| Internet-Draft | DTNMA | March 2022 |
| Birrane, et al. | Expires 7 September 2022 | [Page] |
This document describes a management architecture suitable for deployment in the Delay-Tolerant Networking (DTN) architecture. The DTN environment is characterized by a lack of end-to-end connectivity and communications delays that are both long-lives and unpredictable. A DTN Management Architecture (DTNMA) is needed that can operate without human- or system-in-the-loop synchronous interactivity and without reliance on transport-layer sessions.¶
Within a DTNMA, nodes must provide both eventual data exchange and in-time local protection. This can be accomplished with the use of autonomous and asynchronous configuration and reporting. Despite the significant challenges to data transport, nodes must exhibit behavior that is both determinable and autonomous while maintaining as much compatibility with non-challenged-network operations concepts as possible.¶
The DTNMA is supported by two types of asynchronous behavior. First, the DTNMA does not presuppose any synchronized transport behavior between managed and managing devices. Second, the DTNMA does not support any query-response semantics. In this way, the DTNMA allows for operation in extremely challenging conditions, to include over uni-directional links and cases where delays/disruptions prevent operation over traditional transport layers.¶
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.¶
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.¶
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."¶
This Internet-Draft will expire on 2 September 2022.¶
Copyright (c) 2022 IETF Trust and the persons identified as the document authors. All rights reserved.¶
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Revised BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Revised BSD License.¶
The Delay-Tolerant Networking (DTN) Management Architecture (DTNMA) is designed to provide configuration, monitoring, and local control of application services on a managed device operating within a challenged network.¶
The unique properties of a challenged network are as defined in [RFC7228] and include cases where an end-to-end transport path may not be feasible at any moment in time and delivery delays may prevent timely communications between a network operator and a managed device. These delays may be caused by long signal propagations or frequent link disruptions (as described in [RFC4838]) or by non-environmental factors such as quality-of-service prioritizations and service-level agreements.¶
The DTNMA must operate in the most restrictive environments, to include the following.¶
In these and related scenarios, managed devices need to operate with local autonomy because managing devices may not be available within operationally-relevant timeframes. Managing devices deliver instruction sets that govern the local, autonomous behavior of the managed device. These behaviors include (but are not limited to) collecting performance data, state, and error conditions, and applying pre-determined responses to pre-determined events. The goal is asynchronous and autonomous communication between the device being managed and the manager, at times never expecting a reply, and with knowledge that commands and queries may be delivered much later than the initial request.¶
The DTNMA can leverage transport, network, and security solutions designed for challenged networks but is not bound to any single such protocol or protocol implementation. However the DTNMA is designed to be usable in all environments in which the Bundle Protocol (BPv7) [RFC9171] may be deployed.¶
This document describes the motivation, services, desirable properties, roles/responsibilities, logical data model, and system model that form the DTNMA. These descriptions comprise a concept of operations for management in challenged networks¶
This document is not a normative standardization of a physical data model or any individual protocol. Instead, it serves as informative guidance to authors and users of such models and protocols.¶
The DTNMA is independent of transport and network layers. It does not, for example, require the use of BP, TCP, or UDP. Similarly, it does not pre-suppose the use of IPv4 or IPv6.¶
The DTNMA is not bound to a particular security solution and does not presume that transport layers can exchange messages in a timely manner. It is assumed that any network using this architecture supports services such as naming, addressing, routing, and security that are required to communicate DTNMA messages as would be the case with any other messages in the network.¶
While possible that a challenged network may interface with an unchallenged network, this document does not specifically address compatibility with other management approaches.¶
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 [RFC2119].¶
The remainder of this document is organized into the following seven sections, described as follows.¶
Early work into the rationale and motivation for specialized management for the DTN architecture was captured in [BIRRANE1], [BIRRANE2], and [BIRRANE3]. Some of the properties and feasibility of a management system were adopted from prototyping work done in accordance with the DTN Research Group within the IRTF as documented in [I-D.irtf-dtnrg-dtnmp].¶
The unique nature of challenged networks requires new network capabilities to deliver expected network functions. For example, the unique constraints of the DTN achitecture required the development of BPv7 for transport functions and the Bundle Protocol Security Extensions (BPSec) [RFC9172]. Similarly, new management capabilities are needed to operate in these new environments and when using these new transport and security mechanisms.¶
This section discusses the characteristics of challenged networks and how they may violate the assumptions made by non-DTNMA network management approaches.¶
Constrained networks are defined as those where "some of the characteristics pretty much taken for granted with link layers in common use in the Internet at the time of writing are not attainable." ([RFC7228]). This somewhat information definition captures a variety of potential issues relating to physical, technical, or regulatory constraints to message transmission. Examples of these networks include nodes that regularly reboot or are otherwise turned off for long periods of time, that transmit at low or asynchornous bitrates, and that may have very low computational resources.¶
Separately, challenged network is defined as one that "has serious trouble maintaining what an application would today expect of the end-to-end IP model" ([RFC7228]). This includes cases where there is never simultaneous end-to-end connectivity, when such connectivity is interrupted at planned or unplanned intervals, or when delays exceed those that could be accommodated by IP-based transport. Links in such networks are often unavailable due to attenuations, propagation delays, mobility, occultation, and other limitations imposed by energy and mass considerations.¶
