwarding ....................................... 19 3.5.1.1 Processing an Incoming Request ......................... 19 3.5.1.2 Processing an Incoming Response ........................ 22 3.5.1.3 Processing an Incoming Report Indication ............... 23 3.5.2 Notification Forwarding .................................. 24 4 The Structure of the MIB Modules ............................. 27 4.1 The Management Target MIB Module ........................... 27 4.1.1 Tag Lists ................................................ 28 4.1.2 Definitions .............................................. 28 4.2 The Notification MIB Module ................................ 41 4.2.1 Definitions .............................................. 42 4.3 The Proxy MIB Module ....................................... 53 4.3.1 Definitions .............................................. 53 5 Identification of Management Targets in Notification Originators ............................................... 59 6 Notification Filtering ....................................... 60 7 Management Target Translation in Proxy Forwarder Applications .............................................. 61 7.1 Management Target Translation for Request Forwarding ....... 61 7.2 Management Target Translation for Notification Forwarding ........................................................... 62 8 Intellectual Property ........................................ 63 9 Acknowledgments .............................................. 64 10 Security Considerations ..................................... 65 11 References .................................................. 65 12 Editors' Address ............................................ 67 A. Trap Configuration Example .................................. 68 B. Full Copyright Statement .................................... 70 1. Overview This document describes five types of SNMP applications: - Applications which initiate SNMP Get, GetNext, GetBulk, and/or Set requests, called 'command generators.' - Applications which respond to SNMP Get, GetNext, GetBulk, and/or Set requests, called 'command responders.' - Applications which generate notifications, called 'notification originators.'
- Applications which receive notifications, called 'notification
receivers.'
- Applications which forward SNMP Get, GetNext, GetBulk, and/or
Set requests or notifications, called 'proxy forwarder.'
Note that there are no restrictions on which types of applications
may be associated with a particular SNMP engine. For example, a
single SNMP engine may, in fact, be associated with both command
generator and command responder applications.
1.1. Command Generator Applications
A command generator application initiates SNMP Get, GetNext, GetBulk,
and/or Set requests, as well as processing the response to a request
which it generated.
1.2. Command Responder Applications
A command responder application receives SNMP Get, GetNext, GetBulk,
and/or Set requests destined for the local system as indicated by the
fact that the contextEngineID in the received request is equal to
that of the local engine through which the request was received. The
command responder application will perform the appropriate protocol
operation, using access control, and will generate a response message
to be sent to the request's originator.
1.3. Notification Originator Applications
A notification originator application conceptually monitors a system
for particular events or conditions, and generates Trap and/or Inform
messages based on these events or conditions. A notification
originator must have a mechanism for determining where to send
messages, and what SNMP version and security parameters to use when
sending messages. A mechanism and MIB module for this purpose is
provided in this document.
1.4. Notification Receiver Applications
A notification receiver application listens for notification
messages, and generates response messages when a message containing
an Inform PDU is received.
1.5. Proxy Forwarder Applications
A proxy forwarder application forwards SNMP messages. Note that
implementation of a proxy forwarder application is optional. The
sections describing proxy (4.5, 5.3, and 8) may be skipped for
implementations that do not include a proxy forwarder application.
The term "proxy" has historically been used very loosely, with
multiple different meanings. These different meanings include (among
others):
(1) the forwarding of SNMP requests to other SNMP entities without
regard for what managed object types are being accessed; for
example, in order to forward an SNMP request from one transport
domain to another, or to translate SNMP requests of one version
into SNMP requests of another version;
(2) the translation of SNMP requests into operations of some non-SNMP
management protocol; and
(3) support for aggregated managed objects where the value of one
managed object instance depends upon the values of multiple other
(remote) items of management information.
Each of these scenarios can be advantageous; for example, support for
aggregation of management information can significantly reduce the
bandwidth requirements of large-scale management activities.
However, using a single term to cover multiple different scenarios
causes confusion.
