Network Working Group C. Kalbfleisch Request for Comments: 2564 Verio, Inc. Category: Standards Track C. Krupczak Empire Technologies, Inc. R. Presuhn BMC Software, Inc. J. Saperia IronBridge Networks May 1999 Application Management MIB 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 (1999). All Rights Reserved. Abstract This memo defines a standards track portion of the Management Information Base (MIB) for use with network management protocols in the Internet Community. In particular, it defines objects used for the management of applications. This MIB complements the System Application MIB, providing for the management of applications' common attributes which could not typically be observed without the cooperation of the software being managed. Table of Contents 1. Introduction and Overview ................................... 2 2. The SNMP Management Framework ............................... 4 3. Architecture ................................................ 5 3.1. Relationships to other MIBs ............................... 5 3.1.1. Relationship to the System Application MIB .............. 5 3.1.2. Relationship to the Host Resources MIB .................. 6 3.1.3. Relationship to NSM ..................................... 6 4. MIB Structure ............................................... 6 4.1. The service-level tables .................................. 8 4.1.1. The service name to service instance table .............. 8 4.1.2. The service instance to service name table .............. 9 4.1.3. The service instance to running application element table 9 4.1.4. The running application element to service instance table 9
4.2. The I/O channel group ..................................... 9 4.2.1. The open channels table ................................. 10 4.2.2. The open files table .................................... 10 4.2.3. The open connections table .............................. 11 4.2.4. The transaction stream summary table .................... 12 4.2.5. The transaction flow statistics table ................... 13 4.2.6. The transaction kind statistics table ................... 13 4.3. The former channel group .................................. 13 4.3.1. The former channel control table ........................ 14 4.3.2. The former channel table ................................ 14 4.3.3. The former connection table ............................. 14 4.3.4. The former file table ................................... 14 4.3.5. The transaction history tables .......................... 14 4.4. The running element status and control group .............. 15 4.4.1. The running application element status table ............ 15 4.4.2. The running application element control table ........... 15 5. Definitions ................................................. 16 6. Implementation Issues ....................................... 80 7. Intellectual Property ....................................... 80 8. Acknowledgements ............................................ 81 9. Security Considerations ..................................... 81 10. References ................................................. 82 11. Authors' Addresses ......................................... 84 12. Full Copyright Statement ................................... 86 1. Introduction and Overview This document furthers the work begun in the systems application MIB . The development of the "Host Resources MIB" , "Network Services Monitoring MIB" , "Mail Monitoring MIB" , "Relational Database Management System (RDBMS) Management Information Base (MIB) using SMIv2" , "Entity MIB using SMIv2" , and "Applicability of Standards Track MIBs to Management of World Wide Web Servers"  provides us with a base of experience in making a variety of applications visible to management; this specification abstracts out the common aspects of applications management and provides a generic base usable for the management of almost any application. The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119 . Due to the design decision to not require application instrumentation, many important topics were not handled in system application MIB . The following topics are within the scope of this document:
- Support for generic application throughput measurements; - Providing MIB definitions that allow managed entities to report what they considered to be units of work; - Providing support for generic application response time monitoring capabilities; (Note that APIs for this purpose have already been developed, an example of such an API is to be found in the "Application Response Measurement (ARM) API Guide, Version 2" .) - Provide explicit support for the management of applications distributed within a single managed system ("local" distribution); - Address generic resource management issues, including: - files in use; - I/O statistics (from the application's perspective, not at the operating system or device driver level); - application-layer networking resource usage - Facilities for the control of applications, including: - Stopping application elements - Suspending and resuming application elements; - Requesting reconfiguration (e.g., SIGHUP). Note that these issues are addressed at least in part by other (non- IETF) standards work, including "ITU-T Recommendation X.744 | ISO/IEC IS 10164-18:1996"  and "IEEE P1387.2, POSIX System Administration - Part 2: Software Administration" .
