Ayla Rule Service (ARS)

Created August 28, 2020. Last modified August 28, 2020.

This tech note offers a preview of the Ayla Rule Service which, driven by system events, performs user-defined actions based on user-defined rules. Design, implementation, and APIs are subject to change. The diagram below provides an overview:

The following descriptions support the diagram:

Digital Twin

A digital twin is a cloud-based state machine that reflects the attributes, properties, and schedules of a real-world thing like a thermostat. Device attributes are key-value pairs. Examples include the following:

{
  "device": {
    "id": 100,
    "product_name": "My Device",
    "dsn": "AC000W000000001",
    "oem": "1234abcd",
    "oem_model": "ledevb",
    "sw_version": "Version 1.2",
    "user_uuid": "00000000-0000-0000-0000-000000000000",
    "mac": "aabbcc112233",
    "lat": "00.0000",
    "lng": "00.0000",
    "ip": "000.000.000.000",
    "template_id": 000000,
    "registered": true,
    "activated_at": "0000-00-00T00:00:00Z",
    "connection_status": "Online",
    "connected_at": "0000-00-00T00:00:00Z"
  }
}

Device properties are objects with their own attributes. Typically, a single device would have many properties. Here is an example of one such property showing a subset of attributes:

{
  "property": {
    "key": 200,
    "device_key": 100,
    "name": "temperature",
    "base_type": "decimal",
    "read_only": true,
    "direction": "output",
    "value": 72.0
  }
}

Device schedules are similar to properties. Here is an example:

{
  "schedule": {
    "key": 300
    "device_id": 100,
    "name": "thermostat_schedule",
    "display_name": "thermostat_schedule",
    "active": false,
    "day_occur_of_month": null,
    "direction": "input",
    "duration": null,
    "end_date": "",
    "end_time_each_day": "00:00:00",
    "fixed_actions": true,
    "interval": null,
    "months_of_year": [],
    "days_of_month": [],
    "days_of_week": [],
    "start_date": "",
    "start_time_each_day": "00:00:00",
    "time_before_end": "",
    "utc": false,
    "version": "1"
  }
}

Event

An event is both (a) an occurence and (b) a record of the occurence created by the Ayla Cloud. A device event is a change in the state of a digital twin caused by an edge device, mobile app, RSS feed, schedule, or some other actor. A derived event, issued by a cloud subsystem, is the reinterpretation of a device event. The OTA Update subsystem, for example, issues version derived events based on connectivity, datapoint, and registration device events. The following table lists events of interest:

TypeCategoryDescription
activationDeviceThe value of the activated_at device attribute has changed. An activation event is associated, not with any particular device, but with a DSN. Activation occurs when a reserved DSN is first provisioned (i.e. associated with a device for the first time). Reassociating a DSN with a different device does not cause another activation event.
connectivityDeviceThe value of the connection_status device attribute has changed.
datapointDeviceThe value of a device property has changed.
datapointackDeviceThe cloud has received an ack from a device indicating that the device has changed the value of a to-device property on the device. See also Handling ack-enabled properties.
locationDeviceThe value the lat or lng device attribute has changed.
registrationDeviceThe value of the registered device attribute has changed.
versionDerivedOne of the following device attributes has change:
  • connection_status equals online
  • registered equals true
  • sw_version equals a new value

Rule

A rule includes the following fields:

{
  "rule": {
    "rule_uuid": "00000000-0000-0000-0000-000000000000",
    "name": "",
    "description": "",
    "expression": "",
    "action_ids": [],
    "is_enabled": true
  }
}

Rule expressions are assertions like these:

The Blue_button boolean property value of Device AC000W000000001 equals 1.
The Oxygen_Pct decimal property value of Device AC000W000000001 is less than 90.0.
The cmd string property value of Device AC000W000000001 equals "cmd_on".
The registered boolean attribute value of Device AC000W000000001 is true.

Expressions are written with Ayla Rule Expression Syntax (ARES). Here are the same expressions using ARES:

DATAPOINT(AC000W000000001,Blue_button) == 1
DATAPOINT(AC000W000000001,Oxygen_Pct) < 90.0
str_equals(DATAPOINT(AC000W000000001,cmd),'cmd_on')
REGISTRATION(AC000W000000001,true)

str_equals is a ARES function. Another similar one is str_contains.

Below is an example rule. If ARS finds that an event satisfies a rule expression, it performs the rule action(s).

