Support of streaming/parsing/converting CLOB/BLOB type data structures in DV for larger data elements such as binary images that reside as objects on an S3

Requirement

Integrated Governance

In order to be able to use the CP4D Data Virtualization in combination with Watson Knowledge Catalog as a single, integrated governance layer over structured data (residing in Databases / Flat files) and unstructured data (residing in binary and character files in S3 Object Stores) we require the data types CLOB and BLOB to be available in Data Virtualization.

The Data Virtualization shall be able to read all sort of character and binary formats from files stored on S3 Object Stores. Data Virtualization shall be used to control access to the data of the underlying systems and present it as a CLOB/BLOB like data structure to the user, so they can use and combine it within a single SQL like interface.

This functionality corresponds to the "create external table […] LOCATION 's3a://<bucket>/folder'" functionality available in hive.

Metadata

Additionally object metadata stored in the S3 Object Store should be made available within the Data Virtualization engine, e.g {"Filename":"s3://mybucket/99/99/5412259999_1.TIF","File size":"417","Last mod":"Wed, 20 Oct 2021 13:07:30 GMT","MIME type":"binary/octet-stream","Storage":"STANDARD","MD5 sum":"7cff955d4673633b95ec9517854d1dbb","SSE":"AES256","Policy":"none","CORS":"none","ACL":"everybody"}, so that it can be used in data queries against the DV.

Caching

Caching of the data passed through the Data Virtualization shall be possible in order to speed up access to the objects.

Benefit

Currently WKC / DV can not be used as a single, integrated governance layer across various data locations. Instead the permissions on the S3 Object Storage have to be implemented separately. This process takes extra effort and is error prone.

The described functionality would enable us to use WKC / DV as integrated governance layer.