Network Working Group K. White Request for Comments: 2758 IBM Corp. Category: Experimental February 2000 Definitions of Managed Objects for Service Level Agreements Performance Monitoring Status of this Memo This memo defines an Experimental Protocol for the Internet community. It does not specify an Internet standard of any kind. Discussion and suggestions for improvement are requested. Distribution of this memo is unlimited. Copyright Notice Copyright (C) The Internet Society (2000). All Rights Reserved.Abstract
This memo defines a Management Information Base (MIB) for performance monitoring of Service Level Agreements (SLAs) defined via policy definitions. The MIB defined herein focuses on defining a set of objects for monitoring SLAs and not on replication of the content of the policy definitions being monitored. The goal of the MIB defined within this document is to defined statistics related to a policy rule definition for reporting on the effect that a policy rule has on a system and to defined a method of monitoring this data.Table of Contents
1.0 Introduction . . . . . . . . . . . . . . . . . . . . . . . 2 2.0 The SNMP Network Management Framework . . . . . . . . . . 2 3.0 Structure of the MIB . . . . . . . . . . . . . . . . . . . 3 3.1 Scalar objects . . . . . . . . . . . . . . . . . . . . . . 4 3.2 slapmPolicyNameTable . . . . . . . . . . . . . . . . . . . 5 3.3 slapmPolicyRuleStatsTable . . . . . . . . . . . . . . . . 6 3.4 slapmPRMonTable . . . . . . . . . . . . . . . . . . . . . 6 3.5 slapmSubcomponentTable . . . . . . . . . . . . . . . . . . 8 4.0 Definitions . . . . . . . . . . . . . . . . . . . . . . . 8 5.0 Security Considerations . . . . . . . . . . . . . . . . . 67 6.0 Intellectual Property . . . . . . . . . . . . . . . . . . 67 7.0 Acknowledgments . . . . . . . . . . . . . . . . . . . . . 68 8.0 References . . . . . . . . . . . . . . . . . . . . . . . . 68 9.0 Author's Address . . . . . . . . . . . . . . . . . . . . . 70 10.0 Full Copyright Statement . . . . . . . . . . . . . . . . 71
1.0 Introduction
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, reference [13]. This document's purpose is to define a MIB module for performance management of Service Level Agreements (SLAs). It is assumed that an SLA is defined via policy schema definitions. The policy definitions being modeled with respect to performance management is primarily related to network Quality of Service (QOS). There are a number of methods that exist for defining and administering policy. Definition of these methods is considered out side of the scope of this document. The MIB module defined within this memo has been modeled using the various versions of the schema definitions being developed within the Policy Framework Working Group in the IETF. The content of the MIB defined within this memo has evolved along with the Policy Framework Working Group schema definitions.2.0 The SNMP Network Management Framework
The SNMP Management Framework presently consists of five major components: o An overall architecture, described in RFC 2571 [7]. o 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 [14], STD 16, RFC 1212 [15] and RFC 1215 [16]. The second version, called SMIv2, is described in STD 58, RFC 2578 [3], STD 58, RFC 2579 [4] and STD 58, RFC 2580 [5]. o Message protocols for transferring management information. The first version of the SNMP message protocol is called SNMPv1 and described in STD 15, RFC 1157 [1]. A second version of the SNMP message protocol, which is not an Internet standards track protocol, is called SNMPv2c and described in RFC 1901 [17] and RFC 1906 [18]. The third version of the message protocol is called SNMPv3 and described in RFC 1906 [18], RFC 2572 [8] and RFC 2574 [10]. o Protocol operations for accessing management information. The first set of protocol operations and associated PDU formats is described in STD 15, RFC 1157 [1]. A second set of protocol
operations and associated PDU formats is described in RFC 1905 [6]. o A set of fundamental applications described in RFC 2573 [9] and the view-based access control mechanism described in RFC 2575 [11]. 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.3.0 Structure of the MIB