To avoid such confusion, this document uses the term "proxy" with a
much more tightly defined meaning. The term "proxy" is used in this
document to refer to a proxy forwarder application which forwards
either SNMP requests, notifications, and responses without regard for
what managed objects are contained within requests or notifications.
This definition is most closely related to the first definition
above. Note, however, that in the SNMP architecture [RFC2271], a
proxy forwarder is actually an application, and need not be
associated with what is traditionally thought of as an SNMP agent.
Specifically, the distinction between a traditional SNMP agent and a
proxy forwarder application is simple:
- a proxy forwarder application forwards requests and/or
notifications to other SNMP engines according to the context,
and irrespective of the specific managed object types being
accessed, and forwards the response to such previously
forwarded messages back to the SNMP engine from which the
original message was received;
- in contrast, the command responder application that is part of
what is traditionally thought of as an SNMP agent, and which
processes SNMP requests according to the (names of the)
individual managed object types and instances being accessed,
is NOT a proxy forwarder application from the perspective of
this document.
Thus, when a proxy forwarder application forwards a request or
notification for a particular contextEngineID / contextName pair, not
only is the information on how to forward the request specifically
associated with that context, but the proxy forwarder application has
no need of a detailed definition of a MIB view (since the proxy
forwarder application forwards the request irrespective of the
managed object types).
In contrast, a command responder application must have the detailed
definition of the MIB view, and even if it needs to issue requests to
other entities, via SNMP or otherwise, that need is dependent on the
individual managed object instances being accessed (i.e., not only on
the context).
Note that it is a design goal of a proxy forwarder application to act
as an intermediary between the endpoints of a transaction. In
particular, when forwarding Inform requests, the associated response
is forwarded when it is received from the target to which the Inform
request was forwarded, rather than generating a response immediately
when an Inform request is received.
2. Management Targets
Some types of applications (notification generators and proxy
forwarders in particular) require a mechanism for determining where
and how to send generated messages. This document provides a
mechanism and MIB module for this purpose. The set of information
that describes where and how to send a message is called a
'Management Target', and consists of two kinds of information:
- Destination information, consisting of a transport domain and
a transport address. This is also termed a transport
endpoint.
- SNMP parameters, consisting of message processing model,
security model, security level, and security name information.
The SNMP-TARGET-MIB module described later in this document contains
one table for each of these types of information. There can be a
many-to-many relationship in the MIB between these two types of
information. That is, there may be multiple transport endpoints
associated with a particular set of SNMP parameters, or a particular
transport endpoint may be associated with several sets of SNMP
parameters.
3. Elements Of Procedure The following sections describe the procedures followed by each type of application when generating messages for transmission or when processing received messages. Applications communicate with the Dispatcher using the abstract service interfaces defined in [RFC2271]. 3.1. Command Generator Applications A command generator initiates an SNMP request by calling the Dispatcher using the following abstract service interface: statusInformation = -- sendPduHandle if success -- errorIndication if failure sendPdu( IN transportDomain -- transport domain to be used IN transportAddress -- destination network address IN messageProcessingModel -- typically, SNMP version IN securityModel -- Security Model to use IN securityName -- on behalf of this principal IN securityLevel -- Level of Security requested IN contextEngineID -- data from/at this entity IN contextName -- data from/in this context IN pduVersion -- the version of the PDU IN PDU -- SNMP Protocol Data Unit IN expectResponse -- TRUE or FALSE ) Where: - The transportDomain is that of the destination of the message. - The transportAddress is that of the destination of the message. - The messageProcessingModel indicates which Message Processing Model the application wishes to use. - The securityModel is the security model that the application wishes to use. - The securityName is the security model independent name for the principal on whose behalf the application wishes the message is to be generated. - The securityLevel is the security level that the application wishes to use.
- The contextEngineID is provided by the command generator if it
wishes to explicitly specify the location of the management
information it is requesting.