These networks exhibit the following properties that impact the way in which the function of network management is considered.¶
Finally, it is noted that "all challenged networks are constrained networks ... but not all constrained networks are challenged networks ... Delay-Tolerant Networking (DTN) has been designed to cope with challenged networks" ([RFC7228]).¶
Challenged networks differ from other kinds of constrained networks, in part, in the way in which the topology and, otherwise, roles and responsibilities of the network may evolve over time. From the time at which data is generated to the time at which the data is delivered, the topology of the network and the roles assigned to various nodes, devices, and actors, may have changed. In certain circumstances, the physical node receiving messages for a given logical destination may also have changed.¶
When this topological change impacts the transport of messages, then transports must wait for the incremental connectivity necessary to advance messages along their expected route. Therefore, these networks cannot guarantee that there exist timely data exchange between managing and managed devices. For example, the Bundle Protocol transport protocol for use in DTNs implements this type of store-and-forward operation.¶
When topological change impacts the semantic roles and responsibilities of nodes in the network, then local configuration and autonomy at nodes must be present to determine time-variant changes. For example, the BPSec protocol does not encode security destinations and, instead, requires nodes in a network to identify as verifiers or acceptors when receiving secured messages.¶
When applied to network management, the semantic roles of Agent and Manager may also change with the changing topology of the network. Individual nodes must implement desirable behavior without reliance on a single oracle of configuration or other coordinating function such as an operator-in-the-loop or various supporting infrastructure. This implies that there MUST NOT be a defined relationship between a particular manager and agent in a network. This includes allowing more than one manager per agent, allowing "control from" and "reporting to" managers be independent of one another, and allowing the logical role of a manager to be physically shared among assets and change over time"¶
Network management solutions are prevalent in both local-area and wide-area networks. Their capabilities range from simple configuration and report generation to complex modeling of a device's settings, state, and behavior. While all of these approaches are useful in the domains for which they have been built, they are not all equally functional when placed in a challenged network.¶
Solutions that push and pull large sets of data between managers and agents might fail to operate in a challenged environment as a function of tansmit power, bitrates, and the ability of the network to store and forward large data volumes over long periods of time.¶
Newer network management approaches are exploring the application of autonomy for management, in part to reduce overall data volumes and the need to maintain state between a manager and an agent. However, these focus more on well resourced, unchallenged networks where devices self-configure, self-heal, and self-optimize with other nodes in their vicinity.¶
This section describes some of the well known standardized protocols for network management as well as various proposed solutions. It aims to differentiate their purpose with the needs of challenged network management solutions.¶
Early network management tools for unchallenged networks provided mechanisms for communicating locally-collected data from devices to operators and managing applications, typically using a "pull" mechanism where data must be explicitly requested by a Manager to be produced and transmitted by an Agent.¶
The defacto example of this architecture is the Simple Network Management Protocol (SNMP) [RFC3416]. SNMP utilizes a request/response model to set and retrieve data values such as host identifiers, link utilizations, error rates, and counters between application software on Agents and Managers. Data may be directly sampled or consolidated into representative statistics. Additionally, SNMP supports a model for unidirectional push notification messages, called traps, based on predefined triggering events.¶
SNMP Managers can query Agents for status information, send new configurations, and be informed when specific events have occurred. Traps and queryable data are defined in a data model known as Managed Information Bases (MIBs) which define the information for a particular data standard, protocol, device, or application.¶
While there is a large installation base for SNMP there are several aspects of the protocol that make it inappropriate for use in a challenged network. SNMP relies on sessions with low round-trip latency to support its "pull" model. Complex management can be achieved but only through craftful orchestration using a series of real-time, end-to-end manager-generated query-and-response logic not possible in challenged networks.¶
The SNMP trap model provides some Agent-side processing, however the processing has low fidelity and traps are typically "fire and forget." Adaptive modifications to SNMP to support challenged networks and more complex application-level management, would alter the basic function of the protocol (data models, control flows, and syntax) so as to be functionally incompatible with existing SNMP installations. Therefore, this approach is not suitable for use in challenged networks.¶
Yet Another Next Generation (YANG) [RFC6020] is a data modeling language used to model configuration and state data of managed devices and applications. The YANG model defines a schema for organizing and accessing a device's configuration or operational information. Once a model is developed, it is loaded to both the client and server, and serves as a contract between the two. A YANG model can be complex, describing many containers of managed elements, each providing methods of device configuration or reporting of operational state.¶
YANG allows for the definition of parameterized of Remote Procedure Calls (RPCs) to be executed on managed nodes as well as the definition of push notifications within the model. The RPCs provide a means to execute commands on a device generating an expected, structured response. However, the RPC execution is strictly limited to those issued by the client. Commands are executed immediately and sequentially as they are received by the server, and there is no method to autonomously execute RPCs triggered by specific events or conditions.¶
YANG defines the schema for data to be accessed by network management protocols such as NETCONF [RFC6241], RESTCONF [RFC8040], or CORECONF [I-D.ietf-core-comi]. These protocols provide the mechanisms to install, manipulate, and delete the configuration of network devices.¶
