I2RS protocol strawman

This documents is a strawman for the I2RS Protocol from early I2RS design team discusses. It focuses on the protocol extensions for ephemeral data store. This draft provides suggests the following additions to support the I2RS ephemeral state: Yang ephemeral statement, NETCONF () protocol extensions for the ephemeral data store, RESTCONF () protocol extensions for the ephemeral data store draft-hares-i2rs-protocol-strawman-examples provides provides examples of this strawman protocol use for I2RS. This draft uses a simple thermostat model to illustrate commands. This draft is input to a NETCONF review and design team.

(dean)What edits are allowed in the ephemeral data store. Should those be syntactically correct or syntacticly and semantically? Validation slows things down a lot, so datastore validation needs to be considered carefully. This is where I think routing expertise will help decide how much validation can really be skipped for a particular use-case. (Anu)On priority: So the priority maximum will be same as the max-clients , if we are assigning unique priorities from 1 – max-clilents. (Eg , if max clients is 100 (leaf max clients , range 1 .. max ) then priority range is also from 1 – max ) So the leaf max-clients { .. range 1- 32 } represents priority information also , so it can be max-clients or max-priorities ?? (Andy)The term 'max' in YANG resolves to 4B-1 in this range, not 32. Actually, the text I sent allows for multiple clients to all have the default priority which is not good -- if client-id is needed to resolve collisions then there is no point to requiring a unique priority per client. (Andy)The priority is not required to be densely numbered. Whether there are 1 pane per client or 1 pane per priority or 1 giant blob full of everything, the code will be the same. The goal of "unique priority" is to require that only priority be saved in the meta-data for the ephemeral datastore. Without that, client-id and priority must be saved (per data node).

Currently the configuration systems managed by NETCONF () or RESTCONF () has three types of configuration: candidate, running, and startup running under the config=true flag. The candidate receives configuration changes from NETCONF/RESTCONF. The running configuration is the configuration currently operating on a devices The start-up configuration is the configuration that survives a reboot. The config=false flag has operational data which exists alongside the config=true data. However, at this point there is no datastored defined for configuration false.

........... ........... ........... :Candidate : --> : running : --> :start-up : ........... ........... ........... config true --------------------------------------------- config false =============== | operational | | data | ================ Figure 1 In reality, the running configuration becomes the intended configuration that is intended to be loaded into a device. The loading process of the intended configuation into a devices compares it against the actual devices and creates the actual configuration loaded into a box. Some people denote the actual configuration as applied configuration. The denotes the actual configuration as derived state. This document will use the term actual configuration.

........... ........... ........... :Candidate : --> : running : --> :start-up : ........... ........... ........... ============= | Intended | | config | ============= config true --------------------------------------------- config false ============= | Actual | | config | ============= ______________ | operational | | data | |______________| Figure 2 Recently the has proposed that intended configuration, actual configuration, and the traditional type of operational data included as operational state. Operational data may include: derived state (e.g. negotiated bgp hold timer) operational state for counters or statistics (interface counters) Again, this document will use the definitions above to discuss ephemeral state until the NETCONF WG agrees upon the changes to the state diagrams.

