Platform Dev. Guide - Fundamentals
- Ian Christopher (Deactivated)
The Ed-Fi ODS / API is a fairly large, complex system. This section covers the basic information you’ll want to know before diving in.
Made to Be Secure
Everyone working with student information wants to know that their data is secure and that students’ information is kept private. The Ed-Fi ODS / API was built from the ground up to provide developers with a solution to keep data secure and private.
Communication with the API is encrypted and only sent over HTTPS. The ODS / API uses OAuth 2 to handle client access, requiring clients to provide an access token before any data-related action is invoked. Once an access token has been verified, the ODS / API uses a claims-based system to determine whether the client is authorized to read (or otherwise work with) a particular kind of resource.
The authorization claim system provides a great deal of fine-grained control over what information is shared. Access can be limited by table/entity, row/record, or field/element, whichever is appropriate. For example, a student information system might be granted read/write access to a broad set of student profile and registration information, while an online assessment system might have its access limited to a read-only view of a class’ student roster and read-write access to assessment results information.
See the Security section of this documentation for implementation details.
Made to Be Extended
Out of the box, the ODS / API core data model covers a wide swath of information related to the K–12 domain, with a focus on student achievement. However, the data model is easily extended to handle information specific to your environment.
The database and the API have easy-to-implement design patterns to follow that allow you to do something as simple as add an element to a core object like the student model or as complicated as adding a new domain with its own entities and associations. Further, many of the supporting features are auto-generated. So, for example, the documentation for the API is automatically updated with your additions whenever you extend the core data model.
See the section on Extensibility & Customization in this documentation for details.
An Open Source, Customizable System
The Ed-Fi ODS / API is an enterprise-class customizable system. The source code distribution has a complete set of features and can easily be set up to run on a development machine or test environment – however, some analysis, planning, and development work are essential to put the complete system into production for an enterprise.
The Ed-Fi ODS / API is open source, meaning that all the source code is available to licensees who are free to modify and enhance the system. Many implementers choose to start with a core installation to get running quickly, and then customize or swap out modules to improve fit with their environment and to ready the solution for production deployment. For example, the solution includes log4net, a highly performant, configurable, and free system for event logging – but users can swap that module out and use the logging system of their choice.
The Ed-Fi Alliance also welcomes code contributions from field implementers – and hosts a platform called the Ed-Fi Exchange that allows licensees to share code modules. For a listing of available technology and information about how to contribute code, see exchange.ed-fi.org.
Unique ID Endpoints: Some Assembly Required
Most enterprises have their own existing unique ID system to look up and assign IDs to students, staff, and other individuals associated with their organization. Similar to most unique ID systems, the API assumes that the ID is unique across the enterprise, but that each individual is given only one enterprise-wide unique ID even if they’re in multiple roles (e.g., when a staff member is also a parent).
The Ed-Fi ODS / API comes with unique ID code that makes it easy to set up a development or test environment. However, integration and development are required to connect to an existing unique ID system. The Ed-Fi ODS / API does not replace an organization’s unique ID system, instead providing a standard interface for clients and clear extension points for platform developers to integrate with existing systems.
Platform developers can choose how best to integrate their unique ID system, if at all. The as-shipped configuration of the ODS / API returns an HTTP 501 - Not Implemented response for calls to the unique ID endpoints, so the system is logically complete and correct out of the box. However, most platform hosts will want to consider an integration model.
One integration model is to have the unique ID functionality "integrated" with the API, which means that the platform hosts have wired in their unique ID system to communicate with the ODS / API. As a consequence, the ODS / API contains the enterprise-wide unique ID for person entities, and uses that value as the primary ID attribute for those resources. This integrated approach means that client applications can rely on the API to be aware of that enterprise-wide unique ID when making updates or looking up a record by ID.
Another approach is a "non-integrated" model, which means that the ODS / API is not wired into an enterprise-wide unique ID system, but rather relies on client applications to either assign or obtain a unique ID outside the context of the ODS / API. This unique ID may be an enterprise-wide ID or simply a GUID assigned by clients – the key point just being that client applications are responsible for populating the unique value in the ODS / API.
See the technical article Unique ID System Integration for implementation details.
Keeping the unique ID API endpoints – and behavior – consistent regardless of the underlying Unique ID product allows flexibility for platform implementers. This approach also makes it easy for client application vendors to connect to different implementations of the Ed-Fi ODS / API platform without requiring custom code.
Key Structure
The Ed-Fi ODS/API uses natural keys as its primary means of enforcing uniqueness in records and maintaining relationships between records.
The data in the Ed-Fi ODS / API represents a rich domain with deep relationships. The ODS data model is organized into aggregates based on the principles of Domain-Driven Design, and those aggregates are exposed as API resources. Since the ODS is not the system of record (i.e., the system that creates and manages the data), primary keys are formed from the well-known, natural keys in the domain.
See the technical article Key Structure in the Ed-Fi ODS / API for additional details.
Support for Transactional & Bulk Modes
The API supports both a bulk loading model and a transactional model for updating data in the ODS.
Bulk loading is useful for solutions where data is updated in batches (e.g., for organizations that feed data on a nightly schedule). Bulk loading is also useful for the initial population of data (e.g., when connecting a new system or at the start of the year). A bulk-load API interface as well as command-line tools are provided for pre-loading large quantities of data.
A transactional model is useful once you have data in the system. Individual records and fields can be updated in real-time (or near real-time) by client applications.
The bulk load tools are created using the same code generation methods as are used for the transactional API.
Code Generation Wherever Possible
Much of the data access code and API surface have architectural patterns defined, so the Ed-Fi ODS / API leverages code generation techniques throughout the system.
The Visual Studio solution includes several T4 templates to generate many of the C# data-related objects based on the data model. During the build process, the (possibly extended) Ed-Fi Standard XSD, database structure, and metadata files provide inputs to the code generation process that generates the Ed-Fi ODS / API and Swagger documentation.
The client-facing documentation for the API is also auto-generated and consumed by the Swagger documentation website. The Extensibility & Customization section of this documentation provides additional detail on the code generation used by the solution.
Code generation from authoritative sources helps ensure that the API and Documentation are kept up-to-date in a way that manual maintenance cannot.
Built for Agile Development and Continuous Integration
The Ed-Fi ODS / API ships with unit tests and integration tests that provide assurance that the system functions as intended. Under Agile development methodologies, code introduced to fix existing issues or provide new features must have accompanying test coverage, and should not break any pre-existing tests (unless those tests should break because of changing requirements).
The code was designed to be modular and customizable. The Inversion of Control design pattern is used extensively along with other recognizable software patterns to maximize the flexibility and extensibility of the solution for implementers.
In the Continuous Integration software development methodology, a build and integration server automatically pulls down, builds, deploys, and tests code with every code commit into the version control system. This practice provides an automated authoritative build that can be deployed automatically to test or production servers. The Ed-Fi ODS / API solution is shipped with scripts that can be used as a basis for Continuous Integration.
Built for the Ed-Fi Dashboards
The Ed-Fi ODS / API is an ideal data source to populate an instance of the Ed-Fi Dashboards. The Dashboards display a range of student performance metrics in a compelling interface. The ODS database contains the full spectrum of information needed to populate the Dashboards, and the data structures in the ODS are tuned to the data loaders that ship with the Dashboards. This similarity allows for a simplified data import and metrics computation in organizations where the Ed-Fi Dashboards are used.