2. The SNMP Management Framework The SNMP Management Framework presently consists of five major components: An overall architecture, described in RFC 2571 . Mechanisms for describing and naming objects and events for the purpose of management. The first version of this Structure of Management Information (SMI) is called SMIv1 and described in STD 16, RFC 1155 , STD 16, RFC 1212  and RFC 1215 . The second version, called SMIv2, is described in STD 58, RFC 2578 , RFC 2579  and RFC 2580 . Message protocols for transferring management information. The first version of the SNMP message protocol is called SNMPv1 and described in STD 15, RFC 1157 . A second version of the SNMP message protocol, which is not an Internet standards track protocol, is called SNMPv2c and described in RFC 1901  and RFC 1906 . The third version of the message protocol is called SNMPv3 and described in RFC 1906 , RFC 2572  and RFC 2574 . Protocol operations for accessing management information. The first set of protocol operations and associated PDU formats is described in STD 15, RFC 1157 . A second set of protocol operations and associated PDU formats is described in RFC 1905 . A set of fundamental applications described in RFC 2573  and the view-based access control mechanism described in RFC 2575 . Managed objects are accessed via a virtual information store, termed the Management Information Base or MIB. Objects in the MIB are defined using the mechanisms defined in the SMI. This memo specifies a MIB module that is compliant to the SMIv2. A MIB conforming to the SMIv1 can be produced through the appropriate translations. The resulting translated MIB must be semantically equivalent, except where objects or events are omitted because no translation is possible (use of Counter64). Some machine readable information in SMIv2 will be converted into textual descriptions in SMIv1 during the translation process. However, this loss of machine readable information is not considered to change the semantics of the MIB.
3. Architecture Object-oriented modeling techniques like subclassing and multiple inheritance can be emulated in the SNMP information model through the use of tables with common indexes. The challenge for the developer of management applications is to recognize those situations in which various aspects of a single logical resource are represented in several different tables, possibly defined in different MIBs. Most of the management information defined here may pertain to any number of applications in a managed system. The simplest way of supporting this requirement within the SNMP information model is to use tables. This means that the management information for a particular resource may be found in one or more rows of one or more tables; the fact that this information pertains to a single resource may be inferred from the index values used, possibly with the support of mapping tables. This also means that a single table may contain management information relevant to a number of applications. This has significant implementation implications; see the implementation issues section below for more information. 3.1. Relationships to other MIBs This section outlines the relationships of the components of this MIB (usually in the form of common indexing structures) to: - the systems applications MIB  - the host resources MIB  - the network services monitoring MIB  3.1.1. Relationship to the System Application MIB The system application MIB defines attributes for management of applications which can be realized without instrumenting the application itself. This specification extends that framework to include additional attributes which will typically require instrumentation within the managed resource. The sysApplRunElmtIndex is the key connection between these two MIBs; it is essential that implementations of this MIB and of the system applications MIB running concurrently on a given platform employ a consistent policy for assigning this value to identify running application elements.
3.1.2. Relationship to the Host Resources MIB The Host Resources MIB  supplies information on the hardware, operating system, installed and running software on a host. The Host Resources MIB has three hardware groups ("hrSystem", "hrStorage" and "hrDevice") and three software groups ("hrSWRun", "hrSWRunPerf" and "hrSWInstalled"). Of these, the software groups are of greatest significance to this MIB. The software groups define management information on the software used in the system. The information provided is grouped into (1) the currently running, (2) the performance and (3) the installed applications. The index "hrSWRunIndex" used in the "hrSWRunTable" and other tables to identify running software by process identifier (or equivalent) relates information in the Host Resources MIB to information in the System Applications MIB and this MIB. It is essential that the values assigned to hrSWRunIndex from the Host Resources MIB be consistent with the values used for sysApplRunElmtIndex. 3.1.3. Relationship to NSM The Network Services Monitoring MIB  is defined as the base set of attributes for managing network applications. The Application MIB includes information normally obtainable only from the managed resource itself, rather than the supporting system. Due to differences in index representation, the relationship between the Network Services Monitoring MIB and the Application MIB is not formally defined. 4. MIB Structure This MIB is organized into several groups, which in turn are organized into tables to provide the monitoring and control of information relevant to the management of applications. The groups model: - the service-level view of applications - information on open channels (files, connections, transaction streams) in use by applications - historical information on former channels - process-level status and control information
These groups are organized into various tables. Information for a particular running managed application appears in the form of entries in the appropriate tables. The tables are: - the tables providing a service-level view, including: - the service name to service instance table - the service instance to service name table - the service instance to running application element table - the running application element to service instance table - the tables providing information on I/O channels, including: - the table of open channels - the table of open files - the open connections table - the transaction statistics tables - historical information on I/O channels - the running application element status and control group - the running application element status table - the running application element control table In order to support SNMPv1, SNMPv2, and SNMPv3 environments, in cases where counter objects may potentially advance very rapidly, where sixty-four bit counters have been used thirty-two bit counters reporting the low-order thirty-two bits of the value have also been defined. Since rows in most of these tables will come and go with the running application elements whose information is contained in them, sysUpTime.0 is not appropriate as a discontinuity indicator for counters in these tables. By defining separate discontinuity indicators for the rows in these tables, entries can come and go as needed without causing other objects to appear to have discontinuities. As required by , the discontinuity indicators for the various information objects in these tables are identified in