This section describes the properties of the ephemeral datastore. The ephemeral datastore is not unique to I2RS. This approach to the ephemeral datastore is a panes-of-glass model. This definition of I2RS does not support caching in the I2RS Agents. Future I2RS work may reconsidered supporting caching. The ephemeral data store has the following qualities: Ephemeral state is not unique to I2RS work. The ephemeral datastore is a datastore holds ephemeral configuration information that is intended to not survive a reboot. Configuration information is defined as "config=true nodes". Since Ephemeral is just about data not presisting over a reboot, then in theory every nodes in a yang data model could be ephemeral. The importance of tagging an "ephemeral node" is for conformance checking. Therefore, ephemeral nodes needs to be signalled in the conformance portions of the NETCONF and RESTCONF work. Conformance is signalled in the following ways: The conformance portion of NETCONF () is currently signalled in the <hello>. Yang 1.1 and RESTCONF uses the module library () NETCONF may use the module library in the future. The ephemeral datastore is never locked. The ephemeral datastore is one pane of glass that overrides the running data store. Ephemeral data can occur in three ways: protocol yang module with nodes that can be either non-ephemeral and ephemerally written, protocol yang modules with added nodes which can only be ephemeral protocol independent yang module which designed to be only ephemeral such as I2RS RIB, I2RS Topology models, and I2RS FB-RIB. However, ephemeral data nodes cannot have non-ephemeral data nodes within the subtree. Ephemeral sub-modules cannot have non-ephemeral data nodes wihin the module. Ephemeral modules cannot have non-ephemeral sub-modules or nodes within the module. Ephemeral nodes will be denoted by an "ephemeral config statement in the yang protocol at the node level and at the module level. Ephemeral provides two additional error handling features: Ephemeral data store allows for reduced error handling that removes the requirements for leafref checking, MUST clauses, and instance identifier. Ephemeral data store allows for priority preemption of the write operation. Priority preemption means each I2RS client of the ephemeral I2RS agent (netconf server) is associated with a priority. Priority preemption occurs when a I2RS client with a higher priority writes a node which has been written by an I2RS client (with the lower priority). At this point, the I2RS agent (netconf server) allows the write and provides a notification indication to the notification publication/subscription service.

Error handling is an I2RS protocol features. Normal error handling of I2RS Agent for an I2RS client's information examines the following: syntax validation for nodes of data model, referential validation for nodes of data model, grouping of data within a data models or across data models to accomplish tasks, permission to write nodes of data model, grouping, priority to write nodes of data model being higher than existing priority This section describes the ephemeral data stores handling of each of these error functions.

Syntax validation of the message should be done according to the NETCONF or RESTCONF protocol features. New features for ephemeral datastore should provide the error handling with the feature. Syntax validation of the data model included in the ephemeral data store should be done by the

The ephemeral data store normal processing does not do the following referencial checks: leafref, MUST, instance identifier. The removal of this validations allows for intelligence I2RS clients to rapidly read or write data, and handle error conditions at a higher level.

Yang 1.0 and Yang 1.1 provide the ability to group data in groupings, leafref lists, lists and containers. Data model group data in order to group data that is logical associated with one another. Data models may logical group data across groupings. One example of such an association is the association of a static route with an interface. The concepts of groupings apply to both ephemeral and non-ephemeral nodes within a data model. Error handling on writes of the ephemeral datastore is different for nodes that are grouped versus orthogonal. Group nodes may need to be all changed or all removed (all-or-nothing). In contrast, writing orthogonal data nodes in the same data module or betwen data models need to be added or deleted in sync. The specifies three types of error handling for a partial write operation: "all-or-nothing", "stop-on-error", or "continue-on-error". Partial write operations of "stop-on-error" or "continue-on-error" are allowed only for data writes which are not a part of a grouping within a data model. The definition of the I2RS error conditions are: stop-on-error - means that the configuration process stops when a write to the configuration detects an error due to write conflict. continue-on-error - means the configuration process continues when a write to the configuration detects an error due to write process, and error reports are transmitted back to the client writing the error. all-or-nothing - means that all of the configuration process is correctly applied or no configuration process is applied. (Inherent in all-or-nothing is the concept of checking all changes before applying.)

NETCONF does not support a mandated sequencing of edit functions or write functions. Without this mandated sequences, NETCONF cannot support partial edits.

RESTCONF has a complete set of operations per message. The RESTCONF patch can support accesing multiple data messages.

The initial releases of I2RS will only require "all-or-nothing" in the I2RS Agent.

I2RS protocol uses priority to resolve two I2RS clients having permissions to write the same pieces of data in an I2RS agent (NETCONF server). If two (or more) I2RS clients attempt to write the same data, the the one with the highest priority is enable to write the data. In the case of two clients with teh sample priority attempting to write data, the the first one to request write wins. Each client has a unique priority. Client identities and priorities are assigned outside of I2RS by exterior mechanisms such as AAA or adminstrative interfaces. A valid I2RS client must have both an identity and a priority. A sample container for I2RS client information is shown below.

container i2rs-clients { leaf max-clients { config false; mandatory true; type uint32 { range "1 .. max"; } } list i2rs-client { key name; unique priority; leaf name { ... } leaf priority { ... } } } Figure 3

The data modules supporting the Ephemeral datastore can use the Yang module to describe their datastore. Figure x shows the module library data structure. One part of the features of a module is the type of ephemeral support (module level, submodule level, or node level with a list of nodes). A feature list gives the reference to the identifier for thea ephemeral support. The feature references may allow for vendor extensions to ephemeral support for a specific model. Similarly, the deviation may point to a deviation for a ephemeral state model.

