Is is recommended to use JSON Schema Validation in the place of Bean Validation JSR303 for Spring Boot Rest APIs?

Question

Can we use JSON Schema Validation in the place of Java Bean Validation JSR303 for Spring Boot Rest APIs for Enterprise Applications? Which one is more efficient to validate request Payload to Spring Boot Rest APIs?

(i.e. performance wise, cross-validation wise and RegEx pattern based validation)

Answer 1

It is a good question and there are no definitive answers for it as, perhaps, it is dependent on the application domain and remains subjective to that. At the base level (which usually covers 90%) of all use cases of validating user input to the REST service, both have the equivalent facility to validate data adequately. Both support constraints which can be used to achieve the same result.

However, on one front Bean Validation stands out is its ability to define custom validators, which can be used to validate very specific domain/application dependent constraints. For example, if there is case where a class which has 3 attributes (say, A,B and C) and a constraint is required that is either A occurs or B & C occurs but not both, then it is not really possible to put a constraint in JSON schema directly, it has to be handled through design of the schema (similarly in XML, actually it is more complicated with XML).

On the other hand in Bean Validation a custom validator can be written to handle this situation quite easily.

However, these kind of cases are few and far between.

Another point to consider is the integration of the Bean Validation in the underlying framework e.g. Spring, Jersey, RESTEasy etc., JSON schema validation is not yet integrated in the frameworks.

Therefore, given the support for the tech, it is perhaps better to stick with Bean Validation 2.0 and leverage the underlying frameworks capability to validation (this is, however, purely my view).

Answer 2

From an application development prospect, Java bean validator is sufficient for the business needs. From a system integration point, JSON schema externalizes the business rules and provides a platform independent interface control. So if your system involves many subsystems, JSON schema gives a better way to verify message payload.