How Anaplan became so fast and flexible – thank you Steve Jobs

Share this with your network

Since our launch a few days ago I’ve been flooded with questions as to how we got to be so fast and flexible. Our primary goal building Anaplan was to be the fastest and easiest to use multi-dimensional modeling platform by Orders of Magnitude from existing ‘old-world’ solutions.

Technology advances over the last 20 years have certainly helped – 64 bit in-memory processing utilizing TB’s of Ram is a giant leap forward compared to disk processing. Innovating calculation update dependency graphs (the process of calculating millions of cells) provided another significant step forward.

But the real breakthrough came when we examined the Apple engineering model – do it ALL yourself to deliver exceptional performance. Apple has repeatedly delivered remarkable product innovation and performance thru its excellence in design and by developing all its hardware and software internally. Google has achieved similar performance with BigTable – its own database, to manage many of its applications.

Thus, if we were to deliver significant improvements to our multi-dimensional scalability and overall performance we had to build our own multi-dimensional database tightly coupled to our blazing fast calculation engine and feeding our own JAVASCRIPT grid object available in any browser.

The results have exceeded expectations. As many of you observed last week we now COMPLETELY update 700m cells of dense data < 1 sec, slicing and dicing at whim and changing/implementing model structures in seconds. Thank you Steve for showing us the way.

Leave a Reply