NETCONF is a stateful, XML-based protocol that provides a RPC syntax to retrieve, edit, copy, or delete any data nodes or exposed functionality on the server. It requires that underlying transport protocols support long-lived, reliable, low-latency, sequenced data delivery sessions. NETCONF connections are required to provide authentication, data integrity, confidentiality, and replay protection through secure transport protocols such as SSH or TLS. No data is transferred without first establishing a bi-directional NETCONF session.¶
NETCONF provides the ability to udpate and fetch many data elements simultaneously using verbose XML files which are not easily compressed.¶
RESTCONF is a stateless RESTful protocol based on HTTP that uses JSON encoding to GET, POST, PUT, PATCH, or DELETE data nodes within YANG modules. It requires secure transport such as TLS.¶
RESTCONF configures or retrieves individual data elements or containers within YANG data models passing JSON over REST. However, large configuration changes of many data elements or collection of information depends greatly on synchronous communication.¶
CORECONF is also stateless and built atop the Constrained Application Protocol (CoAP) [RFC7252] which defines the messaging construct developed to operate specifically on constrained devices and networks by limiting message size and fragmentation. CORECONF defines the security requirements of either DTLS or OSCORE. COAP provides the ability to achieve store and forward operation similar to DTN, however it operates strictly at the application layer and requires specification of pre-determined proxies and moments of bi-directional communication.¶
CORECONF leverages the Concise Binary Object Representation (CBOR) [RFC8949] of YANG modules [I-D.ietf-core-yang-cbor] and provides further compressibility through the use of YANG Schema Item iDentifiers (SIDs) [I-D.ietf-core-sid]. While this offers various reductions in encoded size, it is still dependent on underlying transport networks and the organization of the YANG schema.¶
YANG notifications are promising for challenged network management, defined as both subscriptions to YANG notifications [RFC8639] and YANG PUSH notifications [RFC8641]. In this model, the client may subscribe to delivery of specific containers or data nodes defined in the model, either on a periodic or "on change" basis. The notification events can be filtered according to XPATH or subpath filtering as described in [RFC8639] Section 2.2. However generation events from the model are still captured and filtering only limits the amount of information sent to the receiver. Regardless, when a subscription to a notification element has been formed, those events will be streamed continuously until the session fails or the subscription is canceled. What is possibly the most limiting feature of YANG PUSH for challenged DTN, is the inability to trigger the generation of notifications based on external stimuli.¶
While the YANG model provides great flexibility for configuring a homogeneous network of devices, it becomes a burden in challenged networks where concise encoding is necessary. The YANG schema allows the developer of the model great flexibility in the organization of data. The YANG schema supports a broad range of data types noted in [RFC6991]. All the data nodes within a YANG model are referenced by verbose string based path of the module, sub-module, container, and any data nodes such as lists, leaf-lists, or leafs, without explicit hierarchical organization based on data or object type. Recent efforts for compressibility using CBOR and SIDs to address YANG data nodes through integer identifiers, however these lack any formal hierarchical structure. All mapping of SIDs to YANG modules and data nodes is preformed manually which limits the portability of models and further increases the size of any encoding scheme.¶
The future of network operations require more autonomous behavior including self-configuration, self-management, self-healing, and self-optimization. One approach to this is coined Autonomic Networking [RFC7575] and includes many recent efforts describe Autonomic architecture and protocols [RFC8993]. Challenged networks as described need similar degrees of autonomy however lack the ability to depend on complex coordination between nodes and both centralized and distributed supporting infrastructure. Policy-based management is another approach that has been around for quite some time, involving business and operations support systems that monitor and manage devices and networks in real-time. These systems leverage various existing network management protocols such as YANG Service module classification types [RFC8199] describe abstract services and support configuration of service level agreements. These then enact additional control over devices using network element modules. This approach is quite comprehensive but requires sufficient supporting infrastructure and synchronous access.¶
While the various protocols described above are quite useful and well deployed for various applications, they simply do not meet the various requirements for challenged networks. However, that does not mean various features from each do not contribute to the design of DTNMA. The concept of a data model for describing network configuration elements has long been useful for various protocols and ensures compliance between managers and managed devices. It provides error checking and bounds on operation which is necessary when controlling mission critical devices. The SNMP MIBs provide a well organized, hierarchical OIDs which support great compressibility needed by challenged DTNs. YANG, NETCONF, and RESTCONF support notifications much needed by DTN network management, however provide limited features for describing autonomous execution and behavior. The CORECONF provides CBOR encoding and concise reference via SIDS however without any hierarchical structure or authoritative planning to allocation, this will become too verbose and prove limiting in the future. the sections below describe desirable features of an DTNMA.¶
This section describes those design properties that are desirable when defining an architecture that must operate across challenged links in a network. These properties ensure that network management capabilities are retained even as delays and disruptions in the network scale. Ultimately, these properties are the driving design principles for the DTNMA.¶
An DTNMA built to support DTN must be agnostic of the underlying physical topology, transport protocols, security solutions, and supporting infrastructure. The DTNMA shall be limited to only the network management protocols, message structure, and information content, including but not limited to the type of objects to manage and the expected behavior and interaction upon access or execution of those objects. There shall be no prescribed association between between a manager and an agent other than those defined in the responsibilities associated with each in this document. There should be no limitation to the number of managers that can control an agent, the number of managers that an agent should report to, or any requirement that a manager and agent relationship implies a pair.¶
