[SANJUKTA DAS SMITH] If you think about failures and repairs as arrival processes and departure processes that’s random. Completely random. So given that sort of uncertainty level of uncertainty, resource planning becomes a very critical and messy problem to deal with. So that’s where we started in sort of digging into this whole notion of okay how much redundancy is sufficient based on what the service level agreement, what the contract is specifying. Which, by the way, also includes the notion of penalty, right? So you on one hand you want to manage your cost, you don’t want to provide an abundance of redundancy which is going to eat into your profits. But on the other hand, if you don’t provide sufficient redundancy you run into the notion of penalties and, you know, loss of good will and all those things then ensue from that. So we developed a set of algorithms. So basically when we talk about algorithms what do we mean. It’s think of it as you could think of it as a software program that people can enter parameters into. And it is the algorithm is geared towards answering a very specific problem, okay. Which in our case boils down to this notion of redundancy that given that you have the system, given that you have this contract and this is these are all the things that your client is demanding, and if you don’t need those considerations then you
are facing this particular penalty. Now how much redundancy should you provision for. So you basically plug in information about the system that you are encountering, about the cloud data center that you’re running into that software system and out will come a prescription. So you could think of the algorithm as a way of doing prescriptive analytics in some sense. That you enter the parameters of your system and out pops an answer for what you should be doing.