the relevant DESCRIPTION clauses. Note that a discontinuity in one of these counters does not imply a sysUpTime.0 discontinuity, nor does a sysUpTime.0 discontinuity imply a discontinuity in any of these counters. 4.1. The service-level tables The service-level tables permit the identification of one or more instances of named services on a system, and the association of running application elements to these services. Service names are represented as human-readable strings, using values assigned by IANA where possible. The allocation of unique values for service instance identifiers is a local administrative issue; the values allocated must be constant for the lifetime of the service instance, and re-use of values should be avoided. It is important to understand that a service is not the same thing as a protocol. Rather, some services may be at least partially described by the protocol(s) used to provide that service. In deciding what should or should not be considered a service, the following factors merit consideration: - is there an identifiable set of resources associated with providing this service? - is there a reasonably long-lived server or client process? Following this reasoning, one can see where SMTP and HTTP service providers would be good candidates for classification as services for purposes of application management, where finger probably would not. Of course, implementors of this MIB are free to define additional services. An applicability statement may be an appropriate vehicle for standardizing how a specific service's information is reported using this MIB. 4.1.1. The service name to service instance table The service name to service instance table uses the service name as its primary key, and the service instance identifier as its secondary key. It facilitates the identification and lookup of the instances of a given service in a system.
4.1.2. The service instance to service name table The service instance to service name table uses the service instance identifier as its primary key, and the service name as its secondary key. Given a service instance identifier, it facilitates the lookup of the name of the service being provided. 4.1.3. The service instance to running application element table The service instance to running application element table uses the service instance identifier as its primary key, and the running application element index as its secondary key. This facilitates the identification of the set of running application elements providing a given instance of a service. 4.1.4. The running application element to service instance table The running application element to service instance table uses the running application element index as its primary key and the service instance identifier as its secondary key. It identifies the set of services provided by a given running application element. 4.2. The I/O channel group Information processed by an application can be modeled using the concept of a channel. Two kinds of channels, for example, are files and network connections. +-------+ | File | +---------+ /+-------+ +-------------+ | Generic | / | transaction |----| I/O |-------< | stream | | Channel | \ +------------+ +-------------+ +---------+ \ | open or | \| listening | | connection | +------------+ For each entry in the open channel table, there will be a corresponding entry in either the open file table or the open connection table. The information flowing on a channel may be structured as transactions. When the information flow on a channel is being monitored as a transaction stream, an entry in the transaction stream table will represent this fact and the associated information about
that stream. To facilitate traversal of these tables and retrieval of information relevant to a specific running application element or service instances, the initial indexes of these tables are the same. In each case, the first index determines whether the second index is interpreted as a running application element identifier or as a service instance identifier. The third index serves to uniquely identify a channel (and consequently, an open connection or file) in the context of a running application element or service instance. The transaction stream summary table contains per-stream summaries of transaction statistics. The transaction flow statistics table contains statistics broken into both transmit and receive counts for requests and responses on each stream. The transaction kind statistics table contains information further broken down by transaction kind. The transaction tables have a common structure for their indexing, with additional indexes added for increasing detail. The initial three indexes are the same as all the other tables in this group, serving to uniquely identify each transaction stream. 4.2.1. The open channels table The following information is available in this table: - time at which the channel was opened - number of read requests - number of bytes read - time at which most recent read operation was initiated - number of write requests - number of bytes written - time at which most recent write operation was initiated 4.2.2. The open files table The open files table contains one entry for each file in use by a manageable running application element. (See "Definitions of System-Level Managed Objects for Applications"  for a detailed definition of a running application element.) The purpose of this table is to identify the files in use and to record information
peculiar to files not already covered in the open channel table. If multiple running application elements open the same file, there will be an entry for each running application element opening that file. Similarly, if a running application element opens a file multiple times, there will be an entry in this table for the file corresponding to each open. The task of combining the information for file activity from this table (organized by running application element) into per-application statistics can be accomplished by a manager using the System Application MIB's  sysApplInstallPkgTable to find the installed application, the sysApplRunTable to find the running instances of that application, and the sysApplElmtRunTable to find the relevant values of sysApplElmtRunIndex. The manager, armed with a set of values for sysApplElmtRunIndex, is now able to retrieve the relevant portions of the applOpenFileTable and other tables in this MIB. The following information is available in this table: - file name - file size - current mode (read/write) of this file By convention, the names "stdin", "stdout" and "stderr" are used when these streams cannot be resolved to actual file names. 4.2.3. The open connections table This table provides information on channels that are open connections or listeners. The following information is available for each connection: - identification of the transport protocol in use - near-end address and port - far-end address and port - identification of the application layer protocol in use