A means to define a shared contract between agent and manager has long been an approach to network management solutions. A model is a schema that defines this contract and defines all sources of information that can be retrieved, configured, or executed, as well as the various functions for parameterization, filtering, or event driven behavior. A model gives way to concise representation of information, intelligent suffixing, and patterning. The DTNMA model shall be designed with a limited set of object and data types to allow and be organized hierarchally to provide for highly compressible and concise encoding. This allows the agents and managers to infer context with limited link utilization necessary in DTN.¶
Pull management mechanisms require that a Manager send a query to an Agent and then wait for the response to that query. This practice implies a control-session between entities and increases the overall message traffic in the network. Challenged networks cannot guarantee that the round-trip data-exchange will occur in a timely fashion. In extreme cases, networks may be comprised of solely uni-directional links which drastically increases the amount of time needed for a round-trip data exchange. Therefore, pull mechanisms must be avoided in favor of push mechanisms.¶
Push mechanisms, in this context, refer to the ability of Agents to leverage rule-based criteria to determine when and what information should be sent to managers. This could be based solely off logic applied to existing VARs or EDDs, based off operations applied to data elements, or triggered as a function of relative time. Such mechanisms do not require round-trip communications as Managers do not request each reporting instance; Managers need only request once, in advance, that information be produced in accordance with a predetermined schedule or in response to a predefined state on the Agent. In this way information is "pushed" from Agents to Managers and the push is "intelligent" because it is based on some internal evaluation performed by the Agent.¶
Protocol designers must balance message size versus message processing time at sending and receiving nodes. Verbose representations of data simplify node processing whereas compact representations require additional activities to generate/parse the compacted message. There is no asynchronous management advantage to minimizing node processing time in a challenged network. However, there is a significant advantage to smaller message sizes in such networks. Compact messages require smaller periods of viable transmission for communication, incur less re-transmission cost, and consume less resources when persistently stored en-route in the network. A DTN Management Protocol (DTNMP) should minimize PDUs whenever practical, to include packing and unpacking binary data, variable-length fields, and pre-configured data definitions.¶
Elements within the management system must be uniquely identifiable so that they can be individually manipulated. Identification schemes that are relative to system configuration make data exchange between Agents and Managers difficult as system configurations may change faster than nodes can communicate.¶
Consider the following common technique for approximating an associative array lookup. A manager wishing to do an associative lookup for some key K1 will (1) query a list of array keys from the agent, (2) find the key that matches K1 and infer the index of K1 from the returned key list, and (3) query the discovered index on the agent to retrieve the desired data.¶
Ignoring the inefficiency of two pull requests, this mechanism fails when the Agent changes its key-index mapping between the first and second query. Rather than constructing an artificial mapping from K1 to an index, an AMP must provide an absolute mechanism to lookup the value K1 without an abstraction between the Agent and Manager.¶
Custom definition of new data from existing data (such as through data fusion, averaging, sampling, or other mechanisms) provides the ability to communicate desired information in as compact a form as possible. Specifically, an Agent should not be required to transmit a large data set for a Manager that only wishes to calculate a smaller, inferred data set. These new defined data elements could be calculated and used both as parameters for local stimulus-response rules-based criteria or simply serve to populate custom reports and tables. Since the identification of custom data sets is likely to occur in the context of a specific network deployment, AMPs must provide a mechanism for their definition.¶
Aggregation of controls and custom formatting of reports and tables are equally important. Custom reporting provides the flexibility allowing the manager to define the desired format of all information to be sent over the challenged network from the agents, serving to both save link capacity and increase the value of returned information. Aggregation of controls allows a manager to specify a set of controls to execute, specifying both the order and criteria of execution. This aggregate set of controls can be sent as a single command rather than a series of sequential operands. In this case it is additionally possible to use outputs of one command to serve as an input to the next at the agent.¶
DTNMA network functions must be achievable using only knowledge local to the Agent. Rather than directly controlling an Agent, a Manager configures an engine of the Agent to take its own action under the appropriate conditions in accordance with the Agent's notion of local state and time.¶
Such an engine may be used for simple automation of predefined tasks or to support semi-autonomous behavior in determining when to run tasks and how to configure or parameterize tasks when they are run. Wholly autonomous operations MAY be supported where required. Generally, autonomous operations should provide the following benefits.¶
The DTNMA provides a method of configuring DTNMA Agents with local, autonomous management functions, such as rule-based execution of procedures and generation of reports, to achieve expected behavior when managed devices exist over a challenged network. It further allows for dynamic instantiation and population of Variables and reports through local operations defined by the manager, as well as custom formatting of tables and reports to be sent back. This gives the DTNMA significant flexibility to operate over challenged networks, both providing new degrees of freedom over existing configuration based data models used in synchronous networks and allowing for more concise formatting over constrained networks. This architecture makes very few assumptions on the nature of the network and allow for continuous operation through periods of connectivity and lack of connectivity. The DTNMA deviates from synchronous management approaches because it never requires periods of bi-directional connectivity, and provides the manager flexibility to describe agent behavior that was unpredicted at the time of the data model creation.¶