{
  "rule": {
    "rule_uuid": "00000000-0000-0000-0000-000000000000",
    "name": "Blue_button == 1",
    "description": "",
    "expression": "DATAPOINT(AC000W000000001,Blue_button) == 1",
    "action_ids": [
      "00000000-0000-0000-0000-000000000001",
      "00000000-0000-0000-0000-000000000002",
      "00000000-0000-0000-0000-000000000003"
    ],
    "is_enabled": true
  }
}

The following core set of APIs enables clients to manage rules. See also the API Browser.

POST    /rulesservice/v1/rules
GET     /rulesservice/v1/rules
GET     /rulesservice/v1/rules/:ruleId
PUT     /rulesservice/v1/rules/:ruleId
DELETE  /rulesservice/v1/rules/:ruleId

Action

An action includes the following fields:

{
  "action": {
    "action_uuid": "00000000-0000-0000-0000-000000000000",
    "name": "",
    "description": "",
    "type": "",
    "parameters": {},
    "destination_ids": [],
    "is_enabled": true
  }
}

Action types include (but are not limited to) the ones in the table below. Refer Get Action Types to know the complete list of action types.

NameDescription
DATAPOINTSet a property value.Yes
URLPost data to an endpoint.Yes
AMS_EMAILEmail data.No
AMS_SMSText data.No
AMS_PUSHPush data.No
DATASTREAMStream data to Kinesis.No

Parameters are key-value pairs. Each action type has a set of possible parameter keys. Below, for example, is a DATAPOINT action with one parameter. The key is datapoint, and the value is an ARES expression that means "set input to 5":

{
  "action": {
    "action_uuid": "00000000-0000-0000-0000-000000000000",
    "name": "Set input to 5",
    "description": "",
    "type": "DATAPOINT",
    "parameters": {
      "datapoint": "DATAPOINT(AC000W000000001, input) = 5"
    },
    "is_enabled": true
  }
}

Below is a URL action with two parameters, body and endpoint:

{
  "action": {
    "action_uuid": "00000000-0000-0000-0000-000000000000",
    "name": "URL Action Test",
    "description": "",
    "type": "URL",
    "parameters": {
      "body": "{\"key1\":\"val1\",\"key2\":\"val2\"}",
      "endpoint": "https://docs.aylanetworks.com/api/v1/tests"
    },
    "is_enabled": true
  }
}

The body parameter value can contain placeholders which are replaced by ARS at run time with event and rule information. Here are examples:

"body": "{{{dpl_event}}}"
"body": "{\"event_metadata_dsn\":\"{{event.metadata.dsn}}\",\"event_user_uuid\":\"{{event.user_uuid}}\"}"

Both DATAPOINT and URL actions use parameters rather than destination objects to specify destination information. This may change as the dotted boxes in the diagram suggest:

The other types of actions, however, do use destination objects to specify destination information. Here is an example of an `AMS_EMAIL` action:
{
  "action": {
    "action_uuid": "00000000-0000-0000-0000-000000000000",
    "name": "Email List",
    "description": "",
    "type": "AMS_EMAIL",
    "parameters": {},
    "destination_ids": [
      "00000000-0000-0000-0000-000000000000",
      "00000000-0000-0000-0000-000000000000",
      "00000000-0000-0000-0000-000000000000"
    ],
    "is_enabled": true
  }
}

The following core set of APIs enables clients to manage actions. See also the API Browser.

POST    /rulesservice/v1/actions
GET     /rulesservice/v1/actions
GET     /rulesservice/v1/actions/:actionId
PUT     /rulesservice/v1/actions/:actionId
DELETE  /rulesservice/v1/actions/:actionId

Destination

A destination includes the following fields:

{
  "destination": {
    "uuid": "00000000-0000-0000-0000-000000000000",
    "type": "",
    "provider": ""
  }
}

Destination types include email, sms, and push. Providers are tied to type. email providers include smtp. sms providers include twilio and yunpian. push providers include fcm and apns. Depending on the type, destinations may include additional fields.

The following core set of APIs enables clients to manage destinations. See also the API Browser.

POST    /messageservice/v1/destinations
GET     /messageservice/v1/destinations
GET     /messageservice/v1/destinations/:destinationId
PUT     /messageservice/v1/destinations/:destinationId
DELETE  /messageservice/v1/destinations/:destinationId

Relationships

This section explores the relationships among events, rules, actions, and destinations:

Included in the diagram is ARS, the active process.

Event-to-Rule

ARS determines whether an event satisfies a rule expression. Consider the diagram below featuring a datapoint event. DPL stands for Data PipeLine.