- The contextName is provided by the command generator if it
wishes to explicitly specify the local context name for the
management information it is requesting.
- The pduVersion indicates the version of the PDU to be sent.
- The PDU is a value constructed by the command generator
containing the management operation that the command generator
wishes to perform.
- The expectResponse argument indicates that a response is
expected.
The result of the sendPdu interface indicates whether the PDU was
successfully sent. If it was successfully sent, the returned value
will be a sendPduHandle. The command generator should store the
sendPduHandle so that it can correlate a response to the original
request.
The Dispatcher is responsible for delivering the response to a
particular request to the correct command generator application. The
abstract service interface used is:
processResponsePdu( -- process Response PDU
IN messageProcessingModel -- typically, SNMP version
IN securityModel -- Security Model in use
IN securityName -- on behalf of this principal
IN securityLevel -- Level of Security
IN contextEngineID -- data from/at this SNMP entity
IN contextName -- data from/in this context
IN pduVersion -- the version of the PDU
IN PDU -- SNMP Protocol Data Unit
IN statusInformation -- success or errorIndication
IN sendPduHandle -- handle from sendPDU
)
Where:
- The messageProcessingModel is the value from the received
response.
- The securityModel is the value from the received response.
- The securityName is the value from the received response.
- The securityLevel is the value from the received response.
- The contextEngineID is the value from the received response.
- The contextName is the value from the received response.
- The pduVersion indicates the version of the PDU in the
received response.
- The PDU is the value from the received response.
- The statusInformation indicates success or failure in
receiving the response.
- The sendPduHandle is the value returned by the sendPdu call
which generated the original request to which this is a
response.
The procedure when a command generator receives a message is as
follows:
(1) If the received values of messageProcessingModel, securityModel,
securityName, contextEngineID, contextName, and pduVersion are not
all equal to the values used in the original request, the response
is discarded.
(2) The operation type, request-id, error-status, error-index, and
variable-bindings are extracted from the PDU and saved. If the
request-id is not equal to the value used in the original request,
the response is discarded.
(3) At this point, it is up to the application to take an appropriate
action. The specific action is implementation dependent. If the
statusInformation indicates that the request failed, an appropriate
action might be to attempt to transmit the request again, or to
notify the person operating the application that a failure
occurred.
3.2. Command Responder Applications
Before a command responder application can process messages, it must
first associate itself with an SNMP engine. The abstract service
interface used for this purpose is:
statusInformation = -- success or errorIndication
registerContextEngineID(
IN contextEngineID -- take responsibility for this one
IN pduType -- the pduType(s) to be registered
)
Where:
- The statusInformation indicates success or failure of the
registration attempt.
- The contextEngineID is equal to the snmpEngineID of the SNMP
engine with which the command responder is registering.
- The pduType indicates a Get, GetNext, GetBulk, or Set pdu.
Note that if another command responder application is already
registered with an SNMP engine, any further attempts to register with
the same contextEngineID and pduType will be denied. This implies
that separate command responder applications could register
separately for the various pdu types. However, in practice this is
undesirable, and only a single command responder application should
be registered with an SNMP engine at any given time.
A command responder application can disassociate with an SNMP engine
using the following abstract service interface:
unregisterContextEngineID(
IN contextEngineID -- give up responsibility for this one
IN pduType -- the pduType(s) to be unregistered
)
Where:
- The contextEngineID is equal to the snmpEngineID of the SNMP
engine with which the command responder is cancelling the
registration.
- The pduType indicates a Get, GetNext, GetBulk, or Set pdu.
