Chapter 1: Introduction
- Ten Million X
- The Emerging Landscape of Graph Systems Thinking
- How Enterprise Knowledge Graphs are Transforming Organizations
- Why we need Systems Thinkers
Ten Million X
10 million to one. Eight orders of magnitude. A ten-million fold increase in performance. Response times that are one ten-millionth of the time they used to be.
This is what is happing to our enterprise information systems as we move from legacy relational databases to scalable knowledge graphs. We can break it down into three parts:
- Native graph databases allow for 1,000 fold better relationship traversal than RDBMS JOINs on million-row tables. That is our first three orders of magnitude change. Querys run 1,000x.
- Distributed graph databases allow 100 distinct servers to all work together to evenly divide the problem up and return query results in uniform structures. This gives us additional two orders-of-magnitude performance speedup. That is 100,000x
- New hardware that is specifically designed for optimizing graph traversals at scale will also give us our last three orders of magnitude speedup.
In total that is 3 + 2 + 3 = 8 or a 10,000,000
Thinking that we don't even recognize
Are fish aware of the water they swim in?
If you have been modeling data from many years, you have often been involved in
Old and New Assumptions
Old Assumption 1: JOIN Fear Modeling
- We must model all our data in the fewest number of tables because doing JOINs between two tables will slow our response times down.