The SLAPM-MIB consists of the following components: o scalar objects o slapmPolicyNameTable o slapmPolicyRuleStatsTable (equivalent to the deprecated slapmPolicyStatsTable) o slapmPRMonTable (equivalent to the deprecated slapmPolicyMonitorTable) o slapmSubcomponentTable Refer to the compliance statement defined within SLAPM-MIB for a definition of what objects and notifications MUST be implemented by all systems as opposed to those that MUST be implemented by end systems only. Initially most of the tables defined by the MIB module within this document where directly indexed using a policy's name and a subordinate traffic profile name. Over time the structure and resulting naming has grown more complex and as such has exceeded the capacity of being used as a direct MIB table index. As a result of this the original tables (slapmPolicyStatsTable and
slapmPolicyMonitorTable) have been deprecated and replaced with new tables that use an Unsigned32 index element instead of "names". A new table has been defined, slapmPolicyNameTable, that maps the Unsigned32 index to a unique name associated with a given policy rule definition.3.1 Scalar objects
Global objects defined within SLAPM-MIB: o slapmSpinLock Enables multiple management application access to SLAPM-MIB. An agent MUST implement the slapmSpinLock object to enable management applications to coordinate their use of the SLAPM-MIB. Management application use of slapmSpinLock is OPTIONAL. o slapmPolicyCountQueries, slapmPolicyCountAccesses, slapmPolicyCountSuccessAccesses, and slapmPolicyCountNotFounds Basic statistics on the amount of policy directory access that has occurred at a system. o slapmPolicyPurgeTime Used to prevent the entries in various SLAPM-MIB tables that relate to a policy definition from immediately being deleted when the corresponding policy definition no longer exists. This gives management applications time to discover this condition and close out any polled based interval data that may be being collected. All dependent slapmPRMonTable entries are also deleted when its parent slapmPolicyRuleStatsEntry is removed. Refer to the OBJECT description for slapmPolicyPurgeTime for a more precise description of this function. o slapmPolicyTrapEnable This object enables or suppresses generation of slapmPolicyRuleDeleted or slapmPolicyRuleMonDeleted notifications. o slapmPolicyTrapFilter This object enables suppression of slapmSubcMonitorNotOkay notifications.
3.2 slapmPolicyNameTable
The slapmPolicyNameTable maps a Unsigned32 index to a unique name associated with a given policy rule definition. Currently, the core schema definition being worked on within the Policy Framework working group defines five general classes: policyGroup, policyRule, policyCondition, policyTimePeriodCondition, and policyAction. "Policies can either be used in a stand-alone fashion or aggregated into policy groups to perform more elaborate functions. Stand-alone policies are called policy rules. Policy groups are aggregations of policy rules, or aggregations of policy groups, but not both." Each policy rule consists of a set of conditions and a set of actions. Policy rules may be aggregated into policy groups. "Instances in a directory are identified by distinguished names (DNs), which provide the same type of hierarchical organization that a file system provides in a computer system. A distinguished name is a sequence of relative distinguished names (RDNs), where an RDN provides a unique identifier for an instance within the context of its immediate superior, in the same way that a filename provides a unique identifier for a file within the context of the folder in which it resides." Each of these instances can also be named to fit in with the existing DEN practice with a commonName (cn) attribute as oppose to the classes name attribute. "The cn, or commonName, attribute is an X.500 attribute. It stands for commonName. It specifies a user-friendly name by which the object is commonly known. This name may be ambiguous by itself. This name is used in a limited scope (such as an organization). It conforms to the naming conventions of the country or culture with which it is associated. CN is used universally in DEN as the naming attribute for a class." An slapmPolicyNameEntry contains a single object, slapmPolicyNameOfRule, that contains the unique name associated with a policy rule instance. An slapmPolicyNameEntry is indexed by a Unsigned32 index, slapmPolicyNameIndex, that is assigned by the implementation of this MIB.
