AWS Talks About Scalability
Database scalability has been one of my favorite topics ever since Microsoft’s Scalability Day in 1997, when Bill Gates demoed SQL Server 6.5 on Windows NT processing 1 billion transactions per day and 1 terabyte of data.
Here's what scalability looks like today, as shared by AWS during its Database Modernization Week.
Data volume is growing 10x every 5 years
AWS’s time series database, Amazon Timestream, can store trillions of events per day
AWS migrated 7,500 Oracle databases, and 75 petabytes of data, used by Amazon.com to its own databases. The two-year project was completed in 2019.
As we know, terabytes (TB) grow into petabytes (PB), exabytes (EB), and zettabytes (ZB). IDC forecasts worldwide data will swell to 175 ZB by 2025.
A surprising number of enterprises are dealing with exabytes, AWS CEO Andy Jassy said in 2016. At the time, AWS introduced Snowmobile, a 45-foot shipping container on wheels that it uses to move 100 PB at a time from customer data centers to AWS. That process would take 20 years over a 1 GBPS connection.
Database migrations can be unwieldy. Snowflake recently conceded these projects can take six months or longer. (Details here.) In fact, these challenges may have cost Snowflake a potential customer. Teradata CEO Steve McMillan said in February that an existing customer decided to stay with Teradata after facing “technical challenges and extended migration delays” in a planned move to Snowflake.
Oracle's New Cloud Lift Services
Vendors are introducing tools and services to smooth out the process. Oracle today announced Cloud Lift Services, to help customers like the Seattle Sounders FC and Cargill move data warehouses and other workloads to Oracle Cloud Infrastructure.
Here’s my take: Databases are growing by 10x, 100x, and more. Which means data migrations will continue to be a major “X factor” in the cloud database wars.