To determine whether a `datapoint` event satisfies a `DATAPOINT` rule expression, ARS does the following:
  1. Compares event_type (datapoint) to the rule subject name (DATAPOINT).
  2. Compares dsn (AC000W000000001) to the first argument of the rule subject (AC000W000000001).
  3. Compares property_name (decimal_out) to the second argument of the rule subject (decimal_out).
  4. Compares the value (89) to the constant of the rule expression (90) using the < operator.
  5. Uses the Rule action_ids array to find and perform the associated action(s).

Rule-to-Action

The following diagram applies to both rule-to-action and action-to-destination relationships.

  • One rule can reference zero to many actions.
    • One rule can reference different types of actions.
      • More than one rule can reference the same action.
        • Actions do not reference rules. (This is under consideration.)
          • Deleting a rule does not delete an action.

Action-to-Destination

  • One action can reference zero to many destinations that are the same type as the action.
  • One action cannot reference different types of destinations.
  • Destinations do not reference actions.
  • Deleting an action does not delete a destination.

Example

This example shows you how to create and test one rule and one action.

Use Case

The example involves the following use case:

  1. You have a device called Edge Device (dsn = AC000W000000001) connected to the Ayla Cloud.
  2. Your device has three properties:
    NameTypeDirection
    `decimal_in`decimalTo Device
    `decimal_out`decimalFrom Device
    `Blue_LED`booleanTo Device
  3. Whenever the Edge Device host app receives a new decimal_in value, it sets decimal_out to the same value, and sends the new decimal_out value to the cloud. This is a technique for prompting your device to send certain values to the cloud as if originating at the device, thus simulating changes in temperature, pressure, gas levels, etc.
  4. When decimal_out is less than 90.0, you must set Blue_LED to 1. The first requires a rule, and the second an action.

Procedure

The diagram below, and the steps that follow, outline the procedure:

  1. Using the Aura Mobile App or another client, the user sets the digital twin decimal_in property value to 89.
  2. The Ayla Cloud writes decimal_in=89 to the device. (In reality, the cloud tells the device about the new value, and the device retrieves it.)
  3. The host app on the device sets decimal_out=89, and sends the new decimal_out value to the cloud. (Steps 1 and 2 are simply a mechanism for simulating a from-device property value change.)
  4. The Ayla Cloud creates a DATAPOINT device event representing the decimal_out property value change, and shares the event with the system.
  5. ARS realizes that the event satisfies the assertion in Rule A: decimal_out < 90.0.
  6. ARS initiates the associated set_datapoint action. At this stage, most action types (e.g. email, sms, post, stream) would need to retrieve destination information from an associated destination object. Currently, DATAPOINT and URL actions do not because they already contain destination information in parameters.
  7. The action sets the digital twin Blue_LED property value to 1.
  8. The Ayla Cloud writes Blue_LED=1 to the device.

Directions

To accomplish this use case, you need to create one action and one rule as follows:

  1. Open the API Browser in another tab.
  2. Click Accounts, choose a Region, enter email, password, app_id, and app_secret, click Get Tokens, and close the tab.
  3. Click Devices, select a device, and ensure that it includes the three properties listed above.
  4. Set decimal_in to 100, and set Blue_LED to 0.
  5. Click Rules Service, expand createAction, and create the following action:
    {
      "action": {
        "name": "Set Blue_LED 1",
        "type": "DATAPOINT",
        "parameters": {
          "datapoint": "DATAPOINT(AC000W000000001, Blue_LED) = 1"
        }
      }
    }
    
  6. Under Response Data, click Show, and note the new action_uuid. You need this in the next step.
  7. Expand createRule, and create the following rule using the action_uuid:
    {
      "rule": {
        "name": "decimal_out is less than 90",
        "description": "",
        "expression": "DATAPOINT(AC000W000000001,decimal_out) < 90.0",
        "action_ids": ["00000000-0000-0000-0000-000000000000"]
      }
    }
    
  8. Set decimal_in to 89.
  9. Check Blue_LED. It should be set to 1.

Expressions

ARS expects rule expressions and certain action parameters to be written with Ayla Rule Expression Syntax (ARES) which allows for arbitrary complexity and nesting.