Once the command responder has registered with the SNMP engine, it
waits to receive SNMP messages. The abstract service interface used
for receiving messages is:
processPdu( -- process Request/Notification PDU
IN messageProcessingModel -- typically, SNMP version
IN securityModel -- Security Model in use
IN securityName -- on behalf of this principal
IN securityLevel -- Level of Security
IN contextEngineID -- data from/at this SNMP entity
IN contextName -- data from/in this context
IN pduVersion -- the version of the PDU
IN PDU -- SNMP Protocol Data Unit
IN maxSizeResponseScopedPDU -- maximum size of the Response PDU
IN stateReference -- reference to state information
) -- needed when sending a response
Where:
- The messageProcessingModel indicates which Message Processing
Model received and processed the message.
- The securityModel is the value from the received message.
- The securityName is the value from the received message.
- The securityLevel is the value from the received message.
- The contextEngineID is the value from the received message.
- The contextName is the value from the received message.
- The pduVersion indicates the version of the PDU in the
received message.
- The PDU is the value from the received message.
- The maxSizeResponseScopedPDU is the maximum allowable size of
a ScopedPDU containing a Response PDU (based on the maximum
message size that the originator of the message can accept).
- The stateReference is a value which references cached
information about each received request message. This value
must be returned to the Dispatcher in order to generate a
response.
The procedure when a message is received is as follows.
(1) The operation type is determined from the ASN.1 tag value
associated with the PDU parameter. The operation type should
always be one of the types previously registered by the
application.
(2) The request-id is extracted from the PDU and saved.
(3) If the SNMPv2 operation type is GetBulk, the non-repeaters and
max-repetitions values are extracted from the PDU and saved.
(4) The variable-bindings are extracted from the PDU and saved.
(5) The management operation represented by the SNMPv2 operation type
is performed with respect to the relevant MIB view within the
context named by the contextName, according to the procedures set
forth in [RFC1905]. The relevant MIB view is determined by the
securityLevel, securityModel, contextName, securityName, and SNMPv2
operation type. To determine whether a particular object instance
is within the relevant MIB view, the following abstract service
interface is called:
statusInformation = -- success or errorIndication
isAccessAllowed(
IN securityModel -- Security Model in use
IN securityName -- principal who wants to access
IN securityLevel -- Level of Security
IN viewType -- read, write, or notify view
IN contextName -- context containing variableName
IN variableName -- OID for the managed object
)
Where:
- The securityModel is the value from the received message.
- The securityName is the value from the received message.
- The securityLevel is the value from the received message.
- The viewType indicates whether the PDU type is a read or write
operation.
- The contextName is the value from the received message.
- The variableName is the object instance of the variable for
which access rights are to be checked.
Normally, the result of the management operation will be a new PDU
value, and processing will continue in step (6) below. However, at
any time during the processing of the management operation:
- If the isAccessAllowed ASI returns a noSuchView,
noAccessEntry, or noGroupName error, processing of the
management operation is halted, a PDU value is contructed
using the values from the originally received PDU, but
replacing the error_status with an authorizationError code,
and error_index value of 0, and control is passed to step (6)
below.
- If the isAccessAllowed ASI returns an otherError, processing
of the management operation is halted, a different PDU value
is contructed using the values from the originally received
PDU, but replacing the error_status with a genError code, and
control is passed to step (6) below.
- If the isAccessAllowed ASI returns a noSuchContext error,
processing of the management operation is halted, no result
PDU is generated, the snmpUnknownContexts counter is
incremented, and control is passed to step (6) below.
- If the context named by the contextName parameter is
unavailable, processing of the management operation is halted,
no result PDU is generated, the snmpUnavailableContexts
counter is incremented, and control is passed to step (6)
below.
(6) The Dispatcher is called to generate a response or report message.
The abstract service interface is:
returnResponsePdu(
IN messageProcessingModel -- typically, SNMP version
IN securityModel -- Security Model in use
IN securityName -- on behalf of this principal
IN securityLevel -- same as on incoming request
IN contextEngineID -- data from/at this SNMP entity
IN contextName -- data from/in this context
IN pduVersion -- the version of the PDU
IN PDU -- SNMP Protocol Data Unit
IN maxSizeResponseScopedPDU -- maximum size of the Response PDU
IN stateReference -- reference to state information
-- as presented with the request
IN statusInformation -- success or errorIndication
) -- error counter OID/value if error
Where:
- The messageProcessingModel is the value from the processPdu
call.