This section identifies the services that an DTNMA would provide for management of challenged network resources. These services include configuration, reporting, autonomous parameterized control, and administration.¶
Configuration services update Agent data associated with managed applications and protocols. Some configuration data might be defined in the context of an application or protocol, such that any network using that application or protocol would understand that data. Other configuration data may be defined tactically for use in a specific network deployment and not available to other networks even if they use the same applications or protocols.¶
With no guarantee of round-trip data exchange, Agents cannot rely on remote Managers to correct erroneous or stale configurations from harming the flow of data through a challenged network.¶
Examples of configuration service behavior include the following.¶
Reporting services populate report templates with values collected or computed by an Agent. The resultant reports are sent to one or more Managers by the Agent. The term "reporting" is used in place of the term "monitoring", as monitoring implies a timeliness and regularity that cannot be guaranteed by a challenged network. Reports sent by an Agent provide best-effort information to receiving Managers.¶
Since a Manager is not actively "monitoring" an Agent, the Agent must make its own determination on when to send what Reports based on its own local time and state information. Agents should produce Reports of varying fidelity and with varying frequency based on thresholds and other information set as part of configuration services.¶
Examples of reporting service behavior include the following.¶
Similar to an RPC call, some mechanism MUST exist which allows a procedure to be run on an Agent in order to affect its behavior or otherwise change its internal state. Since there is no guarantee that a Manager will be in contact with an Agent at any given time, the decisions of whether and when a procedure should be run MUST be made locally and autonomously by the Agent. Two types of automation triggers are identified in the DTNMA: triggers based on the internal state of the Agent and triggers based on an Agent's notion of time. As such, the autonomous execution of procedures can be viewed as a stimulus-response system, where the stimulus is the positive evaluation of a state or time based predicate and the response is the function to be executed.¶
The autonomous nature of procedure execution by an Agent implies that the full suite of information necessary to run a procedure may not be known by a Manager in advance. To address this situation, a parameterization mechanism MUST be available so that required data can be provided at the time of execution on the Agent rather than at the time of definition/configuration by the Manager.¶
Autonomous, parameterized procedure calls provide a powerful mechanism for Managers to "manage" an Agent asynchronously during periods of no communication by pre-configuring responses to events that may be encountered by the Agent at a future time.¶
Examples of potential behavior include the following.¶
Administration services enforce the potentially complex mapping of auhorization to configuration, reporting, and control services amongst Agents and Managers in the network. Fine-grained access control can specify which Managers may apply which services to which Agents. This is particularly beneficial in networks that either deal with multiple administrative entities or overlay networks that cross administrative boundaries. Whitelists, blacklists, key-based infrastructures, or other schemes may be used for this purpose.¶
Other administrative services may place practical restrictions on the overall number of items that can be kept in a system. This includes items such as the number of rows kept by an Agent for a given table template or number of entries for a given report template.¶
Examples of administration service behavior include the following.¶
Note that the administrative enforcement of access control is different from security services provided by the networking stack carrying such messages.¶
By definition, Agents reside on managed devices and Managers reside on managing devices. There is however no pre-supposed architecture that connects managers and agents and therefore a single device could assume both roles. This section describes the responsibilities associated with each role and how these roles participate in network management.¶
Agents MUST provide mechanisms for operators in the network to use configuration services to create customized data definitions in the context of a specific network or network use-case. Agents MUST allow for the creation, listing, and removal of such definitions in accordance with whatever security models are deployed within the particular network.¶
Where applicable, Agents MUST verify the validity of these definitions when they are configured and respond in a way consistent with the logging/error-handling policies of the Agent and the network.¶
The DTNMA logical data model captures the types of information that should be collected and exchanged to implement necessary roles and responsibilities. The data model presented in this section does not presuppose a specific mapping to a physical data model or encoding technique; it is included to provide a way to logically reason about the types of data that should be exchanged in a DTN managed network.¶
The elements of the DTNMA logical data model are described as follows.¶
There are three fundamental representations of data in the DTNMA: (1) data whose values do not change as a function of time or state, (2) data whose values change as determined by sampling/calculation external to the network management system, and (3) data whose values are calculated internal to the network management system.¶
Data whose values do not change as a function of time or state are defined as Constants (CONST). CONST values are strongly typed, named values that cannot be modified once they have been defined.¶
Data sampled/calculated external to the network management system are defined as Externally Defined Data" (EDD). EDD values represent the most useful information in the management system as they are provided by the applications or protocols being managed on the Agent. It is RECOMMENDED that EDD values be strongly typed to avoid issues with interpreting the data value. It is also RECOMMENDED that the timeliness/staleness of the data value be considered when using the data in the context of autonomous action on the Agent.¶