Here are basic facts about rule expressions:

  1. Rule expressions evaluate to true or false. Below is a rule expression:
    DATAPOINT(AC000W000000001, Blue_LED) == 1
    
  2. Rule expressions include at least one rule subject. Below, the entire line is a rule subject:
    DATAPOINT(AC000W000000001, Blue_LED)
    
  3. Rule expressions can include many rule subjects. The one below includes two rule subjects:
    DATAPOINT(AC000W000000001, Blue_LED) == 1 && DATAPOINT(AC000W000000001, Green_LED) == 1
    
  4. Rule subject names are uppercase. Below, the rule subject name is CONNECTION:
    CONNECTION(AC000W000000001, online) == 0
    
  5. Rule expression space characters are optional.
  6. A rule subject name is also the rule subject type. It indicates the type of event the rule can evaluate:
    EventRule Subject Type
    activationACTIVATION
    connectivityCONNECTION
    datapointDATAPOINT
    datapointackDATAPOINTACK
    locationLOCATION
    registrationREGISTRATION
    versionVERSION
  7. Rule expressions use the following operators:
    OperatorMeaning
    &&AND
    ||OR
    >GT
    >=GTE
    <LT
    <=LTE
    ==EQ
    +Addition
    -Substraction
    *Multiplication
    /Division
    =Assignment
  8. Some rule expressions require functions to, for example, compare strings:
    str_equals(DATAPOINT(AC000W000000001,cmd),'cmd_on')
    
    See also Functions.
  9. Rule expressions can be concrete or abstract.

Concrete expressions

Here is a concrete rule expression which asserts that the Blue_button property value of device AC000W000000001 equals 1

The following list describes concrete rule expressions:
  1. They contain arguments that target specific entities (e.g. device, user) rather than types of entities (e.g. model, status).
  2. They are limited to the user scope. An EndUser can create concrete rules for owned entities.
  3. Evaluation causes a database query.

Concrete rules evaluate the following event types with the following rule subjects. Strikethroughs indicate not implemented.

Event TypeSubjectType
activationACTIVATION(dsn, activated|deactivated)boolean
connectivityCONNECTION(dsn, online|offline|all)boolean
datapointDATAPOINT(dsn, prop_name)Same as property
datapointackDATAPOINTACK(dsn, prop_name)boolean
locationLOCATION(dsn)(decimal,decimal)
locationLOCATION(uuid)(decimal,decimal)
registrationREGISTRATION(dsn, true|false|all)boolean

Abstract expressions

Here is an abstract rule expression which asserts that the decimal_out property value is LTE 90.0 for any device.

The following list describes abstract rule expressions:
  1. They contain terms that filter for types of entities (e.g. model, status) rather than specific entities (e.g. device, user).
  2. They are limited to oem scope. Only an OEM::Admin can create abstract rules.
  3. Evaluation does not cause a database query because abstract rules are cached in ARS memory.
  4. One OEM can have up to 100 abstract rules.
  5. Tags refer to field names within the Event JSON object.
  6. All terms with the ${ } construct must evaluate to true in order for the rule to evaluate to true.

Abstract rules evaluate the following event types with the following rule subjects. All rule subjects evaluate to true or false. Strikethroughs indicate not implemented.

Event TypeSubjectTags
activationACTIVATION(${})dsn, oem_model, status
connectivityCONNECTION(${})dsn, oem_model, status
datapointDATAPOINT(${})dsn, oem_model, property_name, value
locationLOCATION(${})TBD
registrationREGISTRATION(${})dsn, oem_model, registered
versionVERSION(${})oem_model, ayla_model, ayla_model_version, oem_model_version, base_mod_img_model, base_mod_img_version, from_mod_img_version, from_host_img_version

Events and rules

Under construction.

This section helps you visualize all relevant device events through the use of various techniques for generating events, various concrete and abstract rules, and a URL action similar to the following:

{
  "action": {
    "name": "URL dpl_event",
    "type": "URL",
    "parameters": {
      "body": "{{{dpl_event}}}",
      "endpoint": "https://domain.com/api/v1/ruletests"
    }
  }
}

The approach is illustrated in this diagram:

connectivity

This is a connectivity event:

{
  "dpl_event": {
    "metadata": {
      "oem_id": "1234abcd",
      "oem_model": "ledevb",
      "dsn": "AC000W000000001",
      "event_type": "connectivity"
    },
    "timestamp": "2020-08-28T09:45:36.000+0000",
    "connection": {
      "event_time": "2020-08-28T09:45:36Z",
      "user_uuid": "00000000-0000-0000-0000-000000000000",
      "status": "Online",
      "metadata": {
        "oem_id": "1234abcd",
        "oem_model": "ledevb",
        "dsn": "AC000W000000001",
        "event_type": "connectivity"
      }
    }
  }
}