4.2.4. The transaction stream summary table The transaction stream summary table contains per-stream summaries of transaction statistics. The simple model of a transaction used here looks like this: invoker | Request | performer | - - - - - - > | | | | Response | | < - - - - - - | | | Since in some protocols it is possible for an entity to take on both the invoker and performer roles, information here is accumulated for transmitted and received requests, as well as for transmitted and received responses. Counts are maintained for both transactions and bytes transferred. The information represented in this table includes: - identification of the underlying connection or file used for this transaction stream - a human-readable description of this stream - a human-readable description of this stream's notion of what a unit of work is - the cumulative amount of time spent (as an operation invoker) waiting for responses (from queueing of request to arrival of first response) - the cumulative amount of time spent (as an operation invoker) receiving responses (time from the arrival of the first response to the arrival of the last response in a series of responses to a particular request) - the cumulative amount of time spent (as an operation performer) handling requests (time from receipt of request to queueing of first outgoing response) - the cumulative amount of time spent (as an operation performer) sending responses (time from queuing of first response to the last response in a series of responses to a particular request)
- the cumulative number of transactions initiated (as an invoker) - the cumulative number of transactions processed (as a performer) 4.2.5. The transaction flow statistics table The transaction flow statistics table contains statistics broken into both transmit and receive counts for requests and responses on each stream. In addition to the service instance / running application element and transaction stream identifier indexes, rows in this table are indexed by flow direction (transmit or receive) and role (requests and responses). The information in this table includes: - the number of transactions processed - the number of bytes processed - the time at which the most recent transaction was processed in this flow 4.2.6. The transaction kind statistics table The transaction kind statistics table contains summary information organized by direction, request/response, and transaction kind for each stream. The indexing of this table is like that of the transaction flow table, with the addition of a transaction kind index. - number of transactions processed - number of bytes processed - the time at which the most recent transaction of this kind in this direction in this stream was processed 4.3. The former channel group The former channel group has several tables. The former channel control table controls the retention of history information by a running application element or service instance. The remaining tables parallel the structure of the channel group, with one significant difference in indexing structure. The closed channel index is independent from the open channel index.
4.3.1. The former channel control table The former channel control table provides control over the accumulation of information on former connections for running application elements and service instances. For each one, this table, indexed by the running application element or service instance index, controls whether information on former channels is accumulated, how many of these history records are retained, how long these are retained (within the lifetime of the process), and a count of history entries that were deleted before their expiration time in order to make room for new entries. 4.3.2. The former channel table The former channel table provides historical information on channels that have been closed. The number and lifetime of these entries is controlled, for each running application element or service instance, by the former channel control table. Most of the information in this table corresponds to information in the open channel table. For the connection or file-specific aspects of a given former channel, an entry will exist in the former connection table or in the former file table. 4.3.3. The former connection table For formerly open channels that were connections, connection-specific historical information is kept in the former connection table. For each entry in the former connection table, there will be an identically indexed entry in the former channel table. 4.3.4. The former file table For formerly open channels that were files, file-specific historical information is kept in the former file table. For each entry in the former file table, there will be an identically indexed entry in the former channel table. 4.3.5. The transaction history tables Two tables provide per-transaction-kind breakdowns for channels carrying transaction-structured flows. These tables are analogous to the transaction flow and kind statistics tables, with similar index structures.
4.4. The running element status and control group The running application element status and control group has two tables. 4.4.1. The running application element status table This table provides information for a running application element. Indexed by the sysApplElmtRunIndex, an entry in this table reports useful information on that running element's resource usage. Entries in this table contain: - current heap usage for this running application element - current number of open network connections for this running application element - the most recent error status message issued by this running application element Note that other information, such as the current number of open files for this running application element, is available from the sysapplElmtRunTable in . 4.4.2. The running application element control table This table provides rudimentary control over a running application element. Indexed by the sysApplElmtRunIndex, an entry in this table gives a manager with appropriate permissions the ability to suspend and resume processing by this running element, the ability to request reconfiguration, and the ability to terminate the running element. Variables in this table include: - a suspend/resume control - a reconfiguration request control - a termination request control