3.3 slapmPolicyRuleStatsTable
This table is functionally equivalent to the deprecated slapmPolicyStatsTable. The slapmPolicyStatsTable uses the name of both a policy definition and a traffic profile name to index an entry. The slapmPolicyRuleStatsTable uses an slapmPolicyNameEntry index (Unsigned32) instead. The slapmPolicyRuleStatsTable is the main table defined by SLAPM-MIB. The primary index for this table is slapmPolicyNameSystemAddress that enables support of multiple systems from a single policy agent. The index element, slapmPolicyNameSystemAddress, value must be either the zero-length octet string when at a policy agent only a single system is being support, 4 octets for a ipv4 address, or 16 octets for a ipv6 address. It is possible that on a single system multiple policy agent instances exists. The Entity MIB, refer to [19], should be used to handle the resulting MIBs. With respect to slapmPolicyNameSystemAddress one slapmPolicyRuleStatsEntry exists for each policy rule instance. Entries in this table are not administered via SNMP. An agent implementation for this table MUST reflect its current set of policy rule instances via table entries. The mechanisms for policy administration are outside of the scope of this memo.3.4 slapmPRMonTable
This table is functionally equivalent to the deprecated slapmPolicyMonitorTable. The slapmPolicyMonitorTable uses the name of both a policy definition and a traffic profile name to index an entry. The slapmPRMonTable uses an slapmPolicyNameEntry index (Unsigned32) instead. The slapmPRMonTable provides a method of monitoring the effect of SLA policy being used at a system. A management application creates an slapmPRMonEntry for each collection that it requires. The value of the BITS slapmPRMonControl object determines what type of monitoring occurs, at what level to monitor and whether trap support is enabled: o monitorMinRate(0) Use the value of slapmPRMonInterval as the interval to determine current traffic in and out rates, using slapmPRMonCurrentInRate and slapmPRMonCurrentOutRate, that can be compared to slapmPRMonMinRateLow for determining when to generate a slapmPolicyRuleMonNotOkay notification. The notification
slapmPolicyRuleMonOkay is generated when the problem is resolved. This can be determined by comparing the current rates to slapmPRMonMinRateHigh. o monitorMaxRate(1) Use the value of slapmPRMonInterval as the interval to determine current traffic in and out rate, using slapmPRMonCurrentInRate and slapmPRMonCurrentOutRate, that can be compared to slapmPRMonMaxRateHigh for determining when to generate a slapmPolicyRuleMonNotOkay notification. The notification slapmPolicyRuleMonOkay is generated when the problem is resolved. This can be determined by comparing the current rates to slapmPRMonMaxRateLow. o monitorMaxDelay(2) Use the value of slapmPRMonInterval as the interval to determine the current delay. This can be calculated on an aggregate level by averaging the round trip times for all TCP connections associated with the policy definition. For an individual subcomponent its round trip time can be used directly. Compare this value to slapmPRMonMaxDelayHigh for determining when to generate a slapmPolicyRuleMonNotOkay notification. The notification slapmPolicyRuleMonOkay is generated when the problem is resolved. This can be determined by comparing the current rates to slapmPRMonMaxDelayLow. UDP subcomponents don't support max delay monitoring. o enableAggregateTraps(3) The slapmPRMonitorControl BITS setting, enableAggregateTraps(3), MUST be set in order for any notifications relating to slapmPolicyRuleStatsTable monitoring to be generated. o enableSubcomponentTraps(4) This slapmPRMonControl BITS setting MUST be set in order for any notifications relating to slapmSubcomponetTable monitoring to be generated. The slapmPRMonControl BITS setting monitorSubcomponents(5) MUST be selected in order for this setting to be allowed. o monitorSubcomponents(5) If selected monitor slapmSubcomponentTable entries individually. Note: aggregate policy rule monitoring is always enabled.
The index element slapmPRMonOwnerIndex is used as the first index in slapmPRMonTable in order to enable SNMP VACM security control. The slapmPRMonTable is the only table that supports SNMP RowStatus operations.3.5 slapmSubcomponentTable
Entries are made into this table for the protocol entities (policy traffic profile subcomponents) to indicate actual policy rule usage, provide general statistics at either a TCP connection or UDP listener level, and enable subcomponent monitoring.