This CONNECTION concrete rule references the action that posts the event to the endpoint:

{
  "rule": {
    "name": "CONNECTION rule (concrete)",
    "description": "",
    "expression": "CONNECTION(AC000W000000001, online) == 1",
    "action_ids": ["00000000-0000-0000-0000-000000000000"]
  }
}

datapoint

This is a datapoint event:

{
  "dpl_event": {
    "metadata": {
      "oem_id": "1234abcd",
      "oem_model": "ledevb",
      "dsn": "AC000W000000001",
      "property_name": "Blue_LED",
      "display_name": "Blue_LED",
      "base_type": "boolean",
      "event_type": "datapoint"
    },
    "datapoint": {
      "id": "00000000-0000-0000-0000-000000000000",
      "updated_at": "2020-08-28T09:36:15Z",
      "created_at": "2020-08-28T09:36:15Z",
      "echo": false,
      "closed": false,
      "value": "1",
      "metadata": {},
      "user_uuid": "00000000-0000-0000-0000-000000000000",
      "discarded": false,
      "scope": "user",
      "direction": "input"
    },
    "timestamp": "2020-08-28T09:36:15.000+0000"
  }
}

This DATAPOINT concrete rule references the action that posts the event to the endpoint:

{
  "rule": {
    "name": "DATAPOINT rule (concrete)",
    "description": "",
    "expression": "DATAPOINT(AC000W000000001, Blue_LED) == 1",
    "action_ids": ["00000000-0000-0000-0000-000000000000"]
  }
}

location

This is a location event:

{
  "dpl_event": {
    "metadata": {
      "oem_id": "1234abcd",
      "oem_model": "ledevb",
      "dsn": "AC000W000000001",
      "event_type": "location"
    },
    "location_event": {
      "user_uuid": "00000000-0000-0000-0000-000000000000",
      "dsn": "AC000W000000001",
      "ip": "67.255.234.73",
      "lat": " 44.787100",
      "provider": "user-based",
      "created_at": "2020-08-28T09:11:11Z",
      "long": "-64.123400"
    },
    "timestamp": "2020-08-28T09:11:11.813+0000"
  }
}

This LOCATION concrete rule references the action that posts the event to the endpoint:

{
  "rule": {
    "name": "LOCATION rule (concrete)",
    "description": "",
    "expression": "distance_miles(LOCATION(AC000W000000001),lat_long(42.3601,-71.0589) ) > 10",
    "action_ids": ["00000000-0000-0000-0000-000000000000"]
  }
}

registration

This is a registration event:

{
  "dpl_event": {
    "metadata": {
      "oem_id": "1234abcd",
      "oem_model": "linuxevb",
      "dsn": "AC000W000000001",
      "event_type": "registration"
    },
    "registration_event": {
      "dsn": "AC000W000000001",
      "user_uuid": "00000000-0000-0000-0000-000000000000",
      "registered": true,
      "registration_type": "Same-LAN",
      "registered_at": "2020-08-29T10:32:47Z",
      "unregistered_at": null
    },
    "timestamp": "2020-08-29T10:32:47.000+0000"
  }
}

This REGISTRATION concrete rule references the action that posts the event to the endpoint:

{
  "rule": {
    "name": "REGISTRATION rule (concrete)",
    "description": "",
    "expression": "REGISTRATION(AC000W000000001, true) == 1",
    "action_ids": ["00000000-0000-0000-0000-000000000000"]
  }
}

This REGISTRATION abstract rule references the action that posts the event to the endpoint:

{
  "rule": {
    "name": "REGISTRATION rule (abstract)",
    "description": "",
    "expression": "REGISTRATION(${oem_model == linuxevb, registered == true })",
    "action_ids": ["00000000-0000-0000-0000-000000000000"]
  }
}

Functions

Under construction.

Distance functions

distance_km

distance_km(LOCATION(dsn), LOCATION(uuid))

distance_miles

distance_miles(LOCATION(dsn), LOCATION(uuid))

lat_long

lat_long(latitude,longitude)

Retrieval functions

changed

changed(DATAPOINT(dsn, property_name))

String functions

str_contains

str_contains(DATAPOINT(dsn,prop_name), 'other string')

str_equals

str_equals(DATAPOINT(dsn,prop_name), 'foo')

Time/Date functions

current_time

current_time()

time

time(dsn)

time(uuid)

time('2017-05-26T20:40:22.000Z')