+--ro modules +--ro module*[name revision] +--ro name yang: yang-identifier +--ro revision union; +--ro schema? inet:uri +--ro namespace inet:uri +--ro feature* yang:yang-identifier +--ro deviation* [name revision] | +-- ro name yang:yang-identifier | +-- ro revision union +--ro conformance enumeration +--ro submodules +--ro submodule*[name revision] +--ro name yang:yang-identifier +--ro revision union +--ro schema? inet:uri

NETCONF's XML-based protocol () can operate over the following secure and encrypted transport layer protocols: SSH as defined in , TLS with X.509 authentication RESTCONF's XML-based or JSON data encodings of Yang functions are passed over http with (GET, POST, PUT, PATCH, DELETE, OPTIONS, and HEAD).

The ephemeral database may support insecure protocols for information which is ephemeral state which does not engage configuration. The insecure protocol must be defined in conjunction with a data model or a subdata model.

In this discussion of ephemeral configuration, this draft utilizes a simple thermostat model with the yang configuration found in figure 4.

module thermostat { .. leaf desired-temp { type int32; units "degrees Celsius"; description "The desired temperature"; } leaf actual-temp { type int32; config false; units "degrees Celsius"; description "The measured temperature (operational state)."; } } Figure 4 - Simple thermostat model yabng Figure 5 shows the diagram of the configuration state with the Simple thermostat model being attached to by an I2RS scheduler client receiving query information regarding intended configuration and actual configuration. Scheduler has a schedule set of temperatures to put in the thermostat. Actual temperature is operational state.

........... ............... ........... :Candidate : --> : Desired temp:-->:start-up : ........... ............... ........... | V ============ =========== | Intended |----| I2RS | | config | |scheduler| | | | client | ============ =========== config true ^ ------------------------------- | config false | ============= | | Actual |--------| | config | ============= ______________ | actual temp | |(operational | | state | | (op-state) | --------------- Figure 5 - Scheduler client only Figure 6 shows two I2RS clients talking to this model: scheduler and hold-temp. Scheduler has a schedule set of temperatures to put in the thermostat. Hold-temp holds the temperature at the same value. The hold-temp I2RS client has a higher priority than the scheduler client.

........... ............... ........... :Candidate :---: Desired temp : -- :start-up : ........... ............... ........... | | ============= | |I2rs Client| | |scheduler | V / ============ ................../ ephemeral . '''''''''''''''/. ============== datastore . 'desired-temp'---- |I2RS Client | . '''''''''|'''' . | hold temp | . | . ============== . | . ============ . |---------| intended | . . | config | . . ======||==== config true . . || -------------------------------------- || config false || ============= || | Actual |============ | config | ============= ______________ | actual temp | | (op-state) | ---------------- Figure 6 - Two I2RS clients

Yang needs to add a key word ephemeral at the leaf node that signal allowing a version of desired-temp in the ephemeral datatstore in yang.

module thermostat { .. leaf desired-temp { type int32; units "degrees Celsius"; ephemeral true; description "The desired temperature"; } leaf actual-temp { type int32; config false; units "degrees Celsius"; description "The measured temperature"; } } Figure 7 - Simple Thermostat Yang with ephemeral Figure 7 shows the thermostat model has emphemeral variable desired-temp in the running configuration and the ephemeral data store. The RESTCONF way of addressings is below:

RESTCONF running data store PUT /restconf/data/thermostat:desired-temp {"desired-temp":18} RESTCONF ephemeral datastore PUT /restconf/data/thermostat:desired-temp?datastore=ephemeral {"desired-temp":19 } Figure 8 - RESTCONF setting of ephemeral state The NETCONF way of transmitting this data would be

