The primary purpose of Glue is to scan other services[3] in the same Virtual Private Cloud (or equivalent accessible network element even if not provided by AWS), particularly S3.[citation needed] The jobs are billed according to compute time, with a minimum count of 1 minute.[4] Glue discovers the source data to store associated meta-data (e.g. the table's schema of field names, types lengths) in the AWS Glue Data Catalog (which is then accessible via AWS console or APIs).[5]