Data that is calculated internal to the network management system is defined as a Variable (VAR). VARs allow the creation of new data values for use in the network management system. New value definitions are useful for storing user-defined information, storing the results of complex calculations for easier re-use, and providing a mechanism for combining information from multiple external sources. It is RECOMMENDED that VARs be strongly typed to avoid issues with interpreting the data value. In cases where a VAR definition relies on other VAR definitions, mechanisms to prevent circular references MUST be included in any actual data model or implementation.¶
Individual data values may be exchanged amongst Agents and Managers in the DTNMA. However, data are typically most useful to a Manager when received as part of a set of information. Ordered collections of data values can be produced by Agents and sent to Managers as a way of efficiently communicating Agent status. Within the DTNMA, the structure of the ordered collection is treated separately from the values that populate such a structure.¶
The DTNMA provides two ways of defining collections of data: reports and tables. Reports are ordered sets of data values, whereas Tables are special types of reports whose entries have a regular, tabular structure.¶
The typed, ordered structure of a data collection is defined as a Report Template (RPTT). A particular set of data values provided in compliance with such a template is called a Report (RPT).¶
Separating the structure and content of a report reduces the overall size of RPTs in cases where reporting structures are well known and unchanging. RPTTs can be synchronized between an Agent and a Manager so that RPTs themselves do not incur the overhead of carrying self-describing data. RPTTs may include EDD values, VARs, and also other RPTTs. In cases where a RPTT includes another RPTTs, mechanisms to prevent circular references MUST be included in any actual data model or implementation.¶
Protocols and applications managed in the DTNMA may define common RPTTs. Additionally, users within a network may define their own RPTTs that are useful in the context of a particular deployment.¶
Unlike tables, reports do not exploit assumptions on the underlying structure of their data. Therefore, unlike tables, operators can define new reports at any time as part of the runtime configuration of the network.¶
Tables optimize the communication of multiple sets of data in situations where each data set has the same syntactic structure and with the same semantic meaning. Unlike reports, the regularity of tabular data representations allow for the addition of new rows without changing the structure of the table. Attempting to add a new data set at the end of a report would require alterations to the report template.¶
The typed, ordered structure of a table is defined as a Table Template (TBLT). A particular instance of values populating the table template is called a Table (TBL).¶
TBLTs describes the "columns" that define the table schema. A TBL represents the instance of a specific TBLT that holds actual data values. These data values represent the "rows" of the table.¶
The prescriptive nature of the TBLT allows for the possibility of advanced filtering which may reduce traffic between Agents and Managers. However, the unique structure of each TBLT along may make them difficult or impossible to change dynamically in a network.¶
Low-latency, high-availability approaches to network management use mechanisms such as (or similar to) RPCs to cause some action to be performed on an Agent. The DTNMA enables similar capabilities without requiring that the Manager be in the processing loop of the Agent. Command execution in the DTNMA happens through the use of controls and macros.¶
A Control (CTRL) represents a parameterized, predefined procedure that can be run on an Agent. While conceptually similar to a "remote procedure call", CTRLs differ in that they do not provide numeric return codes. The concept of a return code when running a procedure implies a synchronous relationship between the caller of the procedure and the procedure being called, which is disallowed in an DTN management system. Instead, CTRLs may create reports which describe the status and other summarizations of their operation, and these reports may be sent to the Manager(s) calling the CTRL.¶
Parameters can be provided when running a command from a Manager, pre-configured as part of a response to a time-based or state-based rule on the Agent, or auto-generated as needed on the Agent. The success or failure of a control MAY be inferred by reports generated for that purpose.¶
NOTE: The DTNMA term control is derived in part from the concept of Command and Control (C2) where control implies the operational instructions that must be undertaken to implement (or maintain) a commanded objective. A DTN management function controls an Agent to allow it to fulfill its commanded purpose in a variety of operational scenarios. For example, attempting to maintain a safe internal thermal environment for a spacecraft is considered "thermal control" (not "thermal commanding") even though thermal control involves "commanding" heaters, louvers, radiators, and other temperature-affecting components. That said, CTRLs should be developed for specific autonomous and deterministic behavior where possible. Some controls may be designed to set configuration parameters or load complex policies, but there should be no assumption that it will be executed in real time.¶
Often, a series of controls must be executed in sequence to achieve a particular outcome. A Macro (MACRO) represents an ordered collection of controls (or other macros). In cases where a MACRO includes another MACRO, mechanisms to prevent circular references and maximum nesting levels MUST be included in any actual data model or implementation.¶
The DTNMA data model contains EDDs and VARs that capture the state of applications on an Agent. The model also contains controls and macros to perform actions on an Agent. A mechanism is needed to relate these two capabilities: to perform an action on the Agent autonomously in response to the state of the Agent. This mechanism in the DTNMA is the "rule" and can be activated based on Agent internal state (state-based rule) or based on the Agent's notion of relative time (time-based rule).¶
State-Based Rules (SBRs) perform actions based on the Agent's internal state, as identified by EDD and VAR values. An SBR represents a stimulus-response pairing in the following form: IF predicate THEN response The predicate is a logical expression that evaluates to true if the rule stimulus is present and evaluates to false otherwise. The response may be any control or macro known to the Agent.¶
An example of an SBR could be to initiate a thermal control self check if some internal temperature is greater than a threshold: IF (current_temp > maximum_temp) THEN thermal_control_self_check¶
