In my recent travels, I’ve been speaking with database users at various meetups and trade shows worldwide. Very often, I got questions centering around the best use cases for our products, be it TokuDB, our MySQL storage engine, or, TokuMX, our distribution of MongoDB. Over 90% of the time, I responded Cloud, Big Data or both. You see, in the software industry we’re like kindergartners, we like things to fit into neat categories. If you know any software sales people, you’ll recognize this as a fitting analogy (at least in terms of energy and attention span), but I digress. This strategy helps allocate resources where they are most likely to make an impact, and, thus, optimize our return on investment. In this blog series, I’m going to go slightly against the grain and explain why the Fractal Tree makes databases work in these environments.
Before I get into more detail, let me start at the very foundation of what we do here at Tokutek; index data efficiently. Put simply, make databases do more with fewer resources. In the early to mid 2000s, our founders conceptualized and implemented a more efficient algorithm to ingest, sort and query data. This efficient algorithm became, what is now known as, the Fractal Tree or Fractal Tree Indexing. Since the initial implementation, the Fractal Tree has been iterated upon and become even more fast, efficient and durable. Fractal Tree Indexing is the cornerstone of both of our database products, TokuDB and TokuMX.
I’ve taken the time to explain this because I’m going to speak generally about the benefits of Fractal Tree Indexing instead of the benefits of either TokuDB or TokuMX. The NoSQL vs SQL (TokuMX vs TokuDB) debate is mainly about the best approach for a particular use case and has been the subject of many discussions that can be easily found via google search. Continuing over the next few weeks, I will be writing several posts explaining why the Fractal Tree is important and under what circumstances it can best be utilized.
In my next blog post, I’ll be demystifying the idea of big data, detailing why big data is expensive and explaining how the Fractal Tree will help. I will arm you with the knowledge, but if I may paraphrase my favorite “superheroes,” “…knowing is half the battle.” So, act from what you learn and download today!
I have an idea:
if fractal tree is made to be a file system(tokufs),then even I use innodb, I could get a high insert sppeed?Exactly, every application in sysytem could benefit from a high insert speed?