Hacker News Reader

What is your strategy for storage? It seems that because you're offering this as SaaS, your customers both set a high bar for durability of their data and your COGS would be very high if you just kept a sheet this big in memory all the time.

Do you have any plans for models which exceed the size that can be reasonably processed on one machine?

View on HN

Right now we keep our paid customer models' in memory. Models in free accounts are just evaluated every time (these models are usually much smaller). For now we want to avoid the complexity of distributed systems and just run this on very powerful machines with lots of RAM. For reference, one of our enterprise competitors has a calculation engine that can do 40B cells on a single host (using Java!).

that'd be Anaplan I assume?

Correct

How are you ensuring the durability of the data?

We store the model representation (formulas, data) in Postgres. We just keep it in memory for computing updates faster.

Reply on news.ycombinator.com