SAP developed MPP technology over 3 years in parallel, then discontinued the SDCA leaving the TPA in dispute.
# Teradata sued. # SAP counter sued.
Judge Orrick whittled both parties claims down to a narrow set focused on Teradata’s Batched Merge method page
> Merge Statement in Teradata Merge statement is also referred as Upsert that > inserts a record to a table in a database if the record doesn’t exist. if the record already exists , update the existing record. > It combines the UPDATE and INSERT statements into a single statement with two conditional test clauses. > WHEN MATCHED, UPDATE. > WHEN NOT MATCHED, INSERT. > We can also delete the rows if the record matched by specifying : WHEN MATCHED,DELETE. The source and target table should have the same PI and PPI that helps to process the merge statement faster in Teradata. > WHEN MATCHED: > It updates the matched target table row with the values taken from the source table row. Also it deletes the matching target table row. > WHEN NOT MATCHED: > It inserts the current source row into the target table.
Teradata moves for summary judgment and argues that counterclaim-plaintiff SAP SE’s U.S. Patent No. 8,214,321 (“’321 Patent ”) is invalid under 35 U.S.C. § 101.
Summary judgment is a judgment entered by a court for one party and against another party without a full trial. - “hypothetical monopolist” test (“HMT”) tying claim - Inventive Concept Identified During the Hearing - Remaining Claim Elements (2003) - CONCLUSION - IT IS SO ORDERED.
Note on Insert vs Upsert
Syntax of Merge Statement in Teradata... provides great insight into big data in enterprise vs social.
Enterprise databases are optimized for immediate consistency to solve the Byzantine generals problem, especially in banking settlements (prevent withdrawal from 2 ATMs at the same time).
Social databases are optimized for immediacy without consistency. Immediacy provides instant messages without delay, the Byzantine generals problem, is a feature not a bug.
Blockchain separates optimizing for consistency from performing updates.
SAP sought out Teradata expertise at identifying matched data from unmatched data.
Matched data must be updated with consistency (merged). Unmatched data may be inserted (upserted) without consistency. page
The BigTable paper, the DynamoDB paper switched this around by treating all data as unmatched, inserting it into the database table, then looping around to merge any consistency columns with updates, thus decentralizing database consistency (my SWAG).
By marking any column cells that are currently inconsistent, we can defer any transactions that require enterprise consistency, by waiting for a few moments. Unnoticeable to most.
Perhaps, an abstraction blueprint to implement object capabilities as a database service.