- The securityModel is the value from the processPdu call.
- The securityName is the value from the processPdu call.
- The securityLevel is the value from the processPdu call.
- The contextEngineID is the value from the processPdu call.
- The contextName is the value from the processPdu call.
- The pduVersion indicates the version of the PDU to be
returned. If no result PDU was generated, the pduVersion is
an undefined value.
- The PDU is the result generated in step (5) above. If no
result PDU was generated, the PDU is an undefined value.
- The maxSizeResponseScopedPDU is a local value indicating the
maximum size of a ScopedPDU that the application can accept.
- The stateReference is the value from the processPdu call.
- The statusInformation either contains an indication that no
error occurred and that a response should be generated, or
contains an indication that an error occurred along with the
OID and counter value of the appropriate error counter object.
Note that a command responder application should always call the
returnResponsePdu abstract service interface, even in the event of an
error such as a resource allocation error. In the event of such an
error, the PDU value passed to returnResponsePdu should contain
appropriate values for errorStatus and errorIndex.
3.3. Notification Originator Applications
A notification originator application generates SNMP notification
messages. A notification message may, for example, contain an
SNMPv2-Trap PDU or an Inform PDU. However, a particular
implementation is not required to be capable of generating both types
of messages.
Notification originator applications require a mechanism for
identifying the management targets to which notifications should be
sent. The particular mechanism used is implementation dependent.
However, if an implementation makes the configuration of management
targets SNMP manageable, it MUST use the SNMP-TARGET-MIB module
described in this document.
When a notification originator wishes to generate a notification, it
must first determine in which context the information to be conveyed
in the notification exists, i.e., it must determine the
contextEngineID and contextName. It must then determine the set of
management targets to which the notification should be sent. The
application must also determine, for each management target, whether
the notification message should contain an SNMPv2-Trap PDU or Inform
PDU, and if it is to contain an Inform PDU, the number of retries and
retransmission algorithm.
The mechanism by which a notification originator determines this
information is implementation dependent. Once the application has
determined this information, the following procedure is performed for
each management target:
(1) Any appropriate filtering mechanisms are applied to determine
whether the notification should be sent to the management target.
If such filtering mechanisms determine that the notification should
not be sent, processing continues with the next management target.
Otherwise,
(2) The appropriate set of variable-bindings is retrieved from local
MIB instrumentation within the relevant MIB view. The relevant MIB
view is determined by the securityLevel, securityModel,
contextName, and securityName of the management target. To
determine whether a particular object instance is within the
relevant MIB view, the isAccessAllowed abstract service interface
is used, in the same manner as described in the preceding section.
If the statusInformation returned by isAccessAllowed does not
indicate accessAllowed, the notification is not sent to the
management target.
(3) A PDU is constructed using a locally unique request-id value, an
operation type of SNMPv2-Trap or Inform, an error-status and
error-index value of 0, and the variable-bindings supplied
previously in step (2).
(4) If the notification contains an SNMPv2-Trap PDU, the Dispatcher is
called using the following abstract service interface:
statusInformation = -- sendPduHandle if success
-- errorIndication if failure
sendPdu(
IN transportDomain -- transport domain to be used
IN transportAddress -- destination network address
IN messageProcessingModel -- typically, SNMP version
IN securityModel -- Security Model to use
IN securityName -- on behalf of this principal
IN securityLevel -- Level of Security requested
IN contextEngineID -- data from/at this entity
IN contextName -- data from/in this context
IN pduVersion -- the version of the PDU
IN PDU -- SNMP Protocol Data Unit
IN expectResponse -- TRUE or FALSE
)
Where:
- The transportDomain is that of the management target.
- The transportAddress is that of the management target.
