Google Spanner is a major breakthrough in distributed database architecture and functionality. Azure Cosmos DB is not far behind. Microsoft engineer Dharma Shukla wants Cosmos DB “to be the database of the future” that will “last for many decades”. Both are revolutionary compared to the traditional relational database model.
The relational model was invented by IBM Scientist Edgar Codd in 1970, more than 40 years ago. IBM was slow in its go-to-market strategy, so Mr. Oracle Larry Ellison swooped in and the rest is history. Since that time, the relational model has gone through a thousand iterations, which leads us to Google Spanner and Azure Cosmos DB. Is Spanner and Cosmos DB really a revolutionary database that will support the technology industry for the next decade, although both have conceptual similarities to a 40+ year old idea?
We see Spanner and Cosmos DB as a short term solution and an incremental upgrade to the original relational model. Concepts such as CAP Theorem and ACID have been invented to describe the limitations of the current database model. CAP Theorem, which means Consistency, Availability and Partition Tolerance describes how a database can only provide 2 of the 3 attributes at one time. ACID stands for Atomicity, Consistency, Isolation and Durability, where a single transaction must be ACID compliant in order to avoid problems with consistency, duplication, and so on.
The relational model which was created in a pre-cloud, pre-IoT, pre-machine learning, pre-quantum computing and pre-iPhone time period (compared to today) is not the future, since they’re an updated version of a 40+ year old concept. What is needed is a radical new approach and concept, invented from the ground up that will support the technology industry for the next forty years. We don’t want to critique anyone’s work without contributing ideas in some form or another. Therefore, we’re going to be introducing a high level concept in a few months. Stay tuned.