Rules may construct their stimuli from the full set of values known to the network management system. Similarly, responses may be constructed from the full set of controls and macros that can be run on the Agent. By allowing rules to evaluate the variety of all known data and run the variety of all known controls, multiple applications can be monitored and managed by one Agent instance.¶
Time-Based Rules (TBR) perform actions based on the Agent's notion of the passage of time. A possible TBR construct would be to perform some action at 1Hz on the Agent.¶
A TBR is a specialization of an SBR as the Agent's notion of time is a type of Agent state. For example, a TBR to perform an action every 24 hours could be expressed using some type of predicate of the form: IF (((current_time - base_time) % 24_hours) == 0) THEN ... However, time-based events are popular enough that special semantics for expressing them would likely significantly reduce the computations necessary to represent time functions in a SBR.¶
Actions such as computing a VAR value or describing a rule predicate require some mechanism for calculating the value of mathematical expressions. In addition to the aforementioned DTNMA logical data objects, Literals, Operators, and Expressions are used to perform these calculations.¶
A Literal (LIT) represents a strongly typed datum whose identity is equivalent to its value. An example of a LIT value is "4" - its identifier (4) is the same as its value (4). Literals differ from constants in that constants have an identifier separate from their value. For example, the constant PI may refer to a value of 3.14. However, the literal 3.14159 always refers to the value 3.14159.¶
An Operator (OP) represents a mathematical operation in an expression. OPs should support multiple operands based on the operation supported. A common set of OPs SHOULD be defined for any Agent and systems MAY choose to allow individual applications to define new OPs to assist in the generation of new VAR values and predicates for managing that application. OPs may be simple binary operations such as "A + B" or more complex functions such as sin(A) or avg(A,B,C,D). Additionally, OPs may be typed. For example, addition of integers may be defined separately from addition of real numbers.¶
An Expression (EXPR) is a combination of operators and operands used to construct a numerical value from a series of other elements of the DTNMA logical model. Operands include any DTNMA logical data model object that can be interpreted as a value, such as EDD, VAR, CONST, and LIT values. Operators perform some function on operands to generate new values.¶
This section describes the notional data flows and control flows that illustrate how Managers and Agents within an DTNMA cooperate to perform network management services.¶
The DTNMA identifies three significant data flows: control flows from Managers to Agents, reports flows from Agents to Managers, and fusion reports from Managers to other Managers. These data flows are illustrated in Figure 1.¶
DTNMA Control and Data Flows¶
+---------+ +------------------------+ +---------+ | Node A | | Node B | | Node C | | | | | | | |+-------+| |+-------+ +-------+| |+-------+| || ||=====>>||Manager|====>>| ||====>>|| || || ||<<=====|| B |<<====|Agent B||<<====|| || || || |+--++---+ +-------+| ||Manager|| || Agent || +---||-------------------+ || C || || A || || || || || ||<<=========||==========================|| || || ||===========++========================>>|| || |+-------+| |+-------+| +---------+ +---------+
In this data flow, the Agent on node A receives Controls from Managers on nodes B and C, and replies with Report Entries back to these Managers. Similarly, the Agent on node B interacts with the local Manager on node B and the remote Manager on node C. Finally, the Manager on node B may fuse Report Entries received from Agents at nodes A and B and send these fused Report Entries back to the Manager on node C. From this figure it is clear that there exist many-to-many relationships amongst Managers, amongst Agents, and between Agents and Managers. Note that Agents and Managers are roles, not necessarily different software applications. Node A may represent a single software application fulfilling only the Agent role, whereas node B may have a single software application fulfilling both the Agent and Manager roles. The specifics of how these roles are realized is an implementation matter.¶
This section describes three common configurations of Agents and Managers and the flow of messages between them. These configurations involve local and remote management and data fusion.¶
The notation outlined in Table 1 describes the types of control messages exchanged between Agents and Managers.¶
| Term | Definition | Example |
|---|---|---|
| EDD# | EDD definition. | EDD1 |
| V# | Variable definition. | V1 = EDD1 + V0. |
| DEF([ACL], ID,EXPR) | Define ID from expression. Allow managers in access control list (ACL) to request this ID. | DEF([*], V1, EDD1 + EDD2) |
| PROD(P,ID) | Produce ID according to predicate P. P may be a time period (1s) or an expression (EDD1 > 10). | PROD(1s, EDD1) |
| RPT(ID) | A report identified by ID. | RPT(EDD1) |
This is a nominal configuration of network management where a Manager interacts with a set of Agents. The control flows for this are outlined in Figure 2.¶
Serialized Management Control Flow¶
+----------+ +---------+ +---------+
| Manager | | Agent A | | Agent B |
+----+-----+ +----+----+ +----+----+
| | |
|-----PROD(1s, EDD1)--->| | (1)
|----------------------------PROD(1s, EDD1)-->|
| | |
| | |
|<-------RPT(EDD1)------| | (2)
|<----------------------------RPT(EDD1)-------|
| | |
| | |
|<-------RPT(EDD1)------| |
|<----------------------------RPT(EDD1)-------|
| | |
| | |
|<-------RPT(EDD1)------| |
|<----------------------------RPT(EDD1)-------|
| | |
In a simple network, a Manager interacts with multiple Agents.¶
In this figure, the Manager configures Agents A and B to produce EDD1 every second in (1). Upon receiving and configuring this message, Agents A and B then build a Report Entry containing EDD1 and send those reports back to the Manager in (2). This behavior then repeats this action every 1s without requiring other inputs from the Manager.¶
This is a challenged configuration of network management where Agent B temporarily looses connectivity between the agent and the Manager. Flows in this case are outlined in Figure 3.¶
Challenged Management Control Flow¶
+----------+ +---------+ +---------+
| Manager | | Agent A | | Agent B |
+----+-----+ +----+----+ +----+----+
| | |
|-----PROD(1s, EDD1)--->| | (1)
|----------------------------PROD(1s, EDD1)-->|
| | |
| | |
|<-------RPT(EDD1)------| | (2)
|<----------------------------RPT(EDD1)-------|
| | |
| | |
|<-------RPT(EDD1)------| |
|<----------------------------RPT(EDD1)-------|
| | |
| | |
|<-------RPT(EDD1)------| |
|< RPT(EDD1)| (3)
| | |
| | |
|<-------RPT(EDD1)------| |
|< RPT(EDD1)|