- The messageProcessingModel is that of the management target.
- The securityModel is that of the management target.
- The securityName is that of the management target.
- The securityLevel is that of the management target.
- The contextEngineID is the value originally determined for the
notification.
- The contextName is the value originally determined for the
notification.
- The pduVersion is the version of the PDU to be sent.
- The PDU is the value constructed in step (3) above.
- The expectResponse argument indicates that no response is
expected.
Otherwise,
(5) If the notification contains an Inform PDU, then:
a) The Dispatcher is called using the sendPdu abstract service
interface as described in step (4) above, except that the
expectResponse argument indicates that a response is expected.
b) The application caches information about the management
target.
c) If a response is received within an appropriate time interval
from the transport endpoint of the management target, the
notification is considered acknowledged and the cached
information is deleted. Otherwise,
d) If a response is not received within an appropriate time
period, or if a report indication is received, information
about the management target is retrieved from the cache, and
steps a) through d) are repeated. The number of times these
steps are repeated is equal to the previously determined retry
count. If this retry count is exceeded, the acknowledgement
of the notification is considered to have failed, and
processing of the notification for this management target is
halted.
Responses to Inform PDU notifications will be received via the
processResponsePDU abstract service interface.
3.4. Notification Receiver Applications
Notification receiver applications receive SNMP Notification messages
from the Dispatcher. Before any messages can be received, the
notification receiver must register with the Dispatcher using the
registerContextEngineID abstract service interface. The parameters
used are:
- The contextEngineID is an undefined 'wildcard' value.
Notifications are delivered to a registered notification
receiver regardless of the contextEngineID contained in the
notification message.
- The pduType indicates the type of notifications that the
application wishes to receive (for example, SNMPv2-Trap PDUs
or Inform PDUs).
Once the notification receiver has registered with the Dispatcher,
messages are received using the processPdu abstract service
interface. Parameters are:
- The messageProcessingModel indicates which Message Processing
Model received and processed the message.
- The securityModel is the value from the received message.
- The securityName is the value from the received message.
- The securityLevel is the value from the received message.
- The contextEngineID is the value from the received message.
- The contextName is the value from the received message.
- The pduVersion indicates the version of the PDU in the
received message.
- The PDU is the value from the received message.
- The maxSizeResponseScopedPDU is the maximum allowable size of
a ScopedPDU containing a Response PDU (based on the maximum
message size that the originator of the message can accept).
- If the message contains an SNMPv2-Trap PDU, the stateReference
is undefined and unused. Otherwise, the stateReference is a
value which references cached information about the
notification. This value must be returned to the Dispatcher
in order to generate a response.
When an SNMPv2-Trap PDU is delivered to a notification receiver
application, it first extracts the SNMP operation type, request-id,
error-status, error-index, and variable-bindings from the PDU. After
this, processing depends on the particular implementation.
When an Inform PDU is received, the notification receiver application
follows the following procedure:
(1) The SNMPv2 operation type, request-id, error-status, error-index,
and variable-bindings are extracted from the PDU.
(2) A Response PDU is constructed using the extracted request-id and
variable-bindings, and with error-status and error-index both set
to 0.
(3) The Dispatcher is called to generate a response message using the
returnResponsePdu abstract service interface. Parameters are:
- The messageProcessingModel is the value from the processPdu
call.
- The securityModel is the value from the processPdu call.
- The securityName is the value from the processPdu call.
- The securityLevel is the value from the processPdu call.
- The contextEngineID is the value from the processPdu call.
- The contextName is the value from the processPdu call.
- The pduVersion indicates the version of the PDU to be
returned.
- The PDU is the result generated in step (2) above.
- The maxSizeResponseScopedPDU is a local value indicating the
maximum size of a ScopedPDU that the application can accept.
- The stateReference is the value from the processPdu call.
- The statusInformation indicates that no error occurred and
that a response should be generated.
(page 18 continued on part 2)
