In order to make our commitment to our customers clear, we are disclosing our technical architecture here.
The basis or our architecture is Amazon AWS Lambda, which handles all the API calls. This has multiple benefits:
1. It's infinitely scalable — or scalable to any practical purposes, since Amazon probably has enough computing resources for anything.
2. It's more reliable than managing our/your own servers. Our service should theoretically run infinitely as long as the bills are paid. And we are profitable and we do pay our bills. If you see Amazon going out of business and no word from us, then it's time to worry a little.
3. In the near future, when Amazon starts supporting it, we will replicate our service globally in the most AWS Lambda regions. Currently it cannot be done easily, but Amazon is working on that feature. This will bring more reliability should decrease latency world wide.
As explained above, in the near future there will be no single points of failure in regards of the primary service (that of which AWS Lambda handles). There are still the backend services which handle billing/customer databases, which could become a point of failure. How we are handling these, is that we will not do any billing, or disrupt our main service in case the backend service goes offline. This should bring even more robustness into play.