DCAE MOD (Microservice Onboarding & Design) is composed of a mix of components developed by the ONAP team and other components taken from the Apache Nifi project and modified for appropriate use.
The goal of the project is to stream line the construction, management, and evolution of DCAE flows from role to role, from environment to environment, and from release to release.
DCAE MOD is composed of three functional areas: onboarding, design, and distribution.
Below illustrations describe the architecture of DCAE-MOD and show the usage flow in DCAE MOD-
Following is MOD internal component architecture
It is used to onboard models/components/microservices (spec files) into DCAE MOD.
This is a Java project and offers a tool to generate a Nifi Processor POJO from a DCAE component spec.
It is a component taken from the Apache Nifi Project but modified for use in the MOD project.
Apache NiFi is a dataflow system based on the concepts of flow-based programming. It supports powerful and scalable directed graphs of data routing, transformation, and system mediation logic. NiFi has a web-based user interface for design, control, feedback, and monitoring of dataflows. It is highly configurable along several dimensions of quality of service, such as loss-tolerant versus guaranteed delivery, low latency versus high throughput, and priority-based queuing. NiFi provides fine-grained data provenance for all data received, forked, joined cloned, modified, sent, and ultimately dropped upon reaching its configured end-state.
The NiFi UI provides mechanisms for creating automated dataflows, as well as visualizing, editing, monitoring, and administering those dataflows. The UI can be broken down into several segments, each responsible for different functionality of the application. This section provides screenshots of the application and highlights the different segments of the UI. Each segment is discussed in further detail later in the document.
The users of Apache Nifi will find that it is used very differently than intended to serve our purpose in the DCAE-MOD project.
The Rest Api provides programmatic access to command and control a NiFi instance in real time. You can start and stop processors, monitor queues, query provenance data, and more.
A list of all available endpoints are available at https://nifi.apache.org/docs/nifi-docs/rest-api/index.html
Each endpoint includes a description, definitions of the expected input and output, potential response codes, and the authorizations required to invoke each service.
This component taken from the Apache Nifi project, is a REST API that provides an interface to a registry with operations for saving, versioning, reading NiFi flows and components.
It is written in the Flask framework in Python. It is a HTTP API to manage distribution targets for DCAE design. Distribution targets are DCAE runtime environments that have been registered and are enabled to accept flow design changes that are to be orchestrated in that environment.
NiFi’s fundamental design concepts closely relate to the main ideas of Flow Based Programming [fbp].
For more information on how some of the main NiFi concepts map to FBP, check out https://nifi.apache.org/docs/nifi-docs/html/overview.html
It is written in Java’s Spring Boot framework. It is a HTTP API to support runtime environment for DCAE-MOD. It has two major functionalities:
This tool allows the user to create a blueprint from a component spec json file. This tool is used by the runtime api.
DCAE Inventory is a web service that provides the following:
DCAE Inventory is a composite API that relies on other APIs to obtain resources on underlying components and uses these resources to compose a DCAE service resource. In addition, DCAE Inventory will store data that is unique to the DCAE service level including:
DCAE Inventory has a REST interface to service client requests. It has a well-defined query interface that would filter result sets based upon resource attributes.
Here, think of it as a back end API for the DCAE dashboard. The runtime posts Cloudify Blueprints to this API so they show up in the DCAE dashboard.
The DCAE dashboard provides visibility into running DCAE services for operational purposes. It queries the DCAE Inventory for aggregate details on all the running DCAE services and for getting up-to-date status information on DCAE services and their components.
A model/component/microservice can be onboarded by a ms Developer by posting a spec file on the onboarding API. Alternatively, an Acumos model can be onboarded using the Acumos Adapter. Once successfully onboarded, the genprocessor converts converts them to jars and onboards them into Nifi i.e DCAE MOD. These artifacts are now available to use from the Modified Nifi Web UI i.e DCAE Designer.
The registry api offers version control and retrieval for flows. The distributor api can be used to set distribution targets. Once a flow is designed and distributed, it goes to the distributor api which is supposed to post graph changes (in accordance with fbp) to the runtime api. The runtime api generates and distributes blueprints based on the change made on the flow-graph. These blueprints received by the DCAE inventory can then be viewed and deployed from the DCAE dashboard.