| | |
| | |
|<-------RPT(EDD1)------| |
|<----------------RPT(EDD1, EDD1, EDD1)-------| (4)
| | |
In a challenged network, an agent must store and forward reports until links are available.¶
In this figure, the Manager configures Agents A and B to produce EDD1 every second in (1). Upon receiving and configuring this message, Agents A and B then build a Report Entry containing EDD1 and send those reports back to the Manager in (2). At step (3) the connection between Agent B and Manager is not available. The agent still generates the reports and stores locally using DTN protocols. At step (4) the link has been restored and all stored reports are successfully delivered to the manager.¶
This is a configuration of network management where Agent B has been configured to deliver two sets of data and demonstrates the Agent's responsibility to consolidate messages for transport. Flows in this case are outlined in Figure 4.¶
Consolidated Management Control Flow¶
+----------+ +---------+ +---------+
| Manager | | Agent A | | Agent B |
+----+-----+ +----+----+ +----+----+
| | |
|-----PROD(1s, EDD1)--->| | (1)
|----------------------------PROD(1s, EDD1)-->|
| | |
| | |
|<-------RPT(EDD1)------| | (2)
|<----------------------------RPT(EDD1)-------|
| | |
| | |
|----------------------------PROD(1s, EDD2)-->| (3)
| | |
| | |
|<-------RPT(EDD1)------| | (4)
|<--------------------------RPT(EDD1,EDD2)----|
| | |
| | |
|<-------RPT(EDD1)------| |
|<--------------------------RPT(EDD1,EDD2)----|
| | |
In a challenged network, Agents shall consolidate messages where possible.¶
In this figure, the Manager configures Agents A and B to produce EDD1 every second in (1). Upon receiving and configuring this message, Agents A and B then build a Report Entry containing EDD1 and send those reports back to the Manager in (2). At step (3), the manager configures Agent B to additionally report EDD2 every second. At step (4) Agent B proceeds to deliver EDD1 and EDD2 in the same report.¶
Networks spanning multiple administrative domains may require multiple Managers (for example, one per domain). When a Manager defines custom Reports/Variables to an Agent, that definition may be tagged with an Access Control List (ACL) to limit what other Managers will be privy to this information. Managers in such networks should synchronize with those other Managers granted access to their custom data definitions. When Agents generate messages, they MUST only send messages to Managers according to these ACLs, if present. The control flows in this scenario are outlined in Figure 5.¶
Multiplexed Management Control Flow¶
+-----------+ +-------+ +-----------+
| Manager A | | Agent | | Manager B |
+-----+-----+ +---+---+ +-----+-----+
| | |
|---DEF(A,V1,EDD1*2)-->|<-DEF(B, V2, EDD2*2)--| (1)
| | |
|---PROD(1s, V1)------>|<---PROD(1s, V2)------| (2)
| | |
|<--------RPT(V1)------| | (3)
| |--------RPT(V2)------>|
|<--------RPT(V1)------| |
| |--------RPT(V2)------>|
| | |
| |<---PROD(1s, V1)------| (4)
| | |
| |---ERR(V1 no perm.)-->|
| | |
|--DEF(*,V3,EDD3*3)--->| | (5)
| | |
|---PROD(1s, V3)------>| | (6)
| | |
| |<----PROD(1s, V3)-----|
| | |
|<--------RPT(V3)------|--------RPT(V3)------>| (7)
|<--------RPT(V1)------| |
| |--------RPT(V2)------>|
|<-------RPT(V3)-------|--------RPT(V3)------>|
|<-------RPT(V1)-------| |
| |--------RPT(V2)------>|
Complex networks require multiple Managers interfacing with Agents.¶
In more complex networks, any Manager may choose to define custom Reports and Variables, and Agents may need to accept such definitions from multiple Managers. Variable definitions may include an ACL that describes who may query and otherwise understand these definitions. In (1), Manager A defines V1 only for A while Manager B defines V2 only for B. Managers may, then, request the production of Report Entries containing these definitions, as shown in (2). Agents produce different data for different Managers in accordance with configured production rules, as shown in (3). If a Manager requests the production of a custom definition for which the Manager has no permissions, a response consistent with the configured logging policy on the Agent should be implemented, as shown in (4). Alternatively, as shown in (5), a Manager may define custom data with no access restrictions, allowing all other Managers to request and use this definition. This allows all Managers to request the production of Report Entries containing this definition, shown in (6) and have all Managers receive this and other data going forward, as shown in (7).¶
Data fusion reduces the number and size of messages in the network which can lead to more efficient utilization of networking resources. The DTNMA supports fusion of NM reports by co-locating Agents and Managers on nodes and offloading fusion activities to the Manager. This process is illustrated in Figure 6.¶
Data Fusion Control Flow¶
---------------------------------------
| Actor B |
| |
+-----------+ | +-----------+ +---------+ | +---------+
| Manager A | | | Manager B | | Agent B | | | Agent C |
+---+-------+ | +-----+-----+ +----+----+ | +----+----+
| | | | | |
|--------------------DEF(A,V0,EDD1+EDD2)->| | | (1)
|--------------------PROD(EDD1&EDD2,V0)-->| | |
| | | | | |
| | |--PROD(1s,EDD1)->| | | (2)
| | |--------------------PROD(1s, EDD2)->|
| | | | | |
| | |<---RPT(EDD1)----| | | (3)
| | |<--------------------RPT(EDD2)------|
| | | | | |
|<--------------------RPT(A,V0)-----------| | | (4)
| | | | | |
| | | | | |
| |
| |
---------------------------------------
Data fusion occurs amongst Managers in the network.¶
In this example, Manager A requires the production of a Variable V0, from node B, as shown in (1). The Manager role understands what data is available from what agents in the subnetwork local to B, understanding that EDD1 is available locally and EDD2 is available remotely. Production messages are produced in (2) and data collected in (3). This allows the Manager at node B to fuse the collected Report Entries into V0 and return it in (4). While a trivial example, the mechanism of associating fusion with the Manager function rather than the Agent function scales with fusion complexity, though it is important to reiterate that Agent and Manager designations are roles, not individual software components. There may be a single software application running on node B implementing both Manager B and Agent B roles.¶
This protocol has no fields registered by IANA.¶
Security within an DTNMA MUST exist in two layers: transport layer security and access control.¶
Transport-layer security addresses the questions of authentication, integrity, and confidentiality associated with the transport of messages between and amongst Managers and Agents in the DTNMA. This security is applied before any particular Actor in the system receives data and, therefore, is outside of the scope of this document.¶
Finer grain application security is done via ACLs which are defined via configuration messages and implementation specific.¶