<rpc-message-id=101 xmlns="urn:ietf:params:xml:ns:base:1.0"> <edit-config> <target> <ephemeral> </target> <config> <top xmlsns="http:://example.com/schema/1.0/thermostat/config> <desired-temp> 18 </desired-temp> </top> </config> </edit-config> </rpc> Note: config=TRUE; datastore = ephemeral figure 8 NETCONF setting of desired-temp

capability-name: ephemeral-datastore

This capability defines the NETCONF protocol extensions for the ephemeral state. The ephemeral state has the following features: the ephemeral datastore is a datastore holds configuration information (Config=true) that is intended to not survive a reboot. The ephemeral capbility is signalled as a capability for a node, a sub-module, or a module either in the conformance portion of NETCONF (<hello>) or via netconf yang module library () used by Yang 1.1 and RESTCONF. ephemeral data will be doted by an "ephemeral statement at the node, module " The ephemeral datastore is never locked. Each client has a unique priority. The ephemeral data store is one pane of glass that overrides the intended config which is normally the running datastore, but can be designted as the candidate config. Ephemeral data nodes can occur as part of the following types of data modules: protocol dependent data models which mix non-ephemeral and ephemeral configuration data (config=true), protocol dependent data models which have only ephemeral data models, protocol independent data modules with only ephemeral data, However, ephemeral data nodes cannot have non-ephemeral data nodes within the subtree. Ephemeral sub-modules cannot have non-ephemeral data nodes wihin the module. Ephemeral modules cannot have non-ephemeral sub-modules or nodes within the module. ephemeral error checking allows for two additional options: reduced error checking that remove the requirement for leafref checking, MUST clauses, and instance identifier validation. write operation with a priority premption by a higher priority client of the lower priority clients write where the overwrite triggers a notification by the I2RS agent to the lower priority client.

The followign are the dependencies for ephemeral support: The Yang data modules must be flag with the ephemeral data store at the node, sub-module and model. The Yang data modules must be flag with the ephemeral data store. The Yang modules must support the notification of write-conflicts.

The ephemeral-datastore capability is identified by the following capability string: (capability uri)

The <link-ephemeral> allows the ephemeral datastore to be a pane of glass that impacts either the running-config configuration pane of glass or the candidate configuration pane of glass.

The bulk-write goes here if we need one. So far, editor cannot find a case.

The bulk-read goes here if we need one, so far the editor cannot find a case.

The capability for :ephemeral-datastore modifies the target for existing operations.

The :ephemeral-datastore capability modifies the <edit-config> to accept the <ephemeral> as a target for source, and allows the filters focused on a particular module, submodule, or node.

The :ephemeral-datastore capability modifies the <edit-config> to accept the <ephemeral> as a target for source with filters. The operations of merge, replace, create, delete, and remove are available, but each of these operations is modified by the priority write as follows: <merge> parameter is replaced by - merge-priority. The current data is modified by the new data in a merge fashion only if existing data either does not exist, or is owned by a lower priority client. If any data is replaced, this event is passed to the notification function within the pub/sub and traceability. <replace> is replaced by replace-priority - which only replaces data if the existing data is owned by a lower priority client. If data any data is replaced, this event is passed to the notification function within pub/sub and traceability for notification to the previous client. The success or failure of the event is passed to traceabilty. <create> - the creation of the data node works as in except that the success or failure is passed to pub/sub and traceability functions. <deletion> - the deletion of the data node works as in except event that the success or the error event is passed to the notiication function withi pub/sub and traceability functions. <remove> - the remove of the data node works as in except that all results are forwarded to traceabilty. The existing parameters are modified as follows: <target> - add a target of :emphemeral-datastore <default-operation> -sllows only <merge-priority> or <replace-priority> <error-option> - the I2RS agent agent has "stop-on-error", "continue-on-error", and "all-or-nothing" which follow the validation rules listed above. This also requires I2RS agents that support writes to have a "all-or-nothing"/"rollback-on-error" function. Note: The I2RS minimal function suggests that only error function that is required is the "all-or-nothing" function. positive response - the <ok> is sent for a positive response within an <rpc-reply>. negative response - the <rpc-error> is sent for a negative response within an <rpc-reply>.

Copy config allows for the complete replacement of all the ephemeral nodes within a target. The alternation is that source is the :ephemeral datastore with the fitlering to match the datasore.

The delete will delete all ephemeral nodes out of a datastore. The target must be changed to be ephemeral configuration and filters.

