Posted by Brian McCallion ● Jun 5, 2017 4:47:51 AM

Amazon RedShift Notes and Resources

When AWS announced Amazon RedShift at Reinvent 2012, our founder, Brian McCallion was one of the first people to be interviewed and asked how Redshift would work in a corporate computing environment.

Key Takeways:
In the enterprise there's a tendency to "standardize" on a few data stores. After all, staff, compute, budgeting, utilization can be best allocated towards a few key products. Yet the Cloud changes this. In fact, it turns out the "on-demand" and consumption based model works even better for high level services than it does for infrastructure level services.

RedShift combines both the scale-out, on-demand capability to add data store capacity, it also manages backups. It provides robust encryption, a feature we've found can cost hundreds of thousands of dollars even where a firm has already licensed a legacy data store.

Large Data Sets. RedShift can scale Out to a hundred or so nodes, about 1.7PB.

Cost. RedShift costs about $1000.00/TB / Year! And you only need to add nodes as you add data. Further, RedShift's columnar database architecture aligns extremely well with data compression. RedShift optimizes compression based on column, not row contents, and in this way achieves very high degrees of compression, even as this compression improves performance by reducing the I/O required to read a block of storage.

Topics: Amazon RedShift, AWS, Big Data, Columnar Database