Lock and unlock are not supported with a target of :ephemeral-datastore.

The <get> is altered to allow a target of :ephemeral-datastore and with the filters.

The close session is modified to take a taret of "ephemeral-datastore" and to not release locks. The kill session is modified to take a target of "ephemeral-datastore, and to not change locks. "

defines NETCONF capabilities for writeable-running datastore, candidate config data store, confirmed commit, rollback-on-error, validate, distinct start-up, URL capability, and XPATH capability.

The writeable-running and the candidate datastore can be used in conjunction with the ephemeral data store. Ephemeral database overlays an intended configuration - either the writable-running or the c candidate configuration data store. The <link-ephemeral> operation links the two databases.

Confirmed commit capability is not supported for the ephemeral datastore.

The rollback-on-error when included with ephemeral state allows the error handling to be "all-or-nothing" (roll-back-on-error), "stop-on-error", and "continue-on-error". The error handling with I2RS ephemeral state is described above. Initial implementations of the I2RS agent are only required to support the default "roll-back-on-error". The use of the rollback-on-error capability allows the optional support of more capabiity in enhanced I2RS nodes.

The <validate> key word is expanded to support the following: source: ephemeral-datastore filters: reference to data node, sub-module or module.

This NETCONF capability appears to operate to load write-able running config, running-config, or candidate datastore. The ephemeral state does not change the environment based on this command.

The URL capabilities specify a <url> in the <source> and <target>. The initial suggestion to allow both of these features to work with ephemeral operation.

capability-name: ephemeral-datastore

This capability defines the RESTCONF protocol extensions for the ephemeral state. The ephemeral state has the features described in the previous section on NETCONF.

The ephemeral capabilities have the following dependencies: Yang data nodes, sub-modules, or modules must be flaged with the config datastore flag; The Yang modules must support the notification of write-conflicts. The I2RS Yang modules must support the following: the yang-patch features as specified in . The yang module library feature ,

The ephemeral-datastore capability is identified by the following capability string: (capability uri)

The bulk-write goes here.

The bulk-read goes here.

RESTCONF must be able to support the ephemeral datstore with its rules as part of the "{+restconf}/data" subtree. The "edit collision" features in RESTCONF must be able to provide notification to the I2RS pub/sub facility and the traceability functions. The "timestamp" with a last modified features must support the traceability function.

The current operations in RESTCONF are: OPTIONS, HEAD, GET, POST, PUT, PATCH, and DELETE. This section describes the modification to these exiting operations.

The options methods should be augmented by the information that will provide an indication of what ephemeral state exists in a data modules, or a data modules sub-modules or nodes.

The HEAD in retrieving the headers of a resources. It would be useful to changes these headers to indicate the datastore a node or submodule or module is in. (editor)TBD on how HEAD can be chaned to do this.

GET must be able to read from the URL and a particular datastore. (editor) TBD on how to filter for datastore in read.

POST must simply be able to create resources in ephemeral datastores and invoke operations defined in ephemeral data models using the rules of the ephemeral database.

PUT must be able to reference an ephemeral module, sub-module, and nodes.

Plain PATCH must be able to update or create child resources in an ephemeral datastore. The PATCH for the ephemeral state must be change to provide a merge or update of the original data only if the client's using the patch has a higher priority than an existing datastore's client, or if PATCH requests to create a new node, sub-module or module in the datastore.

The phrase "?datastore=ephemeral" following an element will specify the ephemeral data store when deleting entry.

The query parameters (content, depth, fields, insert, point, start-time, stop-time, and with-defaults (report-all, trim, explicit, report-all-tagged) must support ephemeral datastores described above.

The ephemeral database must support subscribing to receiving notifications as Event stream. The ephemeral database] support in RESTCONF must also support passing error information regarding ephemeral data access over to pub/sub client and traceability client.

TBD

This document is an attempt to distill lengthy conversations on the I2RS proto design team from August Here's the list of the I2RS protocol design team members Alia Atlas Ignas Bagdonas Andy Bierman Alex Clemm Eric Voit Kent Watsen Jeff Haas Keyur Patel Hariharan Ananthakrishnan Dean Bogdanavich Anu Nair Juergen Schoenwaelder Kent Watsen